Death blow for Windows DirectX
David T. Johnson
David Sutherland
<djoh...@isomedia.com> wrote:
>A company named SCI has developed a bootable CD that contains a game and
>a Linux OS that can be booted on any Opteron PC to turn it into a
>powerful gaming console.
>
>[begin article excerpt]
>"SCI says its GameStorm technology fits onto a single CD and essentially
>turns the PC into an embedded Linux based "console-like" gaming system.
>The Linux OS scans the hardware, loads a custom distribution of 64-bit
>embedded Linux, and then runs the game software. The experience for the
>end-user is fast and powerful game playing that boots in under one
>minute, without the usual overhead from the legacy operating systems
>traditionally used in the gaming industry, SCI claims."
>[end article excerpt]
>
>http://www.linuxdevices.com/news/NS3103154102.html
>
>This is a BIG deal.
Is it? Bootable linux distributions have been around for a while.
Bundling a software title with it isn't particularly big news.
>Microsoft has been pushing their 'directx'
>technology as a way to get a high-quality gaming experience while still
>running Windows and is up to 'Directx 9' I think. Those of you who have
>struggled to get cranky directx Windows games to actually run like they
>are supposed to know what a miserable kludge the who directx platform is
How interesting - you're not sure what version of directx is current,
but you know it's a "miserable kludge"? Sounds like
wishful-thinking wrapped around ignorance to me.
>and will immediately see the value in the bootable linux CD as a gaming
>platform.
Why? What API are all the game developers supposed to write for? You
surely don't expect them to go back to writing directly to the video
hardware?
Also, what happens when new hardware comes out that isn't supported by
the version of linux on that CD? You've only got SATAII drives but
your now two year old game CD doesn't have drivers? Uh oh.
>So how will this benefit OS/2? One of the chief obstacles of
>playing Windows Directx games on OS/2 has been lack of access to the
>directx software.
Well yeah, I guess not having directx on OS/2 would hamper your
ability to use directx on OS/2!
> If Linux becomes a serious game platform replacement
>for Directx, it is likely that Linux games will be able to be either
>ported to OS/2
Oh my god, you *really* don't have clue, do you? It hasn't happened
with OpenGL games, what do you think has changed?
The whole point of DirectX is that it provides a standard set of API's
for developers to write towards so they can get on with the job of
creating *games* and adding their own custom effects instead of having
to worry about the basics of drawing a triangle and making it green.
The *hardware* developers can then concentrate on accelerating those
standard routines so that games run *fast*.
>or run on OS/2 with emulation software.
For this to happen you need to have a standard API supported across
both platforms - and the video card drivers need to accelerate it if
you don't want to play your games at one frame every four seconds.
The only graphics API shared by both Linux and OS/2 to my knowledge is
OpenGL. Last time I looked, OpenGL support on most cards for OS/2 was
pretty damn shabby where it wasn't non-existent. Linux is better, but
still not as good as Windows.
The other problem with OpenGL of course, is that the current standard
is years old, and lacks a lot of the cool features that DirectX allows
(particularly shader support) - the next OpenGL standard is due soon
which will address this, but the chances of it ever seeing the light
of day for the OS/2 platform are somewhere between a snowballs chance
in hell and zero.
No support means no ports.
>Microsoft has
>pushed Directx as a replacement for DOS as a gaming platform but no one
>has ever loved (or even liked) directx
*cough* - really? And on what evidence do you base this assertion?
DirectX and OpenGL have paved the way for modern gaming. What
developers used to hate was having to write the whole damn render
engine from scratch for DOS.
>and now it looks as if Linux has
>skewered it right through its gizzard.
LOL. Sure - maybe on your planet.
Regards,
David Sutherland
(note **ANTI-SPAM** in reply field)
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
> If Linux becomes a serious game platform replacement
>for Directx, it is likely that Linux games will be able to be either
>ported to OS/2 or run on OS/2 with emulation software.
I forgot to mention: If you have the OS and game on one
self-contained CD, why on earth would anybody go to weeks or months of
efforts to port to a *different* OS such as OS/2? The whole point is
that you don't *need* a pre-installed OS - it's bundled with it's own!
Your whole "this is good news for OS/2" song and dance routine is just
silly. OS/2 is rendered utterly irrelevant by such a product.
David T. Johnson
> On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
>
>
>>A company named SCI has developed a bootable CD that contains a game and
>>a Linux OS that can be booted on any Opteron PC to turn it into a
>>powerful gaming console.
>>
>>[begin article excerpt]
>>"SCI says its GameStorm technology fits onto a single CD and essentially
>>turns the PC into an embedded Linux based "console-like" gaming system.
>>The Linux OS scans the hardware, loads a custom distribution of 64-bit
>>embedded Linux, and then runs the game software. The experience for the
>>end-user is fast and powerful game playing that boots in under one
>>minute, without the usual overhead from the legacy operating systems
>>traditionally used in the gaming industry, SCI claims."
>>[end article excerpt]
>>
>>http://www.linuxdevices.com/news/NS3103154102.html
>>
>>This is a BIG deal.
>
>
> Is it? Bootable linux distributions have been around for a while.
> Bundling a software title with it isn't particularly big news.
AFAIK, this is the first 64-bit game ever bundled with a 64-bit
operating system that is provided complete on one bootable CD-ROM.
>
>
>>Microsoft has been pushing their 'directx'
>>technology as a way to get a high-quality gaming experience while still
>>running Windows and is up to 'Directx 9' I think. Those of you who have
>>struggled to get cranky directx Windows games to actually run like they
>>are supposed to know what a miserable kludge the who directx platform is
>
>
> How interesting - you're not sure what version of directx is current,
> but you know it's a "miserable kludge"? Sounds like
> wishful-thinking wrapped around ignorance to me.
Sounds like first-hand experience with DirectX to me...
>
>
>>and will immediately see the value in the bootable linux CD as a gaming
>>platform.
>
>
> Why? What API are all the game developers supposed to write for? You
> surely don't expect them to go back to writing directly to the video
> hardware?
<laugh> Not hardly. There are very few details available but the fact
that the game is only available in a 64-bit version running on the
Opteron and requires a platform called "Gamestorm" suggests that the
"gamestorm" platform is somehow doing something different than the old
32-bit OpenGl stuff. It is the new 'gamestorm' 64-bit platform that
apparently allows the bootable CD to provide a high-quality gaming
experience without the need for the OpenGL and video drivers for
specific hardware. This would be required if the game was to work
properly on the wide variety of video hardware that would be expected.
I would speculate that it has something to do using direct access to the
64-bit system memory for the video display in some way, as a guess. I
am just guessing, though.
>
> Also, what happens when new hardware comes out that isn't supported by
> the version of linux on that CD? You've only got SATAII drives but
> your now two year old game CD doesn't have drivers? Uh oh.
They seem to be doing something with the 64-bit hardware that makes the
'drivers' less critical. I will await further information, though.
>
>
>>So how will this benefit OS/2? One of the chief obstacles of
>>playing Windows Directx games on OS/2 has been lack of access to the
>>directx software.
>
>
> Well yeah, I guess not having directx on OS/2 would hamper your
> ability to use directx on OS/2!
yes.
>
>
>>If Linux becomes a serious game platform replacement
>>for Directx, it is likely that Linux games will be able to be either
>>ported to OS/2
>
>
> Oh my god, you *really* don't have clue, do you? It hasn't happened
> with OpenGL games, what do you think has changed?
the new 'gamestorm' 64-bit platform, whatever that is, which does not
seem to be using OpenGL.
>
> The whole point of DirectX is that it provides a standard set of API's
> for developers to write towards so they can get on with the job of
> creating *games* and adding their own custom effects instead of having
> to worry about the basics of drawing a triangle and making it green.
> The *hardware* developers can then concentrate on accelerating those
> standard routines so that games run *fast*.
perhaps the 64-bit memory addressing and hypertransport of the Opteron
allows some sort of new design paradigm?
>
>
>>or run on OS/2 with emulation software.
>
>
> For this to happen you need to have a standard API supported across
> both platforms - and the video card drivers need to accelerate it if
> you don't want to play your games at one frame every four seconds.
If the new 'gamestorm' 64-bit platform can boot independently and run on
any Opteron hardware without the need for special drivers, it may be
possible for it to also boot it on OS/2 in some way similar to the way
OS/2 can boot a 'DOS image' or 'boot from drive A' and such. I admit
that I am just wildly speculating here, though. Since there seems to be
no detailed information available, that is all there is to do.
>
> The only graphics API shared by both Linux and OS/2 to my knowledge is
> OpenGL. Last time I looked, OpenGL support on most cards for OS/2 was
> pretty damn shabby where it wasn't non-existent. Linux is better, but
> still not as good as Windows.
There does not seem to be a requirement for the 'gamestorm' 64-bit
platform to use OpenGL or else perhaps they have implemented it in some
sort of virtual hardware implementation or something. Obviously, the
'gamestorm' platform is doing something with 64-bit memory addressing
that was not possible to do with 32-bit addressing.
>
> The other problem with OpenGL of course, is that the current standard
> is years old, and lacks a lot of the cool features that DirectX allows
> (particularly shader support) - the next OpenGL standard is due soon
> which will address this, but the chances of it ever seeing the light
> of day for the OS/2 platform are somewhere between a snowballs chance
> in hell and zero.
>
> No support means no ports.
Perhaps OpenGL is morphing into something new and different or is being
superseded by something new and better. In any event, DirectX seems to
be disappering from the equation if it is possible to insert a CD-ROM
into an Opteron machine and boot and run a high-performance game on any
hardware.
>
>
>>Microsoft has
>>pushed Directx as a replacement for DOS as a gaming platform but no one
>>has ever loved (or even liked) directx
>
>
> *cough* - really? And on what evidence do you base this assertion?
> DirectX and OpenGL have paved the way for modern gaming. What
> developers used to hate was having to write the whole damn render
> engine from scratch for DOS.
>
>
>>and now it looks as if Linux has
>>skewered it right through its gizzard.
>
>
> LOL. Sure - maybe on your planet.
We'll see as more details become available.
--
David T. Johnson
Marty
>
> David Sutherland wrote:
> > On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
> > <djoh...@isomedia.com> wrote:
> >
> >>A company named SCI has developed a bootable CD that contains a game and
> >>a Linux OS that can be booted on any Opteron PC to turn it into a
> >>powerful gaming console.
> >>
> >>[begin article excerpt]
> >>"SCI says its GameStorm technology fits onto a single CD and essentially
> >>turns the PC into an embedded Linux based "console-like" gaming system.
> >>The Linux OS scans the hardware, loads a custom distribution of 64-bit
> >>embedded Linux, and then runs the game software. The experience for the
> >>end-user is fast and powerful game playing that boots in under one
> >>minute, without the usual overhead from the legacy operating systems
> >>traditionally used in the gaming industry, SCI claims."
> >>[end article excerpt]
> >>
> >>http://www.linuxdevices.com/news/NS3103154102.html
> >>
> >>This is a BIG deal.
> >
> > Is it? Bootable linux distributions have been around for a while.
> > Bundling a software title with it isn't particularly big news.
>
> AFAIK, this is the first 64-bit game ever bundled with a 64-bit
> operating system that is provided complete on one bootable CD-ROM.
Never heard of a Playstation 2, Dreamcast, or GameCube, eh? <laugh>
> <laugh> Not hardly. There are very few details available but the fact
> that the game is only available in a 64-bit version running on the
> Opteron and requires a platform called "Gamestorm" suggests that the
> "gamestorm" platform is somehow doing something different than the old
> 32-bit OpenGl stuff. It is the new 'gamestorm' 64-bit platform that
> apparently allows the bootable CD to provide a high-quality gaming
> experience without the need for the OpenGL and video drivers for
> specific hardware.
Yeah. It uses magical pixie dust instead.
> If the new 'gamestorm' 64-bit platform can boot independently and run on
> any Opteron hardware without the need for special drivers, it may be
> possible for it to also boot it on OS/2 in some way similar to the way
> OS/2 can boot a 'DOS image' or 'boot from drive A' and such. I admit
> that I am just wildly speculating here, though. Since there seems to be
> no detailed information available, that is all there is to do.
OS/2 doesn't even fully virtualize 386 hardware. There's no chance that
this could work in a VDM or VMB.
> Perhaps OpenGL is morphing into something new and different or is being
> superseded by something new and better. In any event, DirectX seems to
> be disappering from the equation if it is possible to insert a CD-ROM
> into an Opteron machine and boot and run a high-performance game on any
> hardware.
That's great if you want a standalone gaming box like a game console.
For normal users, they like their computers to do other things while
running their games.
Fascinating that you've taken such a sudden interest in gaming, though.
<laugh>
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>>
>>>A company named SCI has developed a bootable CD that contains a game and
>>>a Linux OS that can be booted on any Opteron PC to turn it into a
>>>powerful gaming console.
>>>
>>>[begin article excerpt]
>>>"SCI says its GameStorm technology fits onto a single CD and essentially
>>>turns the PC into an embedded Linux based "console-like" gaming system.
>>>The Linux OS scans the hardware, loads a custom distribution of 64-bit
>>>embedded Linux, and then runs the game software. The experience for the
>>>end-user is fast and powerful game playing that boots in under one
>>>minute, without the usual overhead from the legacy operating systems
>>>traditionally used in the gaming industry, SCI claims."
>>>[end article excerpt]
>>>
>>>http://www.linuxdevices.com/news/NS3103154102.html
>>>
>>>This is a BIG deal.
>>
>>
>> Is it? Bootable linux distributions have been around for a while.
>> Bundling a software title with it isn't particularly big news.
>
>AFAIK, this is the first 64-bit game ever bundled with a 64-bit
>operating system that is provided complete on one bootable CD-ROM.
>
That's great, but it's still just a bootable OS bundled with a game.
Self-contained bootable Linuxes have been around for years. The
"bitness" doesn't change it's nature. Particularly when the title you
are so excited about is already available for 32bit Linux, and it's
just been recompiled for 64bit Linux. Big deal, right :)
>>
>>
>>>Microsoft has been pushing their 'directx'
>>>technology as a way to get a high-quality gaming experience while still
>>>running Windows and is up to 'Directx 9' I think. Those of you who have
>>>struggled to get cranky directx Windows games to actually run like they
>>>are supposed to know what a miserable kludge the who directx platform is
>>
>>
>> How interesting - you're not sure what version of directx is current,
>> but you know it's a "miserable kludge"? Sounds like
>> wishful-thinking wrapped around ignorance to me.
>
>Sounds like first-hand experience with DirectX to me...
>
Yes, I have considerable first-hand experience with both DirectX and
OpenGL games. That's why I feel qualified to comment.
You just got done telling us that you don't know what the current
version of DirectX is. Hmmmm....
>>
>>
>>>and will immediately see the value in the bootable linux CD as a gaming
>>>platform.
>>
>>
>> Why? What API are all the game developers supposed to write for? You
>> surely don't expect them to go back to writing directly to the video
>> hardware?
>
><laugh> Not hardly. There are very few details available but the fact
>that the game is only available in a 64-bit version running on the
>Opteron and requires a platform called "Gamestorm" suggests that the
>"gamestorm" platform is somehow doing something different than the old
>32-bit OpenGl stuff.
Is that what it "suggests"? Really?
How about you tell us what exactly it is that is so different about
this platform? What new cross-platform API's does it introduce that
is going to make porting to OS/2 such a breeze?
It suggests that it's the 64bit version of Linux, which won't run on a
32bit processor. Pretty simple, really.
What it also suggests to me is that this company is trying to
capitalise on the recent launch of 64-bit consumer CPU's - in case
that had passed you by.
Can you also educate me about what makes "Gamestorm" different from
standard Linux? What's included in there beyond what's already
available? Or are you just eating up those marketing blurbs as if
they really mean something?
> It is the new 'gamestorm' 64-bit platform that
>apparently allows the bootable CD to provide a high-quality gaming
>experience without the need for the OpenGL and video drivers for
>specific hardware.
LOL - right. Gamestorm magically supports every bit of hardware out
there without the need for specific drivers. Nvidia and ATI can fire
their software people and never write another line of code - just
release the hardware and those utterly magical 64bit CPU's completely
eliminate the need for drivers. David T Johnson says so, so it must
be true right?
Do you read and think about what you write before you post? It sure
doesn't look like it.
This is just a version of Linux. Linux needs hardware specific
drivers. Unless these guys have written a driver-less OS from the
ground-up, then you are spouting nonsense.
>This would be required if the game was to work
>properly on the wide variety of video hardware that would be expected.
>I would speculate that it has something to do using direct access to the
>64-bit system memory for the video display in some way, as a guess. I
>am just guessing, though.
>
Yes, you are - guessing. And the very real problem of requiring
drivers to access the functionality of modern hardware in order to get
acceptable performance makes your guesses look wildly off base - if
not downright ridiculous.
>>
>> Also, what happens when new hardware comes out that isn't supported by
>> the version of linux on that CD? You've only got SATAII drives but
>> your now two year old game CD doesn't have drivers? Uh oh.
>
>They seem to be doing something with the 64-bit hardware that makes the
>'drivers' less critical. I will await further information, though.
>
ROTFLMAO.
You should try eating less sugar, and swallowing less marketing BS.
Are you *seriously* suggesting that the move to 64bit means nobody
needs drivers any more? *Muahahahahaha*!!!!!
Guess you should let Nvidia and ATI know that they can fire their
development teams who are working on their 64bit drivers.
>>
>>
>>>So how will this benefit OS/2? One of the chief obstacles of
>>>playing Windows Directx games on OS/2 has been lack of access to the
>>>directx software.
>>
>>
>> Well yeah, I guess not having directx on OS/2 would hamper your
>> ability to use directx on OS/2!
>
>yes.
>
I thoroughly agree. But aren't you going to state the obvious again
in case someone missed it the first time around?
>>
>>
>>>If Linux becomes a serious game platform replacement
>>>for Directx, it is likely that Linux games will be able to be either
>>>ported to OS/2
>>
>>
>> Oh my god, you *really* don't have clue, do you? It hasn't happened
>> with OpenGL games, what do you think has changed?
>
>the new 'gamestorm' 64-bit platform, whatever that is, which does not
>seem to be using OpenGL.
>
Is that what it "seems" to be doing?
Then again, given that AA is already released for Linux, a lot of what
"seems" like magic to you is really not all that terribly clever:
http://www.3dgamers.com/games/americasarmy/
Look for the word "linux". Sheesh.
>>
>> The whole point of DirectX is that it provides a standard set of API's
>> for developers to write towards so they can get on with the job of
>> creating *games* and adding their own custom effects instead of having
>> to worry about the basics of drawing a triangle and making it green.
>> The *hardware* developers can then concentrate on accelerating those
>> standard routines so that games run *fast*.
>
>perhaps the 64-bit memory addressing and hypertransport of the Opteron
>allows some sort of new design paradigm?
>
Perhaps sacrificing a chicken at midnight allows a CPU to magically
address a 3D accelerator board without the need for *ANY GRAPHICS
INSTRUCTIONS*.
Or perhaps you are just talking out of your ass?
Hey, are you going to address any of my points? Nah, didn't think so.
It's all about the API, stupid! You might as well suggest that the
Win32API or OS/2's API's are no longer required - it makes just as
much (aka zero) sense.
>>
>>
>>>or run on OS/2 with emulation software.
>>
>>
>> For this to happen you need to have a standard API supported across
>> both platforms - and the video card drivers need to accelerate it if
>> you don't want to play your games at one frame every four seconds.
>
>If the new 'gamestorm' 64-bit platform can boot independently and run on
>any Opteron hardware without the need for special drivers,
Gah - this is so stupid it makes my teeth hurt.
You cannot access the hardware without drivers.
> it may be
>possible for it to also boot it on OS/2 in some way similar to the way
>OS/2 can boot a 'DOS image' or 'boot from drive A' and such.
<WIPES TEARS FROM EYES>
So, you are going to boot a 64-bit OS from within your 32-bit OS/2,
and it's going to be able to use it's magic routines to directly
access all your hardware *without drivers* so that you get the fastest
games machine around!
Well fuck me, if that's not the most revolutionary piece of technology
to arrive in the last twenty years then I'm Britney Spears with knobs
on.
How come this hasn't been on Fox?
> I admit
>that I am just wildly speculating here, though. Since there seems to be
>no detailed information available, that is all there is to do.
>
ROTFL.
You *could* try applying simple logic or even reading up on what the
latest 64bit CPU's and OS/s are capable of, *before* you make stupid,
wild, and profoundly ignorant guesses that make you look ridiculous.
>>
>> The only graphics API shared by both Linux and OS/2 to my knowledge is
>> OpenGL. Last time I looked, OpenGL support on most cards for OS/2 was
>> pretty damn shabby where it wasn't non-existent. Linux is better, but
>> still not as good as Windows.
>
>There does not seem to be a requirement for the 'gamestorm' 64-bit
>platform to use OpenGL or else perhaps they have implemented it in some
>sort of virtual hardware implementation or something.
Which just happens to be able to take advantage of accelerated 3D
cards, all without any need for drivers.
Which sentence *precisely* leads you to this conclusion?
>Obviously, the
>'gamestorm' platform is doing something with 64-bit memory addressing
>that was not possible to do with 32-bit addressing.
>
The only obvious thing here is that you haven't the faintest clue what
you are prattling on about. Moving to 64-bit has increased the amount
of memory that can be accessed and has increased the bandwidth
available for calculations ( you can now work on 64 bits of data!).
What it *hasn't* done is eliminate the need for drivers or API's.
>>
>> The other problem with OpenGL of course, is that the current standard
>> is years old, and lacks a lot of the cool features that DirectX allows
>> (particularly shader support) - the next OpenGL standard is due soon
>> which will address this, but the chances of it ever seeing the light
>> of day for the OS/2 platform are somewhere between a snowballs chance
>> in hell and zero.
>>
>> No support means no ports.
>
>Perhaps OpenGL is morphing into something new and different or is being
>superseded by something new and better.
There's no "perhaps" about it. Try google - you might *learn*
something.
Try "OpenGL" and "2.0" for starters.
> In any event, DirectX seems to
>be disappering from the equation if it is possible to insert a CD-ROM
>into an Opteron machine and boot and run a high-performance game on any
>hardware.
>
Utter nonsense. As I've already pointed out, there are major
limitations with hardware support once you are reliant upon a fixed
point-in-time OS to run the game. How do you support future hardware
that doesn't exist yet? When new drivers come out that increase
performance or fix bugs how do you benefit?
This product has some appeal - it's a clever idea, and if the game is
optimised for linux you will end up with a very solid performant
platform. Just don't expect it to work on the machine you own five
years from now, and don't expect it to support every piece of hardware
out there. Also don't expect to be able to connect to the internet to
play multiplayer games without setting up all your configuration
details each and every time you boot. Hmmm....this is starting to
sound icky.
>>
>>
>>>Microsoft has
>>>pushed Directx as a replacement for DOS as a gaming platform but no one
>>>has ever loved (or even liked) directx
>>
>>
>> *cough* - really? And on what evidence do you base this assertion?
LOL - none. No response. You haven't changed at all, have you David?
>> DirectX and OpenGL have paved the way for modern gaming. What
>> developers used to hate was having to write the whole damn render
>> engine from scratch for DOS.
>>
No comment, David?
>>
>>>and now it looks as if Linux has
>>>skewered it right through its gizzard.
>>
>>
>> LOL. Sure - maybe on your planet.
>
>We'll see as more details become available.
I've seen. The difference is, *I* understood.
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>> [snip]
>>
>>
>>>If Linux becomes a serious game platform replacement
>>>for Directx, it is likely that Linux games will be able to be either
>>>ported to OS/2 or run on OS/2 with emulation software.
>>
>>
>> I forgot to mention: If you have the OS and game on one
>> self-contained CD, why on earth would anybody go to weeks or months of
>> efforts to port to a *different* OS such as OS/2? The whole point is
>> that you don't *need* a pre-installed OS - it's bundled with it's own!
>
>
>If the new 'gamestorm' 64-bit platform can boot independently and run on
>any Opteron hardware without the need for special drivers,
That's some "if" - <laugh>
> it may be
>possible for it to also boot it on OS/2 in some way similar to the way
>OS/2 can boot a 'DOS image' or 'boot from drive A' and such. I admit
>that I am just wildly speculating here, though. Since there seems to be
>no detailed information available, that is all there is to do.
a) you are repeating yourself.
b) it was nonsense the first time around.
OS/2 cannot provide direct access to the hardware for other 32-bit
OS's (think Windows, if you are struggling to keep up) - it is plain
*stupid* to think it can do so for a 64-bit OS that didn't even exist
at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
that it can act as the bootstrap loader? Would anyone bother even if
it could?
OS/2 has *NO ROLE* with such a self-contained bootable system. That
is it's entire raison d'etre - rendering the idea of the OS
irrelevant. It's a *very* strange choice to try and pin your hopes
for OS/2 upon!
Jason Bowen
link in response to you.
http://slashdot.org/comments.pl?sid=20503&cid=2194363
John Carmack doesn't agree with DTJ but who is John Carmack and ID
software... pah ;-)
A point he makes: D3D used to suck and Microsoft has worked
hard at improving it.
David Sutherland
Bowen) wrote:
>For whatever reason, DTJ's post hasn't shown up here
You are *so* not missing much :)
>so I'll post this
>link in response to you.
>
>http://slashdot.org/comments.pl?sid=20503&cid=2194363
>
>John Carmack doesn't agree with DTJ but who is John Carmack and ID
>software... pah ;-)
>
Who he?
>A point he makes: D3D used to suck and Microsoft has worked
>hard at improving it.
It's worth noting that this article is old (last year maybe?) and that
the programmability he wishes for is now available as part of the DX9
spec - dutifully accelerated by hardware (ATI) and drivers (NV) ;)
For anyone who has been living under a rock and doesn't know what the
state of play is for PC gaming:
http://www.beyondhalflife.com/gallery/HL2
(DX9)
http://www.gamershell.com/game/FPS_Doom_3.shtml
(OpenGL)
Two good API's. Neither of which are going away any time soon!
Stilz
he did use most of his modifiers (i.e. "I would speculate",
"apparently", "seems to be", ...) :-)))))
Tried a Knoppix bootable CD?
Mike
David Sutherland wrote:
>
> On Wed, 01 Oct 2003 12:55:00 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
<snipped>
Jack Troughton
>On 1 Oct 2003 22:18:00 GMT, bow...@rintintin.colorado.edu (Jason
>Bowen) wrote:
>
>>For whatever reason, DTJ's post hasn't shown up here
>
>You are *so* not missing much :)
>
>>so I'll post this
>>link in response to you.
>>
>>http://slashdot.org/comments.pl?sid=20503&cid=2194363
>>
>>John Carmack doesn't agree with DTJ but who is John Carmack and ID
>>software... pah ;-)
>>
>
>Who he?
Oh you know... that guy.
>>A point he makes: D3D used to suck and Microsoft has worked
>>hard at improving it.
>
>It's worth noting that this article is old (last year maybe?) and that
>the programmability he wishes for is now available as part of the DX9
>spec - dutifully accelerated by hardware (ATI) and drivers (NV) ;)
>
>For anyone who has been living under a rock and doesn't know what the
>state of play is for PC gaming:
>
>http://www.beyondhalflife.com/gallery/HL2
>(DX9)
>
>http://www.gamershell.com/game/FPS_Doom_3.shtml
>(OpenGL)
I am sooo looking forward to when these come out for real. The only
problem is I don't have enough box to run them; I'm going to have to
see what I can do about that.
I'm gonna have to wait a bit until I can get a system that can
handle them... though of course warp will run like hell on the
hardware that can support them.
>Two good API's. Neither of which are going away any time soon!
Indeed. I just wish the OpenGL dev license wasn't so expensive; if
it was cheaper someone might be able to do something about bringing
it over.
--
-------------------------------------------------------------------
* Jack Troughton jake at consultron.ca *
* http://consultron.ca irc.ecomstation.ca *
* Kingston Ontario Canada news://news.consultron.ca *
-------------------------------------------------------------------
David Sutherland
>Top Post - Oh come on D.S., DTJ doesn't really understand this stuff and
>he did use most of his modifiers (i.e. "I would speculate",
>"apparently", "seems to be", ...) :-)))))
>
LOL - how very right you are. I cracked up laughing at some of DTJ's
evidence for leaping to his "death blow" claims:
"somehow doing something different"
"in some way, as a guess. I am just guessing, though."
"They seem to be doing something"
"which does not seem to be"
"some sort of virtual hardware implementation or something"
Wow, all very compelling stuff!. DTJ obviously has a firm grasp on
the topic - well, he's got a firm grasp on *something* :)
>Tried a Knoppix bootable CD?
>
I know of it, but I haven't used it. Looks nice, though. DTJ will
claim that it can't exist because gamestorm is the first such platform
ever ;)
MMI
> If the new 'gamestorm' 64-bit platform can boot independently and run on
> any Opteron hardware without the need for special drivers, it may be
> possible for it to also boot it on OS/2 in some way similar to the way
> OS/2 can boot a 'DOS image' or 'boot from drive A' and such. I admit
> that I am just wildly speculating here, though. Since there seems to be
> no detailed information available, that is all there is to do.
It is impossible to run protected mode OS under other protected mode
OS without full virtualization of the x86 CPU or some heavy metal
hacks. Problems of DOS virtualization are almost nonsignificant
compared to protected mode virtualization. Don't even think about
booting such CD in OS/2.
I hate to admit it, but I prefer reality. :-)
Cheers,
Martin
Menno Willemse
> On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
>
>
>>A company named SCI has developed a bootable CD that contains a game and
>>a Linux OS that can be booted on any Opteron PC to turn it into a
>>powerful gaming console.
>>
>>This is a BIG deal.
>
David Sutherland wrote:
> Is it? Bootable linux distributions have been around for a while.
> Bundling a software title with it isn't particularly big news.
What is new about it is that you no longer need Windows to run your
game. Apparently the "Gamestorm" rendering engine developed by SCI is
good enough to develop a game like "America's Army".
> How interesting - you're not sure what version of directx is current,
> but you know it's a "miserable kludge"? Sounds like
> wishful-thinking wrapped around ignorance to me.
Never had a game crash on you because another game insisted on upgrading
DirectX? That's the whole problem with Windows game development. You are
dependent on a system that can change under your fingers at any moment.
If you make the game bootable, you have full control of the system.
Nobody is going to change the OS on you. If you do it under Linux, this
will also remove your dependency on Microsoft. Linux you can stick on
your CD for nothing.
> Why? What API are all the game developers supposed to write for? You
> surely don't expect them to go back to writing directly to the video
> hardware?
"Gamestorm" apparently. Developed by SCI.
> Also, what happens when new hardware comes out that isn't supported by
> the version of linux on that CD? You've only got SATAII drives but
> your now two year old game CD doesn't have drivers? Uh oh.
CD writers are your friend. You could download a fresh ISO image and
burn it. Or you could arrange to park fresh drivers on your harddisk
somewhere. I downloaded Knoppix and am running it under VPC now. eCS
Electronic Delivery is under way. I think that shrink-wrapped software
is in its last days.
Illegal copying wouldn't be a problem. All of these games will probably
be tied to your gaming account that lives on the Internet. I wouldn't be
surprised if you could get the game for free, and play it offline. Just
to give you a taste of what lies on the Internet. And then you'll need a
login on the game server, which is what you pay for.
>>Microsoft has
>>pushed Directx as a replacement for DOS as a gaming platform but no one
>>has ever loved (or even liked) directx
>
> *cough* - really? And on what evidence do you base this assertion?
> DirectX and OpenGL have paved the way for modern gaming. What
> developers used to hate was having to write the whole damn render
> engine from scratch for DOS.
Well, now MS has pissed them off to such an extent that they are going
back to developing their own rendering engines. That should tell you
something about how much fun it is to develop for DirectX.
I think this is a good trend. Traditionally, one of the great selling
points for Windows has been that you needed it for games (like "how many
tries to get this printed out correctly"). Now, you no longer need
Windows to play games. Which will make people think about what you need
Windows for at all.
Cheers/2,
Menno
Menno Willemse
David Sutherland wrote:
> Your whole "this is good news for OS/2" song and dance routine is just
> silly. OS/2 is rendered utterly irrelevant by such a product.
As is Windows. We're working on the assumption here that everything
that's bad news for MS is good news for the rest of the industry.
The selling point "Yeah but you need It to play games" is falling away.
Conversely, the OS/2 disadvantage "But there ain't no games for it" is
also falling away because you can now play games despite the fact that
you have OS/2 on your PC and not Windows. Hence, this development is
good for us and bad for Microsoft. And about time too!
It's also good for Windows users, because it is another lever that
breaks MS' stranglehold on your PC. The less stuff you actually *need*
Windows for, the better.
Cheers/2,
Menno.
Jazz
> Tried a Knoppix bootable CD?
That's nothing. Go here:
http://morphix.sourceforge.net/modules/news/
--
Jazz.
"Curculionem Darl McBride minusculum habet."
(Cato - Pro Lino; Apocrypha)
Irv Spalten
David Sutherland wrote:
I had to check my calendar, thought it might have been April first?
Then I remembered, your typing to a non-programmer. Don't expect him to
have any great insight in how programs work or what needs to be done to
create ANY program. He'd hardly be expected to know how to move a sprite
across a screen than how to put the letter 'A' up on the screen in a set
location, much less than how to even get a single byte off a hard disk
or CD.
Irv
Irv Spalten
Menno Willemse wrote:
> Hello World,
>
> > On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
> > <djoh...@isomedia.com> wrote:
> >
> >
> >>A company named SCI has developed a bootable CD that contains a game and
> >>a Linux OS that can be booted on any Opteron PC to turn it into a
> >>powerful gaming console.
> >>
> >>This is a BIG deal.
> >
>
> David Sutherland wrote:
> > Is it? Bootable linux distributions have been around for a while.
> > Bundling a software title with it isn't particularly big news.
>
> What is new about it is that you no longer need Windows to run your
> game. Apparently the "Gamestorm" rendering engine developed by SCI is
> good enough to develop a game like "America's Army".
>
> > How interesting - you're not sure what version of directx is current,
> > but you know it's a "miserable kludge"? Sounds like
> > wishful-thinking wrapped around ignorance to me.
>
> Never had a game crash on you because another game insisted on upgrading
> DirectX? That's the whole problem with Windows game development. You are
> dependent on a system that can change under your fingers at any moment.
This was true in earlier versions of DirectX. The problem is/was that
the game you bought came bundled with a specific version of DirectX, but
the install NEVER checked which version you had installed of DirectX. It
either automatically installed DirectX packaged with it or ASKED you if
you wanted to install DirectX, impying it was neeeded.
BTW, both OS/2 and WinXX have/had a similar problem with DLL's. Early
on, MS shipped DLL source with the SDK and end-user implementers were to
add API calls to the DLL's and ship the DLL's. Works great, until some
other program provider ships the SAME DLL without YOUR calls. Where did
your DLL go, into the system DLL directory or main system directory in
the path.
>
> If you make the game bootable, you have full control of the system.
> Nobody is going to change the OS on you. If you do it under Linux, this
> will also remove your dependency on Microsoft. Linux you can stick on
> your CD for nothing.
>
> > Why? What API are all the game developers supposed to write for? You
> > surely don't expect them to go back to writing directly to the video
> > hardware?
>
> "Gamestorm" apparently. Developed by SCI.
If they can get enough game developers to write to the API set they
create, and the big if, will it be complete enough.
> > Also, what happens when new hardware comes out that isn't supported by
> > the version of linux on that CD? You've only got SATAII drives but
> > your now two year old game CD doesn't have drivers? Uh oh.
>
> CD writers are your friend. You could download a fresh ISO image and
> burn it. Or you could arrange to park fresh drivers on your harddisk
> somewhere. I downloaded Knoppix and am running it under VPC now. eCS
> Electronic Delivery is under way. I think that shrink-wrapped software
> is in its last days.
I think the device being new isn't the problem but accessing it? Drivers
alone may not be enough, you may need API support to properly use the
new h/w.
> Illegal copying wouldn't be a problem. All of these games will probably
> be tied to your gaming account that lives on the Internet. I wouldn't be
> surprised if you could get the game for free, and play it offline. Just
> to give you a taste of what lies on the Internet. And then you'll need a
> login on the game server, which is what you pay for.
>
> >>Microsoft has
> >>pushed Directx as a replacement for DOS as a gaming platform but no one
> >>has ever loved (or even liked) directx
> >
> > *cough* - really? And on what evidence do you base this assertion?
> > DirectX and OpenGL have paved the way for modern gaming. What
> > developers used to hate was having to write the whole damn render
> > engine from scratch for DOS.
>
> Well, now MS has pissed them off to such an extent that they are going
> back to developing their own rendering engines. That should tell you
> something about how much fun it is to develop for DirectX.
>
> I think this is a good trend. Traditionally, one of the great selling
> points for Windows has been that you needed it for games (like "how many
> tries to get this printed out correctly"). Now, you no longer need
> Windows to play games. Which will make people think about what you need
> Windows for at all.
Many of the games today ship in many versions, especially the 'arcade'
type. Many are for the consoles and you don't need windows for them.
Nothing has changed, other than, at best, another platform has arrived
to write too.
If there are COMPLETE tools, and it is relatively cheap, in both
purchase and usage costs, then you might have a true new environment.
However, IBM had porting tools too for OS/2 from Windows. There was a
Family Game Pack it produced of a few programs. SimCity was back level
for instance. Whatever it took to create the OS/2 version wasn't that
easy to do, otherwise other games would have been ported. It never went
anywhere basically.
The main problem is having the game writer write FIRST to your platform,
whatever it is. Stardock tried that, OS/2 only, and essentially failed
to garner enough support or interest to continue, at least compared to
Windows game producers.
What makes this a COMPELING platform, especially if the game tools may
not exist to compete with Windows?
Irv
> Cheers/2,
> Menno
>
Menno Willemse
Irv Spalten wrote:
>
> If they can get enough game developers to write to the API set they
> create, and the big if, will it be complete enough.
Actually, no. Whether anyone else would use the API would be besides the
point. SCI can use it regardless to produce games that are independent
of the underlying OS. SCI has the assets. They can use it to produce CD
bootable games. As long as those games look good, SCI can compete. They
could, of course, sell the license to others.
> I think the device being new isn't the problem but accessing it? Drivers
> alone may not be enough, you may need API support to properly use the
> new h/w.
Which they would then have to produce and add to the ISO image. CDs may
go old, but you can always burn new ones that have been updated for the
latest hardware.
> The main problem is having the game writer write FIRST to your platform,
> whatever it is. Stardock tried that, OS/2 only, and essentially failed
> to garner enough support or interest to continue, at least compared to
> Windows game producers.
>
> What makes this a COMPELING platform, especially if the game tools may
> not exist to compete with Windows?
Well, the compelling reason to do it this way is that you can use normal
PCs, which everybody already has in the house. (Well when Opterons
become common, that is). The second compelling reason is that you don't
have to deal with the MS platform difficulties. You take your own
platform with you. No royalties, no spurious changes.
Of course, you need to keep up to date with the current display and
sound hardware, but the Linux community has a way of doing that.
Especially if you inject the technology you invent back into the
community. Nonetheless, keeping up to date with new hardware is going to
be the biggest challenge for this setup.
cheers/2,
Menno
Irv Spalten
Menno Willemse wrote:
> Hello World,
>
> Irv Spalten wrote:
> >
> > If they can get enough game developers to write to the API set they
> > create, and the big if, will it be complete enough.
>
> Actually, no. Whether anyone else would use the API would be besides the
> point. SCI can use it regardless to produce games that are independent
> of the underlying OS. SCI has the assets.
Yes, but.... today, a lot of games live on different platforms and
consoles. That is where the money is, universal usage. Single platform
games don't do as well as multi-platform.
The 'real' question should be 'will' the other game producers abandon
the DirectX support and shift. Why would they if they already can
produce the games that work on Win machines and now create a bootable CD
that would work on the same machine?
Also, take into account the many libraries, tools, and routines these
existing companies have for creating games on the existing platforms.
Can these still be used or will time and money need to be spent to
create the same for the new guy?
> They can use it to produce CD
> bootable games. As long as those games look good, SCI can compete. They
> could, of course, sell the license to others.
Licensing doesn't usually sell well, unless CHEAP.
However, you are talking about a single company, not an industry. When
the game industry embraces it, then you have something. Until then, it
is just another 'attempt'.
>
> > I think the device being new isn't the problem but accessing it? Drivers
> > alone may not be enough, you may need API support to properly use the
> > new h/w.
>
> Which they would then have to produce and add to the ISO image.
You make it sound simple, easy, and cheap. The question 'why' would
they? Yes, roll it out in a NEW release that they can sell, makes sense.
Go back and update a 2 year old game with hardly any new sales, doesn't
make sense, unless it is just a DD, not a programming change.
> CDs may
> go old, but you can always burn new ones that have been updated for the
> latest hardware.
>
> > The main problem is having the game writer write FIRST to your platform,
> > whatever it is. Stardock tried that, OS/2 only, and essentially failed
> > to garner enough support or interest to continue, at least compared to
> > Windows game producers.
> >
> > What makes this a COMPELING platform, especially if the game tools may
> > not exist to compete with Windows?
>
> Well, the compelling reason to do it this way is that you can use normal
> PCs, which everybody already has in the house. (Well when Opterons
> become common, that is).
Same reasoning holds for any other platform or underlying OS.
> The second compelling reason is that you don't
> have to deal with the MS platform difficulties. You take your own
> platform with you. No royalties, no spurious changes.
>
Hmm, any platform will give you problems? One 'feature' of DirectX or
OpenGL is the non-need to do the hardware things, thus eliminating both
serious coding work and making the program see the h/w as a transparent
item, no need to enumerate or write code for specific h/w.
Royalties? Like what? DirectX doesn't have royalties, does it?
Spurious changes, I assume you mean DirectX version changes causing
incompatibilies?
I've got a bookcase of at least 20 to 30 games. Each has been loaded at
one point onto a computer and the CD's put away (except for those
blasted ones, like MS FS 2004, that require a CD to play [there is a
crack to make this a non-issue]). Seems more handy to me than needed a
different CD each time I want to play a game.
I don't 'do' consoles, but they require the game usually to be inserted
before you can play it, no? I guess this will turn a PC into a game
console?
I do occasionally RUN something in the background and play a game. Handy
feature that is usefull. I like it that way too.
> Of course, you need to keep up to date with the current display and
> sound hardware, but the Linux community has a way of doing that.
> Especially if you inject the technology you invent back into the
> community. Nonetheless, keeping up to date with new hardware is going to
> be the biggest challenge for this setup.
An OS is far easier to keep current than a completed game with little
financial returns for updating I'd say.
>
> cheers/2,
> Menno
>
Irv
David T. Johnson
David T. Johnson
I haven't seen another game that came with its own OS on a bootable CD
although there might one somewhere. DOS games came close to that when
you would boot DOS from a floppy and run the game. The problem with
doing the bootable CD/OS game on newer systems, though, is support for
the hardware. The OS on the CD has to boot and recognize the hardware,
provide the proper drivers and software, configure itself, and then run
the game and it has to do that for the wide variety of hardware that is
out there. Apparently that has been sufficiently difficult to do that no
one has wanted to do that.
>
>
>>>
>>>>Microsoft has been pushing their 'directx'
>>>>technology as a way to get a high-quality gaming experience while still
>>>>running Windows and is up to 'Directx 9' I think. Those of you who have
>>>>struggled to get cranky directx Windows games to actually run like they
>>>>are supposed to know what a miserable kludge the who directx platform is
>>>
>>>
>>>How interesting - you're not sure what version of directx is current,
>>>but you know it's a "miserable kludge"? Sounds like
>>>wishful-thinking wrapped around ignorance to me.
>>
>>Sounds like first-hand experience with DirectX to me...
>>
>
>
> Yes, I have considerable first-hand experience with both DirectX and
> OpenGL games. That's why I feel qualified to comment.
>
> You just got done telling us that you don't know what the current
> version of DirectX is. Hmmmm....
I don't care what the current version of DirectX is, for sure. <laugh>
>
>
>
>>>
>>>>and will immediately see the value in the bootable linux CD as a gaming
>>>>platform.
>>>
>>>
>>>Why? What API are all the game developers supposed to write for? You
>>>surely don't expect them to go back to writing directly to the video
>>>hardware?
>>
>><laugh> Not hardly. There are very few details available but the fact
>>that the game is only available in a 64-bit version running on the
>>Opteron and requires a platform called "Gamestorm" suggests that the
>>"gamestorm" platform is somehow doing something different than the old
>>32-bit OpenGl stuff.
>
>
> Is that what it "suggests"? Really?
>
> How about you tell us what exactly it is that is so different about
> this platform? What new cross-platform API's does it introduce that
> is going to make porting to OS/2 such a breeze?
Hard to say, but they certainly cannot be less accessible than DirectX.
<laugh>
>
> It suggests that it's the 64bit version of Linux, which won't run on a
> 32bit processor. Pretty simple, really.
Yes, it is obviously running 64-bit Linux.
>
> What it also suggests to me is that this company is trying to
> capitalise on the recent launch of 64-bit consumer CPU's - in case
> that had passed you by.
>
> Can you also educate me about what makes "Gamestorm" different from
> standard Linux? What's included in there beyond what's already
> available? Or are you just eating up those marketing blurbs as if
> they really mean something?
No details about "Gamestorm" are available yet although a lot of people
are interested. It seems to have the capability of providing a
high-quality 3D game on Linux without using either DirectX or OpenGL and
appears to be able to run on any hardware. See below.
>
>
>>It is the new 'gamestorm' 64-bit platform that
>>apparently allows the bootable CD to provide a high-quality gaming
>>experience without the need for the OpenGL and video drivers for
>>specific hardware.
>
>
> LOL - right. Gamestorm magically supports every bit of hardware out
> there without the need for specific drivers. Nvidia and ATI can fire
> their software people and never write another line of code - just
> release the hardware and those utterly magical 64bit CPU's completely
> eliminate the need for drivers. David T Johnson says so, so it must
> be true right?
When I read the Gamestorm story, my thought was that they are using the
64-bit Opteron CPU to do stuff that the video hardware would otherwise
typically do on 32-bit systems. If so, the system would essentially run
on any video hardware.
>
> Do you read and think about what you write before you post? It sure
> doesn't look like it.
>
> This is just a version of Linux. Linux needs hardware specific
> drivers. Unless these guys have written a driver-less OS from the
> ground-up, then you are spouting nonsense.
Yes, video drivers are needed to run America's Army on 32-bit Linux. If
someone developed a bootable CD with America's Army and 32-bit Linux, it
would need to scan the hardware and find the appropriate driver for the
video hardware that it detected. If the video hardware was newer than
the CD, the driver might not be present. If the game did not have a
driver for the video hardware that it detected, it would not be able to
run. Etc. These severe restrictions have prevented the idea of a
bootable CD with the app and OS on it from catching fire. None of this
seems to be the case, however, with the 64-bit version of America's Army
running on the mysterious "Gamestorm" platform. The only hardware
requirement that it lists is that a 64-bit Opteron be used. The
announcement does mention that it scans the hardware, though, but it
also mentions that it boots in one minute so it does not seem to be
doing a lot with the scan.
>
>
>>This would be required if the game was to work
>>properly on the wide variety of video hardware that would be expected.
>>I would speculate that it has something to do using direct access to the
>>64-bit system memory for the video display in some way, as a guess. I
>>am just guessing, though.
>>
>
>
> Yes, you are - guessing. And the very real problem of requiring
> drivers to access the functionality of modern hardware in order to get
> acceptable performance makes your guesses look wildly off base - if
> not downright ridiculous.
I agree that this is true with typical 32-bit systems. The
Gamestorm/Opteron platform, however, has apparently found a way to get
acceptable performance without using hardware-specific drivers.
>
>
>>>Also, what happens when new hardware comes out that isn't supported by
>>>the version of linux on that CD? You've only got SATAII drives but
>>>your now two year old game CD doesn't have drivers? Uh oh.
>>
>>They seem to be doing something with the 64-bit hardware that makes the
>>'drivers' less critical. I will await further information, though.
>>
>
>
> ROTFLMAO.
>
> You should try eating less sugar, and swallowing less marketing BS.
>
> Are you *seriously* suggesting that the move to 64bit means nobody
> needs drivers any more? *Muahahahahaha*!!!!!
>
> Guess you should let Nvidia and ATI know that they can fire their
> development teams who are working on their 64bit drivers.
Software to control hardware (i.e. drivers) will always be needed.
However, this software does not seem to be hardware-specific on the
Gamestorm/Opteron platform. As I mentioned, though, the details are not
yet available.
>
>
>>>
>>>>So how will this benefit OS/2? One of the chief obstacles of
>>>>playing Windows Directx games on OS/2 has been lack of access to the
>>>>directx software.
>>>
>>>
>>>Well yeah, I guess not having directx on OS/2 would hamper your
>>>ability to use directx on OS/2!
>>
>>yes.
>>
>
>
> I thoroughly agree. But aren't you going to state the obvious again
> in case someone missed it the first time around?
>
>
>>>
>>>>If Linux becomes a serious game platform replacement
>>>>for Directx, it is likely that Linux games will be able to be either
>>>>ported to OS/2
>>>
>>>
>>>Oh my god, you *really* don't have clue, do you? It hasn't happened
>>>with OpenGL games, what do you think has changed?
>>
>>the new 'gamestorm' 64-bit platform, whatever that is, which does not
>>seem to be using OpenGL.
>>
>
>
> Is that what it "seems" to be doing?
>
> Then again, given that AA is already released for Linux, a lot of what
> "seems" like magic to you is really not all that terribly clever:
>
> http://www.3dgamers.com/games/americasarmy/
>
> Look for the word "linux". Sheesh.
yes, there is a 32-bit version of America's Army for Linux.
>
>
>
>>>The whole point of DirectX is that it provides a standard set of API's
>>>for developers to write towards so they can get on with the job of
>>>creating *games* and adding their own custom effects instead of having
>>>to worry about the basics of drawing a triangle and making it green.
>>>The *hardware* developers can then concentrate on accelerating those
>>>standard routines so that games run *fast*.
>>
>>perhaps the 64-bit memory addressing and hypertransport of the Opteron
>>allows some sort of new design paradigm?
>>
>
>
> Perhaps sacrificing a chicken at midnight allows a CPU to magically
> address a 3D accelerator board without the need for *ANY GRAPHICS
> INSTRUCTIONS*.
My speculation is that the Opteron CPU and the Gamestorm platform are
somehow doing stuff in the CPU that was formerly done by the
'accelerator' board. The Opteron CPU has a lot of new hardware function
that 32-bit CPUs do not. For example, it has more registers, many new
x86-64 instructions, the ability to address more memory, etc.
>
> Or perhaps you are just talking out of your ass?
>
> Hey, are you going to address any of my points? Nah, didn't think so.
> It's all about the API, stupid! You might as well suggest that the
> Win32API or OS/2's API's are no longer required - it makes just as
> much (aka zero) sense.
>
>
>>>
>>>>or run on OS/2 with emulation software.
>>>
>>>
>>>For this to happen you need to have a standard API supported across
>>>both platforms - and the video card drivers need to accelerate it if
>>>you don't want to play your games at one frame every four seconds.
>>
>>If the new 'gamestorm' 64-bit platform can boot independently and run on
>>any Opteron hardware without the need for special drivers,
>
>
> Gah - this is so stupid it makes my teeth hurt.
>
> You cannot access the hardware without drivers.
Agreed. However, the Opteron/Gamestorm platform does not seem to
require separate drivers for each different video card. At least, they
have not listed any particular video hardware requirements in their
announcement which one would expect they would otherwise do.
>
>
>>it may be
>>possible for it to also boot it on OS/2 in some way similar to the way
>>OS/2 can boot a 'DOS image' or 'boot from drive A' and such.
>
>
> <WIPES TEARS FROM EYES>
>
> So, you are going to boot a 64-bit OS from within your 32-bit OS/2,
> and it's going to be able to use it's magic routines to directly
> access all your hardware *without drivers* so that you get the fastest
> games machine around!
>
There has been no claim by SCI that their machine is "fastest" although
it was termed a 'console machine on steroids'.
>
>>I admit
>>that I am just wildly speculating here, though. Since there seems to be
>>no detailed information available, that is all there is to do.
>>
>
>
> ROTFL.
>
> You *could* try applying simple logic or even reading up on what the
> latest 64bit CPU's and OS/s are capable of, *before* you make stupid,
> wild, and profoundly ignorant guesses that make you look ridiculous.
>
>
>>>The only graphics API shared by both Linux and OS/2 to my knowledge is
>>>OpenGL. Last time I looked, OpenGL support on most cards for OS/2 was
>>>pretty damn shabby where it wasn't non-existent. Linux is better, but
>>>still not as good as Windows.
>>
>>There does not seem to be a requirement for the 'gamestorm' 64-bit
>>platform to use OpenGL or else perhaps they have implemented it in some
>>sort of virtual hardware implementation or something.
>
>
> Which just happens to be able to take advantage of accelerated 3D
> cards, all without any need for drivers.
>
> Which sentence *precisely* leads you to this conclusion?
They don't seem to be using the 'accelerators' in the video hardware
because they don't list any particular hardware requirements and they
suggest that it will boot on any Opteron computer: "They can run the
game on a floor model at Best Buy, and leave the hard drive untouched,"
he says."
Other claims by SCI:
"cuts out legacy operating system overhead..."
"without the usual overhead from the legacy operating systems
raditionally used in the gaming industry, SCI claims."
"With the AMD Athlon 64 processor and GameStorm technology, AMD is able
to showcase a fully-integrated 64-bit environment that delivers
performance and realism to the most demanding gamers,"
"the company is now working on porting several more 32-bit and 64-bit
games to the GameStorm platform"
The last sentence is revealing because it suggests that 'porting' is
required to run other games on the Gamestorm platform which suggests
that Gamestorm does not run OpenGL games. Also, there are no 64-bit
x86-64 OpenGL libraries that I have heard of yet (although I think they
are in development) and the game also certainly does not use DirectX
<laugh>.
>
>
>>Obviously, the
>>'gamestorm' platform is doing something with 64-bit memory addressing
>>that was not possible to do with 32-bit addressing.
>>
>
>
> The only obvious thing here is that you haven't the faintest clue what
> you are prattling on about. Moving to 64-bit has increased the amount
> of memory that can be accessed and has increased the bandwidth
> available for calculations ( you can now work on 64 bits of data!).
> What it *hasn't* done is eliminate the need for drivers or API's.
Software is needed to operate hardware but it does not have to
necessarily be specific to every piece of hardware. For example, the
BIOS on my machine can display stuff on any video card that I insert.
It can also operate the keyboard, floppy disk, hard disk, CD-ROM, ZIP
drive, and USB devices without any additional drivers and before the
"operating system" ever even boots. It seems possible that SCI is doing
more with the CPU and less with the video hardware than is done on
32-bit systems.
>
>
>>>The other problem with OpenGL of course, is that the current standard
>>>is years old, and lacks a lot of the cool features that DirectX allows
>>>(particularly shader support) - the next OpenGL standard is due soon
>>>which will address this, but the chances of it ever seeing the light
>>>of day for the OS/2 platform are somewhere between a snowballs chance
>>>in hell and zero.
>>>
>>>No support means no ports.
>>
>>Perhaps OpenGL is morphing into something new and different or is being
>>superseded by something new and better.
>
>
> There's no "perhaps" about it. Try google - you might *learn*
> something.
>
> Try "OpenGL" and "2.0" for starters.
The Gamestorm platform is apparently not using either OpenGL or DirectX.
See above.
>
>
>>In any event, DirectX seems to
>>be disappering from the equation if it is possible to insert a CD-ROM
>>into an Opteron machine and boot and run a high-performance game on any
>>hardware.
>>
>
>
> Utter nonsense.
I am certain that DirectX is not being used by the Gamestorm platform
which means that DirectX is disappearing from the 'picture.' <laugh>
> As I've already pointed out, there are major
> limitations with hardware support once you are reliant upon a fixed
> point-in-time OS to run the game. How do you support future hardware
> that doesn't exist yet? When new drivers come out that increase
> performance or fix bugs how do you benefit?
Indeed. See above. These are significant limitations to the entire
idea of the bootable CD-with-OS that SCI may have overcome by using the
64-bit Opteron.
>
> This product has some appeal - it's a clever idea, and if the game is
> optimised for linux you will end up with a very solid performant
> platform. Just don't expect it to work on the machine you own five
> years from now, and don't expect it to support every piece of hardware
> out there.
We will have to wait to see what SCI says in upcoming weeks.
> Also don't expect to be able to connect to the internet to
> play multiplayer games without setting up all your configuration
> details each and every time you boot. Hmmm....this is starting to
> sound icky.
"It feels like a gaming console on steroids and even allows for online
access so you can connect to online game servers for multi-player
action," said Jesper Jensen, CEO of Super Computer, Inc."
It will be interesting to see how the Gamestorm platform deals with this
issue.
>
>
>>>
>>>>Microsoft has
>>>>pushed Directx as a replacement for DOS as a gaming platform but no one
>>>>has ever loved (or even liked) directx
>>>
>>>
>>>*cough* - really? And on what evidence do you base this assertion?
>
>
> LOL - none. No response. You haven't changed at all, have you David?
>
>
>>>DirectX and OpenGL have paved the way for modern gaming. What
>>>developers used to hate was having to write the whole damn render
>>>engine from scratch for DOS.
>>>
>
>
> No comment, David?
>
>
>>>>and now it looks as if Linux has
>>>>skewered it right through its gizzard.
>>>
>>>
>>>LOL. Sure - maybe on your planet.
>>
>>We'll see as more details become available.
>
>
> I've seen. The difference is, *I* understood.
>
David T. Johnson
David T. Johnson
Irv Spalten
David T. Johnson wrote:
> Irv Spalten wrote:
> > Menno,
> >
> > Menno Willemse wrote:
> >
> >
> >>Hello World,
> >>
> >>Irv Spalten wrote:
> >>
> >>>If they can get enough game developers to write to the API set they
> >>>create, and the big if, will it be complete enough.
> >>
> >>Actually, no. Whether anyone else would use the API would be besides the
> >>point. SCI can use it regardless to produce games that are independent
> >>of the underlying OS. SCI has the assets.
> >
> >
> > Yes, but.... today, a lot of games live on different platforms and
> > consoles. That is where the money is, universal usage. Single platform
> > games don't do as well as multi-platform.
>
> If the game boots from its own CD and provides its own OS on that CD to
> run, the 'platform' is irrelevant.
You didn't UNDERSTAND what I said!
> Even you should be able to recognize
> that reality.
Did you UNDERSTAND what I said, obviously NOT!!!
>The buyer doesn't care what the platform is along as the
> game runs properly on his system.
I'm *NOT* talking about the BUYER, but the PRODUCER here!
Let me say it again and BREAK it down!
> > Yes, but.... today, a lot of games live on different platforms and
> > consoles.
I assume you didn't disagree with this above? What it means is that the
same game, take Madden 200x Football or Tiger Woods Golf for instance.
Essentially the SAME game on mutiple platforms, all released at the same
time.
One can deduce from this that the developer either has individual teams
(including all phases, from design to packaging) doing the work on a
tight singular schedule, or there are TOOLS to allow porting quickly,
efficiently, and error free, to produce more than one platform version
from a single set of source code.
> >That is where the money is, universal usage.
This means what I just said. One set of source code and multiple
platform versions from that one set. Write once, sell many sort of deal.
Not write individually and sell individually. Too expensive and time
consuming to do so.
> > Single platform
> > games don't do as well as multi-platform.
Here this means a SINGLE game for a single platform from a single set of
source code.
Now as an END-USER, you could be correct. However, the proof is in the
game. Eventually though, I'd venture to guess most gamer'ers would be
fed up with a stack of CD's to have to play some games, continually
re-booting to play another. PC Gamer's aren't used to that mode of play.
Matter of fact, most use VIRTUAL CD's just so the CD's are not required
to play or 'hack' the executables to forgo the CD check.
Irv
Jack Troughton
>David Sutherland wrote:
>
>>
>> OS/2 cannot provide direct access to the hardware for other 32-bit
>> OS's (think Windows, if you are struggling to keep up) - it is plain
>> *stupid* to think it can do so for a 64-bit OS that didn't even exist
>> at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
>> that it can act as the bootstrap loader?
>
>I am writing this using OS/2 running on an Opteron system. <laugh>
Can you start a 64 bit system in a VM under OS/2 on this system?
This is the essential point in question.
The answer is no.
You could make something like loadlin that would load such a system
from OS/2, but it would be tantamount to shooting OS/2 in the head
and replacing it with the 64 bit system code.
What the opteron wants is a 64 bit system that can create 32 bit
VMs. From what I've read, the opteron can create complete
fully-virtualised 32 bit machines when running in 64 bits; you could
use this to start a 32 bit system in a VM. However, such a system
would only be able to get virtual access to hardware like video
cards and the like. This would result in a performance penalty for
certain types of operations, and games that stress the performance
of the hardware (such as a 3d video card) would suffer as a result.
It's the complete virtualisation that leads to the possibility that
AMD could come out on top of Intel with this chip; being able to run
multiple VMs virtualised in hardware means that one can do something
like virtual PC with performance on par.
In fact, IIUC, when you're running a 32 bit system on Opteron, the
chip itself is simply running one VM on top of its 64 bit system,
and giving it exclusive access to devices.
David T. Johnson
Marty
> Jack Troughton wrote:
>
>> The answer is no.
>
> I don't see 64-bit code running as a 'VM' under OS/2 but rather as
> 64-bit code.
Boy, those magical pixies have been busy, haven't they? What do you
suppose is going to happen when OS/2 needs to perform a context switch?
It's really getting painful to even read these half-baked guesses.
Jack Troughton
>Jack Troughton wrote:
>> On Thu, 2 Oct 2003 16:01:33 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
>>
>>
>>>David Sutherland wrote:
>>>
>>>
>>>>OS/2 cannot provide direct access to the hardware for other 32-bit
>>>>OS's (think Windows, if you are struggling to keep up) - it is plain
>>>>*stupid* to think it can do so for a 64-bit OS that didn't even exist
>>>>at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
>>>>that it can act as the bootstrap loader?
>>>
>>>I am writing this using OS/2 running on an Opteron system. <laugh>
>>
>>
>> Can you start a 64 bit system in a VM under OS/2 on this system?
>>
>> This is the essential point in question.
>
>Well, actually, Sutherland's questions was 'will OS/2 even run on a
>64-bit cpu?' but I digress.
That was in the part that you snipped.
>> The answer is no.
>
>I don't see 64-bit code running as a 'VM' under OS/2 but rather as
>64-bit code. AMD has designed x86-64 CPUs to be able to switch quite
>easily between 32-bit and 64-bit capabilities. It seems as though it
>would be possible to run 64-bit code from OS/2 that would transparently
>switch the CPU to long mode and execute the code and then switch back to
>legacy mode. This would probably require some sort of simple 'cpu mode
>selector' .sys driver that would be loaded by OS/2 at boot time. The
>real issue would be the hardware access of the cpu while it was running
>in long mode. All 64-bit code has to be marked as 64-bit in the LDT
>table and if code is not marked as 64-bit code it is presumed to be
>32-bit code so I am not sure what it would do with the 16-bit os/2
>hardware drivers. In any case, it seems like it would be possible to at
>least run 64-bit 'plugins' in OS/2, possibly created with the new port
>of GCC for OS/2 which, I believe, has support for x86-64.
Actually, if what I've read is correct, while the CPU is in 64 bit
mode, it can't deal with 16 bit code of any kind. For example, if
you look at page 311 of the "Bios and Kernel Developer's Guide for
the AMD Athlon 64 and AMD Opteron Processors", you'll see that in
Long Mode the Load Access Rights Byte does not include any valid
descriptors for 16 bit segment descriptors, so while the machine is
in long mode there is no way that it can access 16 bit memory; it
quite simply doesn't understand the data stored there.
This would require the CPU to do a mode switch (overhead), read the
16 bit data (more overhead), transform it into either 32 bit or 64
bit data (more overhead), and then copy it into one of the 32 or 64
bit registers on the CPU (more overhead), then do another switch
into long mode (more overhead), then finally perform an operation on
it.
That is how it would handle 16 bit OS/2 devices.
You can of course manipulate 64 bit data right now on an OS/2 system
if you want to; what that requires is splitting the data into two 32
bit chunks and placing them in memory, and then performing
operations on each piece that is congruent with the way such an
operation would be done on a 64 bit CPU, so when you stitch them
back together you end up with a valid piece of 64 bit data. That
necessarily requires quite a few operations to deal with
effectively, and that's how you would have to do it on an OS/2
system running on a 64 bit CPU because the opsys doesn't understand
64 bit "quad words" (to use the nomenclature in the assembly
reference for the AMD 64 line of CPUs).
For example, to load such data, you would need to execute the "Load
Segment Limit" instruction, which copies a segment descriptor from a
16 bit register into another 16 bit register, or into a 32 bit or 64
bit register.
>> You could make something like loadlin that would load such a system
>> from OS/2, but it would be tantamount to shooting OS/2 in the head
>> and replacing it with the 64 bit system code.
>>
>> What the opteron wants is a 64 bit system that can create 32 bit
>> VMs.
>
>No, the Opteron does not want that.
Oh yes it does. At the very least, to be pedantically precise, any
sane user of an opteron based system wants that.
>> From what I've read, the opteron can create complete
>> fully-virtualised 32 bit machines when running in 64 bits;
>
>I have not read of 'fully-virtualized 32-bit machines in Opteron.'
>Perhaps you are thinking of something else. When the Opteron runs in
>'legacy mode' (LMA=0), it *is* a 32-bit machine, not a 'virtualized'
>32-bit machine.
And, at the same time, it can also do that under the control of a 64
bit system.
You obviously have no idea how a virtualised 32 bit virtual machine
on such a system would be implemented. Here's a hint; it requires
software that would be part of the opsys to do it.
The question is, are there opcodes on the chip to support such a
beast?
It appears that there are. Furthermore, AMDs marketing material
makes it clear that the ability to create them was a key goal in the
design of the chip.
One key issue is the ability to invalidate the memory caches
completely; there are several operations to manage this, depending
on whether the contents of the cache are considered important or
not, ranging from straightforward invalidation of caches (contents
just disappear), through fast writeback (CPU continues execution
after internal caches are written back and flushed, but before
external caches are), all the way up to slow writeback and flush
(internal and external caches are written back to main memory,
verified, and then cache contents are invalidated, and then the CPU
resumes execution). This is important during task switching in which
the mode of the CPU changes, which of course would be a key part of
creating a 32 bit VM under a 64 bit system.
BTW- the operation to modify the machine status is called "Load
Machine Status Word" and it copies the binary string you want to use
to tell the CPU which mode it's in into register CR0; that's on page
320 of the reference.
>> you could
>> use this to start a 32 bit system in a VM. However, such a system
>> would only be able to get virtual access to hardware like video
>> cards and the like. This would result in a performance penalty for
>> certain types of operations, and games that stress the performance
>> of the hardware (such as a 3d video card) would suffer as a result.
>
>The operating system has whatever access to the hardware is allowed by
>its design. There is no inherent 'virtualization' of the 32-bit or
>64-bit modes in the x86-64 cpus such as Opteron, Athlon 64, or Athlon FX-51.
>
>>
>> It's the complete virtualisation that leads to the possibility that
>> AMD could come out on top of Intel with this chip; being able to run
>> multiple VMs virtualised in hardware means that one can do something
>> like virtual PC with performance on par.
>
>No, this is all wrong. There is no 'virtualization' of 32-bit or
>64-bit. The Opteron cpu, like the Athlon cpu, and even the old K6 cpu
>translates the x86 instructions into RISC operations (ROPS) in the first
>few stages of the execution pipeline because it is basically a RISC cpu
>but this is not '32-bit virtualization' or '64-bit virtualization'.
>When running in legacy mode, the *hardware* registers are 32-bit and
>when running in long mode they are 64-bit.
Of course it isn't. The fact that it translates CISC operations into
a number of RISC operations doesn't have anything to do with
implementing 32 bit VMs under a 64 bit system. However, there's a
slight confusion on your part. There are 16 bit registers, 32 bit
registers, and 64 bit registers on the CPU. Which ones the CPU can
read depends on which mode it's in. It's not like the register (a
structure on the chip) magically halves or doubles its size
depending on mode; the CPU is able to read some registers or other
registers depending on which mode it's in. In order to do that, you
use the LMSW operation above to change the contents of register CR0.
>> In fact, IIUC, when you're running a 32 bit system on Opteron, the
>> chip itself is simply running one VM on top of its 64 bit system,
>> and giving it exclusive access to devices.
>
>This is completely wrong. When you are running a 32-bit operating
>system on Opteron, it is running in 'legacy' mode and the *hardware* is
>32-bit. The CPU both appears to be 32-bit to the OS and _is_ 32-bit.
>When it switches to long mode, it is 64-bit. There is no virtual
>machine inherent in the Opteron in either 32-bit or 64-bit modes.
There's no VM inherent in a 386 either; it's all in how you
manipulate the opcodes in the chip's assembly language as to whether
you can create one or not.
The big question is whether or not the Athlon 64 line of CPUs does
have the necessary operations in place to implement one in software.
It does, both according to their marketing materials and their
technical documentation.
One interesting question is whether those operations could be used
to create 64 bit VMs as well.
One can, of course, create a virtual machine on any system; the
question is performance. If the virtual machine is able to be run
strictly using the chip's opcodes, then performance will be good. If
some of those operations need to be implemented in software (as, for
example, in Bochs or even in programs like VMWare and VPC) then
there will be a performance penalty for doing so.
Being able to use the chip's native operations to create a VM is
what one is talking about when discussing "VMs virtualised in
hardware", as oppsed to VMs virtualised with hardware and software,
such as is found in VMWare and VPC on the 32 bit platform.
One has to understand about what is meant by the bitness of the
machine as a whole, as opposed to the bitness presented to the
OpSys. When you're running a 32 bit system like warp on opteron, the
chip is using it's 32 and 15 bit registers, some parts of the chip
that perform operations are used across both 32 bit and 64 bit mode,
and some parts of the chip that perform operations are exclusive to
either 64 or 32 bit mode. When you're using a 32 bit system, you're
using the components that are relevant to 32 bit operation, while
the components that aren't simply sit idle. In order to enforce,
this, one loads the machine status word so that it can decide
whether an attempt to access a particular register or operation is
legitimate considering the mode it's in.
Stilz
Good stuff!
FYI - What pulled me toward Knoppix at first was the "full" set of app's
it comes with .... you can try almost everything needed for a
non-developer ("normal" user?) environment and get a feel for Linux
without changing anything on your hard disk. It never failed to boot
and configure on any machine I tried it on.
I've since installed my own build of the lilo and the kernal (why pay
for a distribution when you can do it yourself) on my development
machine.
Mike
David Sutherland
wrote:
>Hello World,
>
>> On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>>
>>>A company named SCI has developed a bootable CD that contains a game and
>>>a Linux OS that can be booted on any Opteron PC to turn it into a
>>>powerful gaming console.
>>>
>>>This is a BIG deal.
>>
>
>David Sutherland wrote:
>> Is it? Bootable linux distributions have been around for a while.
>> Bundling a software title with it isn't particularly big news.
>
>What is new about it is that you no longer need Windows to run your
>game.
How is that "new"? The game was *already* available for Linux along
with many others! In case you haven't noticed there are a *lot* of
games available for platforms other than Windows.
> Apparently the "Gamestorm" rendering engine developed by SCI is
>good enough to develop a game like "America's Army".
>
Umm....what makes you think that SCI wrote the game? Geez, have all
you "advocates" been banned from using google or something?
All SCI have done is taken the *existing* game and bundled it with an
existing Linux.
>> How interesting - you're not sure what version of directx is current,
>> but you know it's a "miserable kludge"? Sounds like
>> wishful-thinking wrapped around ignorance to me.
>
>Never had a game crash on you because another game insisted on upgrading
>DirectX?
Actually no, I haven't. Or at least, not in the last couple of years.
Can you give me an example of where this has happened to you?
>That's the whole problem with Windows game development. You are
>dependent on a system that can change under your fingers at any moment.
>
And which can be updated to the latest spec and drivers as well.
>If you make the game bootable, you have full control of the system.
>Nobody is going to change the OS on you. If you do it under Linux, this
>will also remove your dependency on Microsoft. Linux you can stick on
>your CD for nothing.
>
Sure, and so long as your CD can support every single piece of
hardware out there - including stuff that hasn't even been released
yet - you'll do just fine...
>> Why? What API are all the game developers supposed to write for? You
>> surely don't expect them to go back to writing directly to the video
>> hardware?
>
>"Gamestorm" apparently. Developed by SCI.
>
ROTLFMAO.
Americas Army is built on the Unreal Warfare engine. Unreal comes
from a - *gasp* - Windows background. It uses OpenGL *or* DirectX.
Given that this incarnation is running on Linux it is safe to see that
we are seeing the OpenGL variety here.
http://cleaned.beyondunreal.com/UT2FAQ/UT2FAQ.htm#_Toc6765476
"Apparently", you are quite entirely wrong in your beliefs about what
"gamestorm" *is* :)
>> Also, what happens when new hardware comes out that isn't supported by
>> the version of linux on that CD? You've only got SATAII drives but
>> your now two year old game CD doesn't have drivers? Uh oh.
>
>CD writers are your friend.
Wait a second - suddenly this "no-brainer" way of running a game
requires that you build your OWN FRIGGING RUNTIME CD?
ROTFL!!!
>You could download a fresh ISO image and
>burn it.
Gee, that really sounds like a quick and easy way of making sure all
your games are using the latest drivers. Not.
Hay, want to download a patch to the game? Here's another 600MB iso
to download! Have fun kids!!
>Or you could arrange to park fresh drivers on your harddisk
>somewhere. I downloaded Knoppix and am running it under VPC now.
And you think that you are going to be able to run this 64bit release
in the same way?
>eCS
>Electronic Delivery is under way. I think that shrink-wrapped software
>is in its last days.
>
Eventually that will happen - but this CD is still a physical product
which means it has be to packaged, sold and transported. Not exactly
a great poster-child for your brave new world of download-only
software.
If you want to see a working example of the kind of system I expect to
see in the future, look at www.steampowered.com.
>Illegal copying wouldn't be a problem. All of these games will probably
>be tied to your gaming account that lives on the Internet. I wouldn't be
>surprised if you could get the game for free, and play it offline. Just
>to give you a taste of what lies on the Internet. And then you'll need a
>login on the game server, which is what you pay for.
>
You are assuming that everyone has broadband. They don't. 600MB is
a *LOT* of data for most people. The 3 or 4 CD's of modern games are
not going to be a viable download for Joe Public any time soon.
>>>Microsoft has
>>>pushed Directx as a replacement for DOS as a gaming platform but no one
>>>has ever loved (or even liked) directx
>>
>> *cough* - really? And on what evidence do you base this assertion?
>> DirectX and OpenGL have paved the way for modern gaming. What
>> developers used to hate was having to write the whole damn render
>> engine from scratch for DOS.
>
>Well, now MS has pissed them off to such an extent that they are going
>back to developing their own rendering engines. That should tell you
>something about how much fun it is to develop for DirectX.
>
LOL - if only your entire premise wasn't so deeply wrong. Unreal
Warfare uses OpenGL and DX. Americas Army is built on the Unreal
Warfare engine.
SCI have *not* created any new rendering system!
All your bitching about how dreadful DX is for developers is absolute
nonsense - go away and read what John Carmack has said about it, or
read the comments in the link I provided above.
You'll seem a lot better informed if you do :)
>I think this is a good trend.
Unfortunately for your argument, your "trend" consists of nothing but
your own ignorance of the technologies being used, and your own
wishful thinking.
>Traditionally, one of the great selling
>points for Windows has been that you needed it for games (like "how many
>tries to get this printed out correctly").
Hmmmm....and OS/2 of course has such fantastic printer support that
Windows users should be jealous, right?
>Now, you no longer need
>Windows to play games. Which will make people think about what you need
>Windows for at all.
>
All your claims spring from your deep misunderstanding about what
Gamestorm is. Most people don't make the same mistakes and arrive at
rather different conclusions.
>Cheers/2,
>Menno
David Sutherland
wrote:
>Hello World,
>
>Irv Spalten wrote:
>>
>> If they can get enough game developers to write to the API set they
>> create, and the big if, will it be complete enough.
>
>Actually, no. Whether anyone else would use the API would be besides the
>point. SCI can use it regardless to produce games that are independent
>of the underlying OS. SCI has the assets. They can use it to produce CD
>bootable games. As long as those games look good, SCI can compete. They
>could, of course, sell the license to others.
>
SCI have NOT written any game code.
The game is built on the Unreal Warfare engine, which uses D3D or
OpenGL. Your entire argument is specious.
[flawed arguments born of misapprehension snipped]
David Sutherland
Put that CD in your Xbox or GameCube and get back to us on how well it
runs.
Stilz
David Sutherland wrote:
>
> On Wed, 01 Oct 2003 12:55:00 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
>
> >David Sutherland wrote:
> >> On Wed, 01 Oct 2003 09:29:09 -0700, "David T. Johnson"
> >> <djoh...@isomedia.com> wrote:
> >>
> >>
> >>>A company named SCI has developed a bootable CD that contains a game and
> >>>a Linux OS that can be booted on any Opteron PC to turn it into a
> >>>powerful gaming console.
Did I miss some here? Where in the released documentation this it say
that it "...can be booted on ANY Opteron PC to...."? A call to my AMD
rep. sent me here by the way -
http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543_5730~75278,00.html
where is says (along with a lot more) "The GameStorm technology for AMD
Athlon 64 processors, featuring America's Army, was created in
co-operation between AMD, SCI, NVIDIA and the US Army."
I'll wager it works best with a NVIDIA adapter, "The First Choice of
Gamers around the World".
<snipped>
Stilz
Jack Troughton wrote:
>
> On Thu, 2 Oct 2003 18:28:25 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
>
> >Jack Troughton wrote:
> >> On Thu, 2 Oct 2003 16:01:33 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
<snipped>
Good post Jack but I think it's wasted on DTJ.
Now if I could only find a buyer willing to meet or come close to my
asking price for a Beech Baron 58 with only 5000 TT and less than 50
hours SMOH on each engine (full suite of avionics inc. S-TEC AP), I
might be able to replace these "slow" 3.2GHz/2GByte Intel machines with
some Opteron-based gaming boxes and finally leave the stone age.
Nothing wrong with the Beech by the way, I just find myself in the GS
(G550 ER) more of fen. :-))
M.
David Sutherland
<djoh...@isomedia.com> wrote:
There is. References to the people producing them have already been
provided.
That's why it isn't big news - although it may be to *you*.
>DOS games came close to that when
>you would boot DOS from a floppy and run the game. The problem with
>doing the bootable CD/OS game on newer systems, though, is support for
>the hardware. The OS on the CD has to boot and recognize the hardware,
>provide the proper drivers and software, configure itself, and then run
>the game and it has to do that for the wide variety of hardware that is
>out there. Apparently that has been sufficiently difficult to do that no
>one has wanted to do that.
>
Gosh, you almost sound like you're learning.
>>
>>
>>>>
>>>>>Microsoft has been pushing their 'directx'
>>>>>technology as a way to get a high-quality gaming experience while still
>>>>>running Windows and is up to 'Directx 9' I think. Those of you who have
>>>>>struggled to get cranky directx Windows games to actually run like they
>>>>>are supposed to know what a miserable kludge the who directx platform is
>>>>
>>>>
>>>>How interesting - you're not sure what version of directx is current,
>>>>but you know it's a "miserable kludge"? Sounds like
>>>>wishful-thinking wrapped around ignorance to me.
>>>
>>>Sounds like first-hand experience with DirectX to me...
>>>
>>
>>
>> Yes, I have considerable first-hand experience with both DirectX and
>> OpenGL games. That's why I feel qualified to comment.
>>
>> You just got done telling us that you don't know what the current
>> version of DirectX is. Hmmmm....
>
>I don't care what the current version of DirectX is, for sure. <laugh>
>
That's sweet, but doesn't deflect from your ignorance.
>>
>>
>>
>>>>
>>>>>and will immediately see the value in the bootable linux CD as a gaming
>>>>>platform.
>>>>
>>>>
>>>>Why? What API are all the game developers supposed to write for? You
>>>>surely don't expect them to go back to writing directly to the video
>>>>hardware?
>>>
>>><laugh> Not hardly. There are very few details available but the fact
>>>that the game is only available in a 64-bit version running on the
>>>Opteron and requires a platform called "Gamestorm" suggests that the
>>>"gamestorm" platform is somehow doing something different than the old
>>>32-bit OpenGl stuff.
>>
>>
>> Is that what it "suggests"? Really?
>>
>> How about you tell us what exactly it is that is so different about
>> this platform? What new cross-platform API's does it introduce that
>> is going to make porting to OS/2 such a breeze?
>
>Hard to say, but they certainly cannot be less accessible than DirectX.
> <laugh>
>
"hard to say" - in other words you've got no clue - "<laugh>"
Why don't you provide a list of your grievances with the accessibility
of DirectX, as you claim to be such an expert?
You still haven't explained what magic Gamestorm is performing that
eliminates the need for drivers, David. When are you going to get to
that?
>>
>> It suggests that it's the 64bit version of Linux, which won't run on a
>> 32bit processor. Pretty simple, really.
>
>Yes, it is obviously running 64-bit Linux.
>
And...?
>>
>> What it also suggests to me is that this company is trying to
>> capitalise on the recent launch of 64-bit consumer CPU's - in case
>> that had passed you by.
>>
>> Can you also educate me about what makes "Gamestorm" different from
>> standard Linux? What's included in there beyond what's already
>> available? Or are you just eating up those marketing blurbs as if
>> they really mean something?
>
>No details about "Gamestorm" are available yet although a lot of people
>are interested. It seems to have the capability of providing a
>high-quality 3D game on Linux without using either DirectX or OpenGL and
>appears to be able to run on any hardware. See below.
>
ROTFL. What game engine is America's Army built on, David? And what
API's does that engine support? Do you know? I don't think you do :)
Let me help:
http://www.unrealgaming.com/website/powered.php#americaarmy
http://cleaned.beyondunreal.com/UT2FAQ/UT2FAQ.htm#_Toc6765476
No doubt you will now withdraw all your claims about this new and
magical gaming API from SCI? Also your claims about how developers
such as those writing AA are running screaming from DX?
Or would you prefer to continue making a complete ass of yourself in
public?
>>
>>
>>>It is the new 'gamestorm' 64-bit platform that
>>>apparently allows the bootable CD to provide a high-quality gaming
>>>experience without the need for the OpenGL and video drivers for
>>>specific hardware.
>>
>>
>> LOL - right. Gamestorm magically supports every bit of hardware out
>> there without the need for specific drivers. Nvidia and ATI can fire
>> their software people and never write another line of code - just
>> release the hardware and those utterly magical 64bit CPU's completely
>> eliminate the need for drivers. David T Johnson says so, so it must
>> be true right?
>
>When I read the Gamestorm story, my thought was that they are using the
>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>typically do on 32-bit systems. If so, the system would essentially run
>on any video hardware.
>
If your belief was true, sure. But it's not. You are as wrong as
it's possible to be.
>
>>
>> Do you read and think about what you write before you post? It sure
>> doesn't look like it.
>>
>> This is just a version of Linux. Linux needs hardware specific
>> drivers. Unless these guys have written a driver-less OS from the
>> ground-up, then you are spouting nonsense.
>
>Yes, video drivers are needed to run America's Army on 32-bit Linux. If
>someone developed a bootable CD with America's Army and 32-bit Linux, it
>would need to scan the hardware and find the appropriate driver for the
>video hardware that it detected. If the video hardware was newer than
>the CD, the driver might not be present. If the game did not have a
>driver for the video hardware that it detected, it would not be able to
>run. Etc. These severe restrictions have prevented the idea of a
>bootable CD with the app and OS on it from catching fire. None of this
>seems to be the case, however, with the 64-bit version of America's Army
>running on the mysterious "Gamestorm" platform.
ROTFLMFAO!!!!!!
It's not "mysterious" David, it's the 64bit version of Linux that's
been around for a while now. It's the recompiled version of America's
Army, which has been around for a while. It uses drivers from Nvidia
to drive the video card, just as on a 32bit system.
The only "mysterious" thing here is that you seem to think that anyone
takes you seriously. Go and read up on what you are talking about.
You've taken a marketing release and extrapolated a non-existent
technology from it. Your choking ego won't let you admit that you
could be wrong, so you introduce ever more fantastic tales to try and
plaster over the holes.
Good luck. No one believes you.
>The only hardware
>requirement that it lists is that a 64-bit Opteron be used. The
>announcement does mention that it scans the hardware, though, but it
>also mentions that it boots in one minute so it does not seem to be
>doing a lot with the scan.
>
Guess you haven't seen much of Linux lately.
>>
>>
>>>This would be required if the game was to work
>>>properly on the wide variety of video hardware that would be expected.
>>>I would speculate that it has something to do using direct access to the
>>>64-bit system memory for the video display in some way, as a guess. I
>>>am just guessing, though.
>>>
>>
>>
>> Yes, you are - guessing. And the very real problem of requiring
>> drivers to access the functionality of modern hardware in order to get
>> acceptable performance makes your guesses look wildly off base - if
>> not downright ridiculous.
>
>I agree that this is true with typical 32-bit systems. The
>Gamestorm/Opteron platform, however, has apparently found a way to get
>acceptable performance without using hardware-specific drivers.
>
And where is your evidence for this incredible claim? The same place
as the evidence for your claim that they wrote a whole new gaming API
and game engine from scratch?
Oh gee, that's right - they *didn't*, did they? You are talking
absolute bollocks as usual.
>>
>>
>>>>Also, what happens when new hardware comes out that isn't supported by
>>>>the version of linux on that CD? You've only got SATAII drives but
>>>>your now two year old game CD doesn't have drivers? Uh oh.
>>>
>>>They seem to be doing something with the 64-bit hardware that makes the
>>>'drivers' less critical. I will await further information, though.
>>>
>>
>>
>> ROTFLMAO.
>>
>> You should try eating less sugar, and swallowing less marketing BS.
>>
>> Are you *seriously* suggesting that the move to 64bit means nobody
>> needs drivers any more? *Muahahahahaha*!!!!!
>>
>> Guess you should let Nvidia and ATI know that they can fire their
>> development teams who are working on their 64bit drivers.
>
>Software to control hardware (i.e. drivers) will always be needed.
>However, this software does not seem to be hardware-specific on the
>Gamestorm/Opteron platform. As I mentioned, though, the details are not
>yet available.
>
How odd then that Nvidia are mentioned so prominently. What do they
make again?
You have almost no details - and what few you do have I have proven
wrong - but you are making the claim that it's the death knell for
DirectX? You're a laugh a minute!
Tell you what David - I'll bet you US$100 that the SCI bundle includes
drivers for graphics and sound cards. Care to take me up on it?
>>
>>
>>>>
>>>>>So how will this benefit OS/2? One of the chief obstacles of
>>>>>playing Windows Directx games on OS/2 has been lack of access to the
>>>>>directx software.
>>>>
>>>>
>>>>Well yeah, I guess not having directx on OS/2 would hamper your
>>>>ability to use directx on OS/2!
>>>
>>>yes.
>>>
>>
>>
>> I thoroughly agree. But aren't you going to state the obvious again
>> in case someone missed it the first time around?
>>
>>
>>>>
>>>>>If Linux becomes a serious game platform replacement
>>>>>for Directx, it is likely that Linux games will be able to be either
>>>>>ported to OS/2
>>>>
>>>>
>>>>Oh my god, you *really* don't have clue, do you? It hasn't happened
>>>>with OpenGL games, what do you think has changed?
>>>
>>>the new 'gamestorm' 64-bit platform, whatever that is, which does not
>>>seem to be using OpenGL.
>>>
>>
>>
>> Is that what it "seems" to be doing?
>>
>> Then again, given that AA is already released for Linux, a lot of what
>> "seems" like magic to you is really not all that terribly clever:
>>
>> http://www.3dgamers.com/games/americasarmy/
>>
>> Look for the word "linux". Sheesh.
>
>yes, there is a 32-bit version of America's Army for Linux.
>
You just got done claiming that it was a brand new invention of SCI
and their "mysterious" gamestorm platform. All done using their
all-new, all-magic gaming API that eliminates the need for drivers!
It can't be both an existing product *and* new, David. Try to keep
you story straight as you are already starting to flounder in your own
contradictions.
>>
>>
>>
>>>>The whole point of DirectX is that it provides a standard set of API's
>>>>for developers to write towards so they can get on with the job of
>>>>creating *games* and adding their own custom effects instead of having
>>>>to worry about the basics of drawing a triangle and making it green.
>>>>The *hardware* developers can then concentrate on accelerating those
>>>>standard routines so that games run *fast*.
>>>
>>>perhaps the 64-bit memory addressing and hypertransport of the Opteron
>>>allows some sort of new design paradigm?
>>>
>>
>>
>> Perhaps sacrificing a chicken at midnight allows a CPU to magically
>> address a 3D accelerator board without the need for *ANY GRAPHICS
>> INSTRUCTIONS*.
>
>My speculation is that the Opteron CPU and the Gamestorm platform are
>somehow doing stuff in the CPU that was formerly done by the
>'accelerator' board.
They don't. Nvidia and ATI have already announced 64bit driver
support for the new platforms. Why do you think they have done this?
All Gamestorm gives you is a bundled OS and application. The
application in question is using OpenGL, an existing API. There is no
magic here, David - go down the bottom of your garden and watch out
for the pretty lights instead if that is what you are after.
>The Opteron CPU has a lot of new hardware function
>that 32-bit CPUs do not. For example, it has more registers, many new
>x86-64 instructions, the ability to address more memory, etc.
>
Yes, and it still can't do 3D acceleration as fast as a hardware board
can do it. Or are you suggesting that this is the "death knell" for
ATI and Nvidia as well? Even stranger then, that this bundle was
created with NV's assistance!
>>
>> Or perhaps you are just talking out of your ass?
>>
>> Hey, are you going to address any of my points? Nah, didn't think so.
>> It's all about the API, stupid! You might as well suggest that the
>> Win32API or OS/2's API's are no longer required - it makes just as
>> much (aka zero) sense.
>>
Nothing to say, David?
>>
>>>>
>>>>>or run on OS/2 with emulation software.
>>>>
>>>>
>>>>For this to happen you need to have a standard API supported across
>>>>both platforms - and the video card drivers need to accelerate it if
>>>>you don't want to play your games at one frame every four seconds.
>>>
>>>If the new 'gamestorm' 64-bit platform can boot independently and run on
>>>any Opteron hardware without the need for special drivers,
>>
>>
>> Gah - this is so stupid it makes my teeth hurt.
>>
>> You cannot access the hardware without drivers.
>
>Agreed.
You are contradicting yourself again.
>However, the Opteron/Gamestorm platform does not seem to
>require separate drivers for each different video card. At least, they
>have not listed any particular video hardware requirements in their
>announcement which one would expect they would otherwise do.
They have listed almost *nothing* in their announcement, which reads
like a puff-piece from a marketer who knows a good buzz-word when he
sees one. I'm not surprised you've gone all starry-eyed - you were
always a sucker for a salesman.
If you want to believe that they have eliminated the need for drivers
by injecting magic into the polycarbonate of their CD, then you go
right ahead. And I will go on taking the piss out of you :)
>
>
>>
>>
>>>it may be
>>>possible for it to also boot it on OS/2 in some way similar to the way
>>>OS/2 can boot a 'DOS image' or 'boot from drive A' and such.
>>
>>
>> <WIPES TEARS FROM EYES>
>>
>> So, you are going to boot a 64-bit OS from within your 32-bit OS/2,
>> and it's going to be able to use it's magic routines to directly
>> access all your hardware *without drivers* so that you get the fastest
>> games machine around!
>>
>
>There has been no claim by SCI that their machine is "fastest" although
>it was termed a 'console machine on steroids'.
>
Hmmm....David? You seem to be ignoring the point of what I said. Is
it because you accept that you will never be able to run this
environment from within OS/2? A simple yes or no will do ;)
>>
>>>I admit
>>>that I am just wildly speculating here, though. Since there seems to be
>>>no detailed information available, that is all there is to do.
>>>
>>
>>
>> ROTFL.
>>
>> You *could* try applying simple logic or even reading up on what the
>> latest 64bit CPU's and OS/s are capable of, *before* you make stupid,
>> wild, and profoundly ignorant guesses that make you look ridiculous.
>>
>>
>>>>The only graphics API shared by both Linux and OS/2 to my knowledge is
>>>>OpenGL. Last time I looked, OpenGL support on most cards for OS/2 was
>>>>pretty damn shabby where it wasn't non-existent. Linux is better, but
>>>>still not as good as Windows.
>>>
>>>There does not seem to be a requirement for the 'gamestorm' 64-bit
>>>platform to use OpenGL or else perhaps they have implemented it in some
>>>sort of virtual hardware implementation or something.
>>
>>
>> Which just happens to be able to take advantage of accelerated 3D
>> cards, all without any need for drivers.
>>
>> Which sentence *precisely* leads you to this conclusion?
>
>They don't seem to be using the 'accelerators' in the video hardware
>because they don't list any particular hardware requirements and they
>suggest that it will boot on any Opteron computer:
So it's the LACK of information that has lead you to your conclusions!
I believe that my initial comments were about your ignorance and
wishful thinking, and you've just proven that true.
> "They can run the
>game on a floor model at Best Buy, and leave the hard drive untouched,"
>he says."
>
Just as you can with a distro from other Linux providers. No magic to
see, just code and drivers.
>Other claims by SCI:
>
>"cuts out legacy operating system overhead..."
Such as? Does this mean no more drivers?
>"without the usual overhead from the legacy operating systems
>raditionally used in the gaming industry, SCI claims."
Is this the bit that you think means no more drivers?
>"With the AMD Athlon 64 processor and GameStorm technology, AMD is able
>to showcase a fully-integrated 64-bit environment that delivers
>performance and realism to the most demanding gamers,"
LOL - I believe Nvidia have been making identical claims for some
years now! Did *they* eliminate the drivers?
>"the company is now working on porting several more 32-bit and 64-bit
>games to the GameStorm platform"
>
>The last sentence is revealing because it suggests that 'porting' is
>required to run other games on the Gamestorm platform which suggests
>that Gamestorm does not run OpenGL games.
What it *suggests* to you is irrelevant, because the FACT is that
Americas Army is built on an OpenGL/DX engine, which was launched
initially on Windows.
What you *want* to believe and read into press releases is not the
same thing as reality. Life is a bitch sometimes, isn't it?
>Also, there are no 64-bit
>x86-64 OpenGL libraries that I have heard of yet (although I think they
>are in development) and the game also certainly does not use DirectX
><laugh>.
>
"<laugh>" You should be careful what you believe. You just end up
looking stupid by laughing about your own ignorance :)
If the game exists in a 64bit version (although come to think of it
there is nowhere that you can buy or download such a version) means
that the code is already done.
The game *can* use DirectX as I have already proven - but for Linux
OpenGL is the obvious API. Either way, SCI didn't write it and it is
nothing new - certainly not the holy grail of platform indepence that
you have been boosting it as!
>>
>>
>>>Obviously, the
>>>'gamestorm' platform is doing something with 64-bit memory addressing
>>>that was not possible to do with 32-bit addressing.
>>>
>>
>>
>> The only obvious thing here is that you haven't the faintest clue what
>> you are prattling on about. Moving to 64-bit has increased the amount
>> of memory that can be accessed and has increased the bandwidth
>> available for calculations ( you can now work on 64 bits of data!).
>> What it *hasn't* done is eliminate the need for drivers or API's.
>
>Software is needed to operate hardware but it does not have to
>necessarily be specific to every piece of hardware. For example, the
>BIOS on my machine can display stuff on any video card that I insert.
Yes, by using lowest-common-denominator routines (VGA). Very slow,
very limited.
>It can also operate the keyboard, floppy disk, hard disk, CD-ROM, ZIP
>drive, and USB devices without any additional drivers and before the
>"operating system" ever even boots.
And all with very limited functionality and only because all the
hardware offers bare-bones interfaces analoguous to VGA.
> It seems possible that SCI is doing
>more with the CPU and less with the video hardware than is done on
>32-bit systems.
>
It seems possible to you because you don't have a clue. Read about
the game engine, read about the new CPU's, read about the new OS's
coming down the pike. Then get back to me.
>>
>>
>>>>The other problem with OpenGL of course, is that the current standard
>>>>is years old, and lacks a lot of the cool features that DirectX allows
>>>>(particularly shader support) - the next OpenGL standard is due soon
>>>>which will address this, but the chances of it ever seeing the light
>>>>of day for the OS/2 platform are somewhere between a snowballs chance
>>>>in hell and zero.
>>>>
>>>>No support means no ports.
>>>
>>>Perhaps OpenGL is morphing into something new and different or is being
>>>superseded by something new and better.
>>
>>
>> There's no "perhaps" about it. Try google - you might *learn*
>> something.
>>
>> Try "OpenGL" and "2.0" for starters.
>
>The Gamestorm platform is apparently not using either OpenGL or DirectX.
> See above.
>
See above. "Apparently", you are wrong. "<laugh>"
>>
>>
>>>In any event, DirectX seems to
>>>be disappering from the equation if it is possible to insert a CD-ROM
>>>into an Opteron machine and boot and run a high-performance game on any
>>>hardware.
>>>
>>
>>
>> Utter nonsense.
>
>I am certain that DirectX is not being used by the Gamestorm platform
>which means that DirectX is disappearing from the 'picture.' <laugh>
>
Ho hum - I can always tell when you're toast because the "<laughs>"
increase in frequency as your arguments decrease in coherency.
The game that you are so excited about comes in both OpenGL and
DirectX flavours. It's not magic, David. Sorry to burst your bubble.
>
>> As I've already pointed out, there are major
>> limitations with hardware support once you are reliant upon a fixed
>> point-in-time OS to run the game. How do you support future hardware
>> that doesn't exist yet? When new drivers come out that increase
>> performance or fix bugs how do you benefit?
>
>Indeed. See above. These are significant limitations to the entire
>idea of the bootable CD-with-OS that SCI may have overcome by using the
>64-bit Opteron.
>
In other words - magic. You think that the Opteron is not *just* a
faster higher-bandwidth CPU, but instead it's able to immediately
determine the capabilities of every singe piece of hardware just by
being in proximity with it - and then use it at peak performance:
This sound card has hardware DTS6.1 encoding? No problem, the CPU
will devise routines and the necessary code to use it even though it's
never been seen before. This video card allows for dynamic
overclocking based on a temperature reading? No problem, the
"Mysterious Mythical Optimagicon CPU" will figure that out in a flash
and make your every day a better one! Ignore the smoke coming from
behind the curtain!
LOL - you seriously believe that this is what's happening? You're
cracked.
>>
>> This product has some appeal - it's a clever idea, and if the game is
>> optimised for linux you will end up with a very solid performant
>> platform. Just don't expect it to work on the machine you own five
>> years from now, and don't expect it to support every piece of hardware
>> out there.
>
>We will have to wait to see what SCI says in upcoming weeks.
>
*You* certainly didn't. You jumped to conclusions and sounded the
"death knell" based on nothing but your own fantasies and ignorance.
>
>> Also don't expect to be able to connect to the internet to
>> play multiplayer games without setting up all your configuration
>> details each and every time you boot. Hmmm....this is starting to
>> sound icky.
>
>"It feels like a gaming console on steroids and even allows for online
>access so you can connect to online game servers for multi-player
>action," said Jesper Jensen, CEO of Super Computer, Inc."
>
>It will be interesting to see how the Gamestorm platform deals with this
>issue.
>
Won't it just.
>>
>>
>>>>
>>>>>Microsoft has
>>>>>pushed Directx as a replacement for DOS as a gaming platform but no one
>>>>>has ever loved (or even liked) directx
>>>>
>>>>
>>>>*cough* - really? And on what evidence do you base this assertion?
>>
>>
>> LOL - none. No response. You haven't changed at all, have you David?
>>
>>
>>>>DirectX and OpenGL have paved the way for modern gaming. What
>>>>developers used to hate was having to write the whole damn render
>>>>engine from scratch for DOS.
>>>>
>>
>>
>> No comment, David?
>>
"Apparently" not :)
>>
>>>>>and now it looks as if Linux has
>>>>>skewered it right through its gizzard.
>>>>
>>>>
>>>>LOL. Sure - maybe on your planet.
>>>
>>>We'll see as more details become available.
>>
>>
>> I've seen. The difference is, *I* understood.
>>
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>
>>
>> OS/2 cannot provide direct access to the hardware for other 32-bit
>> OS's (think Windows, if you are struggling to keep up) - it is plain
>> *stupid* to think it can do so for a 64-bit OS that didn't even exist
>> at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
>> that it can act as the bootstrap loader?
>
>I am writing this using OS/2 running on an Opteron system. <laugh>
In 32bit mode - "<laugh>"
I note that you did not address the real issues. I wonder why?
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
>This is completely wrong. When you are running a 32-bit operating
>system on Opteron, it is running in 'legacy' mode and the *hardware* is
>32-bit. The CPU both appears to be 32-bit to the OS and _is_ 32-bit.
Hmmm - so you are saying that 32bit OS/2 does *not* run on a 64bit CPU
as when it runs 32bit code the hardware _is_ 32bit.
Or are you now going to contradict yourself? Again.
"<laugh">
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
>Current
>versions of OS/2 run in legacy mode but it seems possible that OS/2
>could be used to boot to another application/OS in long mode.
>
I think there are other bootloaders on the market that have a smaller
footprint and a faster boot time than OS/2. So why bother?
And I love the "current versions" qualifier. When exactly are you
expecting the next release of OS/2?
David Sutherland
wrote:
>Hello World,
>
>David Sutherland wrote:
>> Your whole "this is good news for OS/2" song and dance routine is just
>> silly. OS/2 is rendered utterly irrelevant by such a product.
>
>As is Windows. We're working on the assumption here that everything
>that's bad news for MS is good news for the rest of the industry.
>
Somehow I don't think Redmond is losing any sleep over a bootable CD
with a game bundled.
>The selling point "Yeah but you need It to play games" is falling away.
You haven't noticed the titles available for Linux - or the consoles?
Do *they* need windows to run?
>Conversely, the OS/2 disadvantage "But there ain't no games for it" is
>also falling away because you can now play games despite the fact that
>you have OS/2 on your PC and not Windows. Hence, this development is
>good for us and bad for Microsoft. And about time too!
>
So you expect people to reboot their machines every time they feel
like spending a few minutes playing a game? Menno, I resent having to
put the CD in the drive to play a game now *without* having to reboot.
If they released a game that meant I had to shut down my system every
time I wanted to play, *I wouldn't buy it*. If you look at the
numbers of people looking for cracks to disable the cd checks, you'd
understand that I am not alone in this.
I like being able to drop out of a game to my desktop, do something
else for a few minutes - maybe check something out on the net, and
then click at the bottom of the screen and I'm straight back where I
left off in the game. A stripped down environment with none of my
customisations and apps holds absolutely zero appeal.
>It's also good for Windows users, because it is another lever that
>breaks MS' stranglehold on your PC. The less stuff you actually *need*
>Windows for, the better.
>
I *choose* Windows because of the mix of software and stability. When
another platform gives a better mix, I'll move.
>Cheers/2,
>Menno.
Jazz
Not to want to put too fine a point on it, but if the producers really
desired so[1], they could easily[2] make it work on an Xbox or a PS2.
GameCubes I don't know. Don't they use some weird file system?
____
[1] But then, why would they? "Look, if you put our game in the PS2, it
will start up and play just like any other native game but get this: It
runs on -- wait for it -- *Linux* !"
[2] More easily on the PS2 than on the Xbox, given that Sony quite
happily aided the development of a Linux version for their game
platform. Microsoft OTOH has, for some reason or other, been less
cooperative vis-a-vis their /extremely/ /popular/ box. Still, you can't
keep a penguin down...
Jason Bowen
David T. Johnson <djoh...@isomedia.com> wrote:
>I haven't seen another game that came with its own OS on a bootable CD
>although there might one somewhere.
Some Segadreamcast CD's came with WinCE as the OS. What you don't know
could fill a warehouse.
>DOS games came close to that when
>you would boot DOS from a floppy and run the game. The problem with
>doing the bootable CD/OS game on newer systems, though, is support for
>the hardware. The OS on the CD has to boot and recognize the hardware,
>provide the proper drivers and software, configure itself, and then run
>the game and it has to do that for the wide variety of hardware that is
>out there. Apparently that has been sufficiently difficult to do that no
>one has wanted to do that.
>
Not too difficult at all, no more difficult than doing it from a hard
drive while booting. There are other reasons for not doing it, such as
whether or not you will get a return on investment.
>
>I don't care what the current version of DirectX is, for sure. <laugh>
>
If you don't, why'd you make comments about it? <laugh>
>
>Hard to say, but they certainly cannot be less accessible than DirectX.
> <laugh>
>
John Carmack disagree with you. I'll stick with the opinion of the
creator of some of the best selling games of all time over a blowhard on
usenets opinion. <laugh>
>
>No details about "Gamestorm" are available yet although a lot of people
>are interested. It seems to have the capability of providing a
>high-quality 3D game on Linux without using either DirectX or OpenGL and
>appears to be able to run on any hardware. See below.
>
So that means somebody has to use a new SDK or learn a new API?
>
>When I read the Gamestorm story, my thought was that they are using the
>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>typically do on 32-bit systems. If so, the system would essentially run
>on any video hardware.
>
Oi, you are ignorant. The rest of your nonsense about the CPU somehow
giving a video card new capabilities snipped. You and Eddie are truly
from the same cloth.
David T. Johnson
> On Thu, 2 Oct 2003 18:28:25 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
>
>
>>Jack Troughton wrote:
>>
>>>On Thu, 2 Oct 2003 16:01:33 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
>>>
>>>
>>>
>>>>David Sutherland wrote:
>>>>
>>>>
>>>>
>>>>>OS/2 cannot provide direct access to the hardware for other 32-bit
>>>>>OS's (think Windows, if you are struggling to keep up) - it is plain
>>>>>*stupid* to think it can do so for a 64-bit OS that didn't even exist
>>>>>at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
>>>>>that it can act as the bootstrap loader?
>>>>
>>>>I am writing this using OS/2 running on an Opteron system. <laugh>
>>>
>>>
>>>Can you start a 64 bit system in a VM under OS/2 on this system?
>>>
>>>This is the essential point in question.
>>
>>Well, actually, Sutherland's questions was 'will OS/2 even run on a
>>64-bit cpu?' but I digress.
>
>
> That was in the part that you snipped.
>
>
>>>The answer is no.
I should have said earlier that the answer to Sutherland's question
'does OS/2 run on opteron?' was 'yes'.
>>
>>I don't see 64-bit code running as a 'VM' under OS/2 but rather as
>>64-bit code. AMD has designed x86-64 CPUs to be able to switch quite
>>easily between 32-bit and 64-bit capabilities. It seems as though it
>>would be possible to run 64-bit code from OS/2 that would transparently
>>switch the CPU to long mode and execute the code and then switch back to
>>legacy mode. This would probably require some sort of simple 'cpu mode
>>selector' .sys driver that would be loaded by OS/2 at boot time. The
>>real issue would be the hardware access of the cpu while it was running
>>in long mode. All 64-bit code has to be marked as 64-bit in the LDT
>>table and if code is not marked as 64-bit code it is presumed to be
>>32-bit code so I am not sure what it would do with the 16-bit os/2
>>hardware drivers. In any case, it seems like it would be possible to at
>>least run 64-bit 'plugins' in OS/2, possibly created with the new port
>>of GCC for OS/2 which, I believe, has support for x86-64.
>
>
> Actually, if what I've read is correct, while the CPU is in 64 bit
> mode, it can't deal with 16 bit code of any kind.
You apparently didn't understand that in the AMD technical documents,
"64-bit mode" is used as a reference to a sub-mode of "long mode." The
x86-64 cpus operate in either 'long' mode or 'legacy' mode depending on
the status of the LMA (long mode active) bit. In 'legacy' mode, the cpu
can actually boot and run DOS in real mode. In 'long' mode, there are
two 'compatibility' sub-modes for 32-bit and 16-bit code. The bottom
line is that the x86-64 CPU can (and does) run 16-bit code when running
in either 'long' mode or 'legacy' mode which are the only two operating
modes that the x86-64 cpus have. You might also be confused about the
difference between 'legacy' mode and the 'compatibility' sub-mode of
'long' mode. Legacy mode is one of two operating modes for the cpus.
Compatibility mode is just a sub-mode of Long mode.
> For example, if
> you look at page 311 of the "Bios and Kernel Developer's Guide for
> the AMD Athlon 64 and AMD Opteron Processors", you'll see that in
> Long Mode the Load Access Rights Byte does not include any valid
> descriptors for 16 bit segment descriptors, so while the machine is
> in long mode there is no way that it can access 16 bit memory; it
> quite simply doesn't understand the data stored there.
>
> This would require the CPU to do a mode switch (overhead), read the
> 16 bit data (more overhead), transform it into either 32 bit or 64
> bit data (more overhead), and then copy it into one of the 32 or 64
> bit registers on the CPU (more overhead), then do another switch
> into long mode (more overhead), then finally perform an operation on
> it.
>
> That is how it would handle 16 bit OS/2 devices.
Again, this is all wrong. From the AMD docs (AMD64 Architecture
Programmer s Manual Volume 1: Application Programming Publication No.
24592 Revision Date 3.09 September 2003):
"In compatibility mode, legacy 16-bit and 32-bit applications run
using legacy x86 protected-mode segmentation semantics. The
16-bit or 32-bit effective addresses generated by programs are
combined with their segments to produce 32-bit virtual (linear)
addresses that are zero-extended to a maximum of 64 bits. The
paging that follows is the same long-mode paging function used
in 64-bit mode. It translates the virtual addresses into physical
addresses. The combination of segment selector and effective
address is also called a logical address or far pointer. The virtual
address is also called the linear address."
Note that the 'compatibility' mode and the '64-bit' modes referred to
here are sub-modes of long mode. The cpu does not do any 'mode
switching' to access 16-bit memory when running in long mode.
>
> You can of course manipulate 64 bit data right now on an OS/2 system
> if you want to; what that requires is splitting the data into two 32
> bit chunks and placing them in memory, and then performing
> operations on each piece that is congruent with the way such an
> operation would be done on a 64 bit CPU, so when you stitch them
> back together you end up with a valid piece of 64 bit data. That
> necessarily requires quite a few operations to deal with
> effectively, and that's how you would have to do it on an OS/2
> system running on a 64 bit CPU because the opsys doesn't understand
> 64 bit "quad words" (to use the nomenclature in the assembly
> reference for the AMD 64 line of CPUs).
>
> For example, to load such data, you would need to execute the "Load
> Segment Limit" instruction, which copies a segment descriptor from a
> 16 bit register into another 16 bit register, or into a 32 bit or 64
> bit register.
All of this is irrelevant to the x86-64 and memory management in legacy
or long modes.
>
>
>>>You could make something like loadlin that would load such a system
>>>from OS/2, but it would be tantamount to shooting OS/2 in the head
>>>and replacing it with the 64 bit system code.
>>>
>>>What the opteron wants is a 64 bit system that can create 32 bit
>>>VMs.
>>
>>No, the Opteron does not want that.
>
>
> Oh yes it does. At the very least, to be pedantically precise, any
> sane user of an opteron based system wants that.
No, the Opteron does not want a 64-bit system that can create 32-bit
virtual machines. The Opteron can execute 64-bit, 32-bit, and 16-bit
code with excellent performance using its existing hardware without the
need to use any 'virtual machines.' Neither sane nor insane users would
want a 32-bit VM for Opteron.
>
>
>>>From what I've read, the opteron can create complete
>>>fully-virtualised 32 bit machines when running in 64 bits;
>>
>>I have not read of 'fully-virtualized 32-bit machines in Opteron.'
>>Perhaps you are thinking of something else. When the Opteron runs in
>>'legacy mode' (LMA=0), it *is* a 32-bit machine, not a 'virtualized'
>>32-bit machine.
>
>
> And, at the same time, it can also do that under the control of a 64
> bit system.
Yes, of course. The only way that the Opteron can even switch to 'long'
mode (LMA-1) is with a 64-bit system. When OS/2 runs on it, it is in
'legacy' mode and appears as a 32-bit cpu.
>
> You obviously have no idea how a virtualised 32 bit virtual machine
> on such a system would be implemented. Here's a hint; it requires
> software that would be part of the opsys to do it.
There is no 32-bit virtual machine on the x86-64 cpu as you claimed. It
is possible, however, to run virtual machine software on the Opteron or
any other compatible cpu. The virtual machine can have any
configuration that the software implements. I am running Virtual PC on
the Opteron but is just another 32-bit application to the Opteron.
Inside the Virtual PC, however, it creates a software CPU, disk
controller, video display, etc, that are all software constructs.
Again, this is all just so much more application software from the point
of view of the real underlying hardware.
>
> The question is, are there opcodes on the chip to support such a
> beast?
>
> It appears that there are. Furthermore, AMDs marketing material
> makes it clear that the ability to create them was a key goal in the
> design of the chip.
I don't know what you are talking about here. The Opteron includes some
new x86 instructions that provide much more powerful 128-bit floating
point operations than the older x87 stuff. Other than that, there are
no significant changes to the x86 instructions afaik. Any new version
of Virtual PC (or some such similar software) would be using essentially
the same x86 instructions as before.
>
> One key issue is the ability to invalidate the memory caches
> completely; there are several operations to manage this, depending
> on whether the contents of the cache are considered important or
> not, ranging from straightforward invalidation of caches (contents
> just disappear), through fast writeback (CPU continues execution
> after internal caches are written back and flushed, but before
> external caches are), all the way up to slow writeback and flush
> (internal and external caches are written back to main memory,
> verified, and then cache contents are invalidated, and then the CPU
> resumes execution). This is important during task switching in which
> the mode of the CPU changes, which of course would be a key part of
> creating a 32 bit VM under a 64 bit system.
A 32-bit VM is a software construct that would be created the same way
under a 32-bit or 64-bit (or 128-bit) cpu. If it were compiled with a
compiler supporting x86-64, it would be likely to be much faster than
the 32-bit version thanks to the improved floating point capabilities
and the new and larger registers present on x86-64 hardware.
You are confused about the x86-64 registers. In long mode, there are 16
general purpose registers that are 64-bits. In legacy mode, there are 8
general purpose registers that are 32-bits. There are also 16 128-bit
SSE2 registers in long mode and 8 128-bit SSE registers in legacy mode.
The 8-bit and 16-bit registers are implemented in the same way that
they were implemented in the 32-bit Athlon and earlier 32-bit cpus. For
the 16-bit data size, the general purpose registers are AX, BX, CX, DX,
DI, SI, BP, and SP. For the 32-bit data size, these are EAX, EBX, ECX,
EDX, EDI, ESI, EBP, and ESP. For the 64-bit data size, these include
RAX, RBX, RCX, RDX, RDI, RSI, RBP, RSP, and R8-R15. There is a good
graphic of this on page 7 here:
http://www.amd.com/us-en/assets/content_type/DownloadableAssets/MPF_Hammer_Presentation.PDF
The registers are in the same place on the silicon real estate but their
number and addressable size changes depending on the status of the LMA
bit.
>
>
>>>In fact, IIUC, when you're running a 32 bit system on Opteron, the
>>>chip itself is simply running one VM on top of its 64 bit system,
>>>and giving it exclusive access to devices.
>>
>>This is completely wrong. When you are running a 32-bit operating
>>system on Opteron, it is running in 'legacy' mode and the *hardware* is
>>32-bit. The CPU both appears to be 32-bit to the OS and _is_ 32-bit.
>>When it switches to long mode, it is 64-bit. There is no virtual
>>machine inherent in the Opteron in either 32-bit or 64-bit modes.
>
>
> There's no VM inherent in a 386 either;
...or in a 286...
> it's all in how you
> manipulate the opcodes in the chip's assembly language as to whether
> you can create one or not.
the VM is a 'software' construct and does not directly access the
underlying hardware. For example, Virtual PC creates a 32-bit virtual
machine on OS/2 that thinks it is a pentium with a trio64 video card or
something. It doesn't directly access any of the underlying hardware
but instead creates 'virtual' hardware of its own using a complex
software model.
>
> The big question is whether or not the Athlon 64 line of CPUs does
> have the necessary operations in place to implement one in software.
> It does, both according to their marketing materials and their
> technical documentation.
>
> One interesting question is whether those operations could be used
> to create 64 bit VMs as well.
A 64-bit VM could be created on 32-bit hardware (and if fact, that's
what AMD did with their 'simulator' program that was available before
the real hardware was) but performance would be (and was) very slow.
>
> One can, of course, create a virtual machine on any system; the
> question is performance. If the virtual machine is able to be run
> strictly using the chip's opcodes, then performance will be good. If
> some of those operations need to be implemented in software (as, for
> example, in Bochs or even in programs like VMWare and VPC) then
> there will be a performance penalty for doing so.
>
> Being able to use the chip's native operations to create a VM is
> what one is talking about when discussing "VMs virtualised in
> hardware", as oppsed to VMs virtualised with hardware and software,
> such as is found in VMWare and VPC on the 32 bit platform.
The "virtual PC" that Connectix sold that runs on OS/2 (or VMWare) was
created entirely with software. There is no 'VM virtualised in
hardware' on the x86-64 cpus such as Opteron which uses a real machine
rather than 'virtualized' machines. I have never seen a 'VM virtualized
in hardware' and that expression makes no sense to me. If the machine
is implemented in hardware, it is a real machine and not a virtual machine.
>
> One has to understand about what is meant by the bitness of the
> machine as a whole, as opposed to the bitness presented to the
> OpSys. When you're running a 32 bit system like warp on opteron, the
> chip is using it's 32 and 15 bit registers, some parts of the chip
> that perform operations are used across both 32 bit and 64 bit mode,
Essentially all parts of the chip are used across both the legacy and
long modes (all except the extra registers). All parts of the chip are
also used across the 32-bit and 64-bit sub-modes of long mode.
> and some parts of the chip that perform operations are exclusive to
> either 64 or 32 bit mode. When you're using a 32 bit system, you're
> using the components that are relevant to 32 bit operation, while
> the components that aren't simply sit idle. In order to enforce,
> this, one loads the machine status word so that it can decide
> whether an attempt to access a particular register or operation is
> legitimate considering the mode it's in.
>
>
--
David T. Johnson
David T. Johnson
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Thu, 02 Oct 2003 11:28:25 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>> [snip]
>>
>>
>>>This is completely wrong. When you are running a 32-bit operating
>>>system on Opteron, it is running in 'legacy' mode and the *hardware* is
>>>32-bit. The CPU both appears to be 32-bit to the OS and _is_ 32-bit.
>>
>>
>> Hmmm - so you are saying that 32bit OS/2 does *not* run on a 64bit CPU
>> as when it runs 32bit code the hardware _is_ 32bit.
>
>The x86-64 cpus can run in either 'legacy' mode or 'long' mode. Long
>mode supports new 64-bit registers, addressing, and instructions. In
>'legacy' mode, the cpu is a high-performance 32-bit cpu and the other
>new features of the cpu such as the hypertransport-connected memory
>controllers are still available.
>
So you are saying that in this mode it is essentially a 32bit CPU with
better memory access.
>>
>> Or are you now going to contradict yourself? Again.
>>
>> "<laugh">
>>
>
>There is no contradiction.
Other than the one where you state that it _is_ 32bit - "<laugh>"
> The Opteron is considered to be a 64-bit
>cpu. The existence of its 'legacy' mode capability does not change that.
I agree - it's just funny to see you contradicting yourself from
paragraph to paragraph :)
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Thu, 02 Oct 2003 09:01:33 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>>
>>>David Sutherland wrote:
>>>
>>>
>>>>OS/2 cannot provide direct access to the hardware for other 32-bit
>>>>OS's (think Windows, if you are struggling to keep up) - it is plain
>>>>*stupid* to think it can do so for a 64-bit OS that didn't even exist
>>>>at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
>>>>that it can act as the bootstrap loader?
>>>
>>>I am writing this using OS/2 running on an Opteron system. <laugh>
>>
>>
>>
>> In 32bit mode - "<laugh>"
>
>Incorrect. It is running in 'legacy' mode.
>
"In 'legacy' mode, the cpu is a high-performance 32-bit cpu"
How is this different from "32bit mode"?
>>
>> I note that you did not address the real issues. I wonder why?
>>
:)
David Sutherland
<djoh...@isomedia.com> wrote:
>Jack Troughton wrote:
>> On Thu, 2 Oct 2003 18:28:25 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
>>
>>
>>>Jack Troughton wrote:
>>>
>>>>On Thu, 2 Oct 2003 16:01:33 UTC, "David T. Johnson" <djoh...@isomedia.com> wrote:
>>>>
>>>>
>>>>
>>>>>David Sutherland wrote:
>>>>>
>>>>>
>>>>>
>>>>>>OS/2 cannot provide direct access to the hardware for other 32-bit
>>>>>>OS's (think Windows, if you are struggling to keep up) - it is plain
>>>>>>*stupid* to think it can do so for a 64-bit OS that didn't even exist
>>>>>>at the time OS/2 was created. Will OS/2 even *run* on a 64bit CPU so
>>>>>>that it can act as the bootstrap loader?
>>>>>
>>>>>I am writing this using OS/2 running on an Opteron system. <laugh>
>>>>
>>>>
>>>>Can you start a 64 bit system in a VM under OS/2 on this system?
>>>>
>>>>This is the essential point in question.
>>>
>>>Well, actually, Sutherland's questions was 'will OS/2 even run on a
>>>64-bit cpu?' but I digress.
>>
>>
>> That was in the part that you snipped.
>>
>>
>>>>The answer is no.
>
>I should have said earlier that the answer to Sutherland's question
>'does OS/2 run on opteron?' was 'yes'.
>
Actually, it was meant as an aside to the real question of whether or
not you seriously believed that OS/2 would be able to host a 64-bit OS
in the equivalent of a floppy dos image, and even be able to give it
direct access to the hardware so that it can harness it's magical
DriverLess(tm) pixie dust to make a blinding games platform.
[snip]
>If the new 'gamestorm' 64-bit platform can boot independently and run on
>any Opteron hardware without the need for special drivers, it may be
>possible for it to also boot it on OS/2 in some way similar to the way
>OS/2 can boot a 'DOS image' or 'boot from drive A' and such. I admit
>that I am just wildly speculating here, though. Since there seems to be
>no detailed information available, that is all there is to do.
>DTJ
It's not like you were overly-optimisitic for a simple OS and games
bundle, or claiming that OS/2 could do something that it can't, is it?
;)
David T. Johnson
Scott Jones
> Jason Bowen wrote:
>> Not too difficult at all, no more difficult than doing it from a hard
>> drive while booting.
>
> The CD is much more difficult because it has to have the drivers for all
> of the hardware that it can expect to encounter present on the CD. The
> hard drive, otoh, can be customized for the particular hardware it is
> installed with.
Why would that be difficult? Look at Knoppix (http://knoppix.org). If
*they* can squeeze 2GiB of OS and applications onto a single CD, surely
it cannot be an impossible task to do the same with an embedded Linux
distribution w/ drivers, X, and a game that takes up less than a gig of
HD space, can it?
How about the already-mentioned Gentoo Games, which produced the
"America's Army" GameCD for Linux-32 (with, incidentally, the help of
Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
Apparently, it is not difficult.
Or is difficult=="DTJ doesn't know how to do it"?
>>>No details about "Gamestorm" are available yet although a lot of people
>>>are interested. It seems to have the capability of providing a
>>>high-quality 3D game on Linux without using either DirectX or OpenGL and
>>>appears to be able to run on any hardware. See below.
>>>
>>
>>
>> So that means somebody has to use a new SDK or learn a new API?
>
> This is unknown. My *guess* is that something very much like OpenGL is
> used so the learning curve might not be steep.
So close, yet so far. It *does* use OpenGL (and SDL). Ryan Gordon
(http://www.icculus.org), who did the original 32-bit Linux port of
"America's Army" also did the Opteron port. Once again, he used
OpenGL and SDL.
>>>When I read the Gamestorm story, my thought was that they are using the
>>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>>typically do on 32-bit systems. If so, the system would essentially run
>>>on any video hardware.
>>>
>>
>>
>> Oi, you are ignorant. The rest of your nonsense about the CPU somehow
>> giving a video card new capabilities snipped. You and Eddie are truly
>> from the same cloth.
>
> The personalities of the people here seem to be of more interest to you
> than the actual topics and discussion points.
No, you are ignorant. If the Opteron had somehow magically done away
with the need for video drivers, there would have no reason for NVIDIA
to be a part of the project.
--
Scott Jones
(scott at exti dot net)
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
>
>
>The personalities of the people here seem to be of more interest to you
>than the actual topics and discussion points.
The only thing of interest in terms of "personalities" is that you
have shown yourself to be utterly ignorant of the subject matter, but
still insist on pretending to be an expert.
What a surprise.
Jason Bowen
David T. Johnson <djoh...@isomedia.com> wrote:
>Jason Bowen wrote:
>> In article <vnoin9i...@corp.supernews.com>,
>> David T. Johnson <djoh...@isomedia.com> wrote:
>>
>>>I haven't seen another game that came with its own OS on a bootable CD
>>>although there might one somewhere.
>>
>>
>> Some Segadreamcast CD's came with WinCE as the OS. What you don't know
>> could fill a warehouse.
>
><laugh> I agree with you on your last statement. The Sega, though, was
>a gaming console, not a pc.
>
They are both computers.
>>
>>
>>>DOS games came close to that when
>>>you would boot DOS from a floppy and run the game. The problem with
>>>doing the bootable CD/OS game on newer systems, though, is support for
>>>the hardware. The OS on the CD has to boot and recognize the hardware,
>>>provide the proper drivers and software, configure itself, and then run
>>>the game and it has to do that for the wide variety of hardware that is
>>>out there. Apparently that has been sufficiently difficult to do that no
>>>one has wanted to do that.
>>>
>>
>>
>> Not too difficult at all, no more difficult than doing it from a hard
>> drive while booting.
>
>The CD is much more difficult because it has to have the drivers for all
>of the hardware that it can expect to encounter present on the CD. The
>hard drive, otoh, can be customized for the particular hardware it is
>installed with.
Because an operating system boots from a hard drive it can support ALL
hardware? I just shot a hole right through your argument. Ever heard of
a DVD?
>
>
>> There are other reasons for not doing it, such as
>> whether or not you will get a return on investment.
>
>It would be hard to lose money on a bootable CD. The CD would cost
>about $.49 to make so as long as you sold it for more than that you
>would be doing okay.
>
R & D costs money. Next...
>>
>>
>>>I don't care what the current version of DirectX is, for sure. <laugh>
>>>
>>
>> If you don't, why'd you make comments about it? <laugh>
>
>My comment was that DirectX was a dead technology if games were
>distributed along with a complete OS on a bootable CD. Of course, that
>wouldn't be true if Microsoft lets game makers sell DirectX games with
>the Windows OS on a bootable CD. <laugh>
>
Direct X is far from Dead.
>>
>>>Hard to say, but they certainly cannot be less accessible than DirectX.
>>> <laugh>
>>>
>>
>> John Carmack disagree with you. I'll stick with the opinion of the
>> creator of some of the best selling games of all time over a blowhard on
>> usenets opinion. <laugh>
>
>I have yet to see Carmack's opinions about the Gamestorm platform and
>its effect on DirectX so I don't know if he agrees or not. I have also
>not seen any opinion by Carmack about the accessibility of the DirectX
>libraries to non-Windows users but I doubt that he disagrees with me
>about that. <laugh>
>
He thinks that game writers are wise to only target Windows. That quote
is in a link I've posted already ;-) <laugh>
>>
>>>No details about "Gamestorm" are available yet although a lot of people
>>>are interested. It seems to have the capability of providing a
>>>high-quality 3D game on Linux without using either DirectX or OpenGL and
>>>appears to be able to run on any hardware. See below.
>>>
>>
>>
>> So that means somebody has to use a new SDK or learn a new API?
>
>This is unknown. My *guess* is that something very much like OpenGL is
>used so the learning curve might not be steep.
>
What you can prove is what holds merit in such a discussion.
>>
>>
>>>When I read the Gamestorm story, my thought was that they are using the
>>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>>typically do on 32-bit systems. If so, the system would essentially run
>>>on any video hardware.
>>>
>>
>>
>> Oi, you are ignorant. The rest of your nonsense about the CPU somehow
>> giving a video card new capabilities snipped. You and Eddie are truly
>> from the same cloth.
>
>The personalities of the people here seem to be of more interest to you
>than the actual topics and discussion points.
>
I interested in the fact that you seem to think you could stick any video
card with an opteron and sprinkle some magic pixie dust and voila... any
graphics operation you want can suddenly appear. You don't understand
just how loudly you show your ignornace towards software development and
hardware knowledge with comments like these.
David T. Johnson
David T. Johnson
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Sat, 04 Oct 2003 06:20:02 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>> [snip]
>>
>>
>>>
>>>The personalities of the people here seem to be of more interest to you
>>>than the actual topics and discussion points.
>>
>>
>> The only thing of interest in terms of "personalities" is that you
>> have shown yourself to be utterly ignorant of the subject matter, but
>> still insist on pretending to be an expert.
>
>I have never claimed, nor 'pretended' to be an expert. I am interested
>in the topic.
The go and read up on it instead of making stupid and ignorant claims
here.
David Sutherland
<djoh...@isomedia.com> wrote:
>Scott Jones wrote:
>> On 2003-10-04, David T. Johnson <djoh...@isomedia.com> wrote:
>>
>>>Jason Bowen wrote:
>>>
>>>>Not too difficult at all, no more difficult than doing it from a hard
>>>>drive while booting.
>>>
>>>The CD is much more difficult because it has to have the drivers for all
>>>of the hardware that it can expect to encounter present on the CD. The
>>>hard drive, otoh, can be customized for the particular hardware it is
>>>installed with.
>>
>>
>> Why would that be difficult? Look at Knoppix (http://knoppix.org). If
>> *they* can squeeze 2GiB of OS and applications onto a single CD, surely
>> it cannot be an impossible task to do the same with an embedded Linux
>> distribution w/ drivers, X, and a game that takes up less than a gig of
>> HD space, can it?
>
>I don't see space as being the problem but rather the need to have
>drivers for all of the hardware that it can expect to encounter.
Only you have claimed that it DOES NOT NEED any drivers. Which is
absurd!
>Obviously, that would be impossible for hardware that is developed after
>the CD is released and it would be difficult to be complete and
>up-to-date for all of the hardware existing at the time of its release.
> The game manufacturer would have to be constantly surveying the
>hardware to be sure they were supporting both new and old hardware when
>they released their games.
>
So why do you keep claiming that drivers aren't necessary? Make up
your mind!
>>
>> How about the already-mentioned Gentoo Games, which produced the
>> "America's Army" GameCD for Linux-32 (with, incidentally, the help of
>> Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>> Apparently, it is not difficult.
>
>Look what Gentoo says about that game on their website: "This CD
>includes the full version of the America's Army game. The game runs on
>recent Intel Pentium and AMD Athlon-based machines, and requires an
>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>
>The very restricted choice of video hardware supported by Gentoo
>supports my point and restricts the number of systems which can play the
>game.
>
Yes, it makes it clear that these tailored distributions may not
support all the hardware out there. Which was raised previously and
ignored by you.
>>
>> Or is difficult=="DTJ doesn't know how to do it"?
>
>See above.
>
LOL - you are full of shit above.
>>
>>
>>>>>No details about "Gamestorm" are available yet although a lot of people
>>>>>are interested. It seems to have the capability of providing a
>>>>>high-quality 3D game on Linux without using either DirectX or OpenGL and
>>>>>appears to be able to run on any hardware. See below.
>>>>>
>>>>
>>>>
>>>>So that means somebody has to use a new SDK or learn a new API?
>>>
>>>This is unknown. My *guess* is that something very much like OpenGL is
>>>used so the learning curve might not be steep.
>>
>>
>> So close, yet so far. It *does* use OpenGL (and SDL). Ryan Gordon
>> (http://www.icculus.org), who did the original 32-bit Linux port of
>> "America's Army" also did the Opteron port. Once again, he used
>> OpenGL and SDL.
>
>There is no mention of 64-bit x86-64 OpenGL libraries at opengl.org nor
>have I seen a reference to anyone having developed these. Please
>provide your URLs to information on the x86-64 opengl libraries here:
>
>[Scott Jone's URLs go here]
>
ROTFLMAO! You were the one claiming that the game was the brand new
creation of SCI, written using their magical API's. You were wrong,
David. Your constant attempts at deflection are lame.
>
>>
>>
>>>>>When I read the Gamestorm story, my thought was that they are using the
>>>>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>>>>typically do on 32-bit systems. If so, the system would essentially run
>>>>>on any video hardware.
>>>>>
>>>>
>>>>
>>>>Oi, you are ignorant. The rest of your nonsense about the CPU somehow
>>>>giving a video card new capabilities snipped. You and Eddie are truly
>>>>from the same cloth.
>>>
>>>The personalities of the people here seem to be of more interest to you
>>>than the actual topics and discussion points.
>>
>>
>> No, you are ignorant. If the Opteron had somehow magically done away
>> with the need for video drivers, there would have no reason for NVIDIA
>> to be a part of the project.
>
>I don't see why not. Exactly the opposite would seem to be the case
>IMO. Nvidia could provide valuable assistance in developing a software
>graphics implementation.
Yeah sure, Nvidia are going to be really interested in developing a
software solution that puts their entire industry out of work! Damn
but you're dim!
Or tell you what - why don't you try proving it instead of making yet
more moronic speculations:
[David T Johnsons URLs go here]
L.O.L.
Scott Jones
>> it cannot be an impossible task to do the same with an embedded Linux
>> distribution w/ drivers, X, and a game that takes up less than a gig of
>> HD space, can it?
>
> I don't see space as being the problem but rather the need to have
> drivers for all of the hardware that it can expect to encounter.
Really? I thought you said that the Opteron magically did away with
the need for hardware drivers. Which is it, David? You can't have
it both ways.
>> How about the already-mentioned Gentoo Games, which produced the
>> "America's Army" GameCD for Linux-32 (with, incidentally, the help of
>> Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>> Apparently, it is not difficult.
>
> Look what Gentoo says about that game on their website: "This CD
> includes the full version of the America's Army game. The game runs on
> recent Intel Pentium and AMD Athlon-based machines, and requires an
> NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>
> The very restricted choice of video hardware supported by Gentoo
> supports my point and restricts the number of systems which can play the
> game.
The restricted choice of video hardware is due to the limited number of
vendors with Linux drivers that support 3D acceleration in hardware.
It's a short list: ATI and nVidia. No serious gamer in their right
mind is going to slow down their machine by using software rendering.
>> Or is difficult=="DTJ doesn't know how to do it"?
>
> See above.
Was there an answer up there? All I saw was a regurgitated quote
from a website that neither supports nor denies your claim as to the
supposed difficulty of the task.
>>>This is unknown. My *guess* is that something very much like OpenGL is
>>>used so the learning curve might not be steep.
>>
>>
>> So close, yet so far. It *does* use OpenGL (and SDL). Ryan Gordon
>> (http://www.icculus.org), who did the original 32-bit Linux port of
>> "America's Army" also did the Opteron port. Once again, he used
>> OpenGL and SDL.
>
> There is no mention of 64-bit x86-64 OpenGL libraries at opengl.org nor
> have I seen a reference to anyone having developed these. Please
> provide your URLs to information on the x86-64 opengl libraries here:
>
> [Scott Jone's URLs go here]
That's Jones'. One is unlikely to find much mention of OpenGL for
x86-64 at opengl.org, as they are a standards body, not a vendor. It
is up to the hard- and software vendors to provide the implementation.
Meanwhile, look at nVidia's drivers for XFree86, which include OpenGL
libraries:<ftp://download.nvidia.com/XFree86/Linux-x86-64/1.0-4499>.
Or Mesa, which implements the OpenGL API <http://mesa3d.org>, and
compiles on an x86-64 system. Doesn't look like SuSE had any problems
doing so for their upcoming x86-64 distribution, or in their previous
beta of SuSE 8.2 for x86-64.
>> No, you are ignorant. If the Opteron had somehow magically done away
>> with the need for video drivers, there would have no reason for NVIDIA
>> to be a part of the project.
>
> I don't see why not. Exactly the opposite would seem to be the case
> IMO. Nvidia could provide valuable assistance in developing a software
> graphics implementation.
So they would willing cut their own throats by simultaneously making
their hardware irrelevant and assisting their competitors? Perhaps.
And if my aunt had balls, she'd be my uncle.
David T. Johnson
> On Sat, 04 Oct 2003 21:00:01 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
>
>
>>Scott Jones wrote:
>>
>>>On 2003-10-04, David T. Johnson <djoh...@isomedia.com> wrote:
>>>
>>>
>>>>Jason Bowen wrote:
>>>>
>>>>
>>>>>Not too difficult at all, no more difficult than doing it from a hard
>>>>>drive while booting.
>>>>
>>>>The CD is much more difficult because it has to have the drivers for all
>>>>of the hardware that it can expect to encounter present on the CD. The
>>>>hard drive, otoh, can be customized for the particular hardware it is
>>>>installed with.
>>>
>>>
>>>Why would that be difficult? Look at Knoppix (http://knoppix.org). If
>>>*they* can squeeze 2GiB of OS and applications onto a single CD, surely
>>>it cannot be an impossible task to do the same with an embedded Linux
>>>distribution w/ drivers, X, and a game that takes up less than a gig of
>>>HD space, can it?
>>
>>I don't see space as being the problem but rather the need to have
>>drivers for all of the hardware that it can expect to encounter.
>
>
> Only you have claimed that it DOES NOT NEED any drivers. Which is
> absurd!
Incorrect. Software to control hardware (i.e. drivers) will always be
needed. However, this software does not seem to be hardware-specific on
the Gamestorm/Opteron platform. As I mentioned, though, the details are
not yet available.
>
>
>>Obviously, that would be impossible for hardware that is developed after
>>the CD is released and it would be difficult to be complete and
>>up-to-date for all of the hardware existing at the time of its release.
>> The game manufacturer would have to be constantly surveying the
>>hardware to be sure they were supporting both new and old hardware when
>>they released their games.
>>
>
>
> So why do you keep claiming that drivers aren't necessary? Make up
> your mind!
Software to control hardware (i.e. drivers) will always be needed.
However, this software does not seem to be hardware-specific on the
Gamestorm/Opteron platform. As I mentioned, though, the details are not
yet available.
>
>
>>>How about the already-mentioned Gentoo Games, which produced the
>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>Apparently, it is not difficult.
>>
>>Look what Gentoo says about that game on their website: "This CD
>>includes the full version of the America's Army game. The game runs on
>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>
>>The very restricted choice of video hardware supported by Gentoo
>>supports my point and restricts the number of systems which can play the
>>game.
>>
>
>
> Yes, it makes it clear that these tailored distributions may not
> support all the hardware out there. Which was raised previously and
> ignored by you.
The Gentoo game distribution clearly specifies a very limited set of
hardware that it will run on. The Gamestorm/Opteron announcement does
not, hence my comment.
>
>
>>>Or is difficult=="DTJ doesn't know how to do it"?
>>
>>See above.
>>
>
>
[personal comment deleted]
A new software-video solution would not put Nvidia out of work. Nvidia
always seems to be at the forefront of new technology. They are not the
kind of company that would attempt to hang on to the status quo to
preserve their short-term revenues. The only constant in the computer
industry is constant change.
>
> Or tell you what - why don't you try proving it instead of making yet
> more moronic speculations:
There are no details that have been released, yet, on the Gamestorm
platform for the Opteron but I think it is very interesting.
>
> [David T Johnsons URLs go here]
>
--
David T. Johnson
David Sutherland
<djoh...@isomedia.com> wrote:
So you both do and do not need drivers for the hardware? Sure thing
David, you just keep on wriggling around and avoid having to confront
your mistakes - it's not like it isn't obvious to everyone! ROTFL!
>
>>
>>
>>>Obviously, that would be impossible for hardware that is developed after
>>>the CD is released and it would be difficult to be complete and
>>>up-to-date for all of the hardware existing at the time of its release.
>>> The game manufacturer would have to be constantly surveying the
>>>hardware to be sure they were supporting both new and old hardware when
>>>they released their games.
>>>
>>
>>
>> So why do you keep claiming that drivers aren't necessary? Make up
>> your mind!
>
>Software to control hardware (i.e. drivers) will always be needed.
>However, this software does not seem to be hardware-specific on the
>Gamestorm/Opteron platform. As I mentioned, though, the details are not
>yet available.
>
Yawn. You're an idiot if you think that drivers aren't hardware
specific.
>>
>>
>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>Apparently, it is not difficult.
>>>
>>>Look what Gentoo says about that game on their website: "This CD
>>>includes the full version of the America's Army game. The game runs on
>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>
>>>The very restricted choice of video hardware supported by Gentoo
>>>supports my point and restricts the number of systems which can play the
>>>game.
>>>
>>
>>
>> Yes, it makes it clear that these tailored distributions may not
>> support all the hardware out there. Which was raised previously and
>> ignored by you.
>
>The Gentoo game distribution clearly specifies a very limited set of
>hardware that it will run on. The Gamestorm/Opteron announcement does
>not, hence my comment.
>
The Gamestorm press release said very little indeed. They did not, for
example, specify that they had failed to create world peace. By your
logic that means that all those nasty problems in the Middle East are
now solved. Oh good.
>>
>>
>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>
>>>See above.
>>>
>>
>>
>[personal comment deleted]
You're an idiot, and you are full of shit. It's not personal David,
it's a statement of fact.
>>
>>
>>>>
>>>>>>>No details about "Gamestorm" are available yet although a lot of people
>>>>>>>are interested. It seems to have the capability of providing a
>>>>>>>high-quality 3D game on Linux without using either DirectX or OpenGL and
>>>>>>>appears to be able to run on any hardware. See below.
>>>>>>>
>>>>>>
>>>>>>
>>>>>>So that means somebody has to use a new SDK or learn a new API?
>>>>>
>>>>>This is unknown. My *guess* is that something very much like OpenGL is
>>>>>used so the learning curve might not be steep.
>>>>
>>>>
>>>>So close, yet so far. It *does* use OpenGL (and SDL). Ryan Gordon
>>>>(http://www.icculus.org), who did the original 32-bit Linux port of
>>>>"America's Army" also did the Opteron port. Once again, he used
>>>>OpenGL and SDL.
>>>
>>>There is no mention of 64-bit x86-64 OpenGL libraries at opengl.org nor
>>>have I seen a reference to anyone having developed these. Please
>>>provide your URLs to information on the x86-64 opengl libraries here:
>>>
>>>[Scott Jone's URLs go here]
>>>
>>
>>
>> ROTFLMAO! You were the one claiming that the game was the brand new
>> creation of SCI, written using their magical API's. You were wrong,
>> David. Your constant attempts at deflection are lame.
>
>When I read the Gamestorm story, my thought was that they are using the
>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>typically do on 32-bit systems. If so, the system would essentially run
>on any video hardware.
>
You thought wrong. They still need 3D boards to do the graphics
grunt work.
>
>>
>>
>>>>
>>>>>>>When I read the Gamestorm story, my thought was that they are using the
>>>>>>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>>>>>>typically do on 32-bit systems. If so, the system would essentially run
>>>>>>>on any video hardware.
>>>>>>>
>>>>>>
>>>>>>
>>>>>>Oi, you are ignorant. The rest of your nonsense about the CPU somehow
>>>>>>giving a video card new capabilities snipped. You and Eddie are truly
>>>>>
>>>>>>from the same cloth.
>>>>>
>>>>>The personalities of the people here seem to be of more interest to you
>>>>>than the actual topics and discussion points.
>>>>
>>>>
>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>to be a part of the project.
>>>
>>>I don't see why not. Exactly the opposite would seem to be the case
>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>graphics implementation.
>>
>>
>>
>> Yeah sure, Nvidia are going to be really interested in developing a
>> software solution that puts their entire industry out of work! Damn
>> but you're dim!
>
>A new software-video solution would not put Nvidia out of work. Nvidia
>always seems to be at the forefront of new technology. They are not the
>kind of company that would attempt to hang on to the status quo to
>preserve their short-term revenues. The only constant in the computer
>industry is constant change.
LOL - Nvidia are a *hardware* manufacturer. Their expertise is in
making 3D chipsets. A generic device will *never* be able to provide
the same kind of performance as custom silicon, and they would DIE if
they did what you are claiming. If you believe otherwise it's just
another example of how ignorant you are about How Things Work.
>
>
>>
>> Or tell you what - why don't you try proving it instead of making yet
>> more moronic speculations:
>
>There are no details that have been released, yet, on the Gamestorm
>platform for the Opteron but I think it is very interesting.
>
I think it's very interesting that without any details at all you
create entire fantasy worlds where the laws of physics do not apply,
and then embarrass yourself by pretending they are real in this ng.
What is that about?
>>
>> [David T Johnsons URLs go here]
>>
Where's your evidence, David?
And hey, why aren't you taking me up on that bet? LOL!
David Sutherland
<djoh...@isomedia.com> wrote:
>Scott Jones wrote:
>> On 2003-10-05, David T. Johnson <djoh...@isomedia.com> wrote:
>>
>>>>it cannot be an impossible task to do the same with an embedded Linux
>>>>distribution w/ drivers, X, and a game that takes up less than a gig of
>>>>HD space, can it?
>>>
>>>I don't see space as being the problem but rather the need to have
>>>drivers for all of the hardware that it can expect to encounter.
>>
>>
>> Really? I thought you said that the Opteron magically did away with
>> the need for hardware drivers. Which is it, David? You can't have
>> it both ways.
>
>When I read the Gamestorm story, my thought was that they are using the
>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>typically do on 32-bit systems. If so, the system would essentially run
>on any video hardware. Software to control hardware (i.e. drivers) will
>always be needed. However, this software does not seem to be
>hardware-specific on the Gamestorm/Opteron platform. As I mentioned,
>though, the details are not yet available.
>
You are cutting and pasting this over and over - and it has already
been refuted. All you are doing is making it painfully obvious how
bankrupt your claims are.
>>
>>
>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>Apparently, it is not difficult.
>>>
>>>Look what Gentoo says about that game on their website: "This CD
>>>includes the full version of the America's Army game. The game runs on
>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>
>>>The very restricted choice of video hardware supported by Gentoo
>>>supports my point and restricts the number of systems which can play the
>>>game.
>>
>>
>> The restricted choice of video hardware is due to the limited number of
>> vendors with Linux drivers that support 3D acceleration in hardware.
>> It's a short list: ATI and nVidia. No serious gamer in their right
>> mind is going to slow down their machine by using software rendering.
>
>I agree that this is true with typical 32-bit systems. The
>Gamestorm/Opteron platform, however, has apparently found a way to get
>acceptable performance without using hardware-specific drivers.
>
Where does it say that? Why were Nvidia involved if there are no
hardware or driver dependencies?
>
>
>>
>>
>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>
>>>See above.
>>
>>
>> Was there an answer up there? All I saw was a regurgitated quote
>> from a website that neither supports nor denies your claim as to the
>> supposed difficulty of the task.
>
>The very restricted choice of video hardware supported by Gentoo
>supports my point and restricts the number of systems which can play the
>game.
>
It doesn't support your point at all. Nvidia's presence in that
press release completely vaporises your "point".
They aren't making any claims except that it's been recompiled and
bundled with Linux. Coincidentally, this has happened at the same
time as the release of the Athlon64. Joint marketing.
It's *you* that's making all the grandiose claims based upon nothing
but a vague press release. If you have any evidence to the contrary
then *present it*
>Also, the absence
>of any specific video hardware requirements for the game seems
>misleading at best.
Misleading to you maybe, but nobody else seems confused by it.
> It's interesting that you mention Mesa, since the
>standard Mesa release supports only software rendering which is what I
>have been suggesting might be the case for the Gamestorm thing.
>
LOL - wow, that thing will be a real dog!
>>
>>
>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>to be a part of the project.
>>>
>>>I don't see why not. Exactly the opposite would seem to be the case
>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>graphics implementation.
>>
>>
>> So they would willing cut their own throats by simultaneously making
>> their hardware irrelevant and assisting their competitors? Perhaps.
>
>No one can stay with the same approach forever. Things change very
>quickly in the computer business. The whole idea of having a dedicated
>video hardware with its own memory sticking out on the PCI/AGP bus
>doesn't seem like a long-term thing to me for video.
Really? So you think that a general purpose CPU can do things just as
fast as dedicated silicon? ROTFL! Damn you're funny!
>The Opteron CPU is
>much more powerful than the video chip
Duh! The video chip wipes the floor with the Opteron for 3D
rendering. There is no comparison. Otherwise, software rendering
would be just as fast as hardware rendering!
id don't even include software renderers in their games since QuakeII
because a general purpose CPU can't keep up. Are you saying that
they are going to start again? ROTFLMAO!
>and the Opteron system memory is
>much faster and has much greater bandwidth than the video memory.
DUH!
9800Pro = 21.8GB/s
Athlon64 = 6.4GB/sec
Don't you ever get bored of parading your stupidity?
> Seems
>like software rendering could ultimately provide better performance when
>the software is developed, even if Gamestorm has not done that.
Seems like you don't know your ass from your elbow.
David Sutherland
<djoh...@isomedia.com> wrote:
>Scott Jones wrote:
>> On 2003-10-05, David T. Johnson <djoh...@isomedia.com> wrote:
>>
>>>>it cannot be an impossible task to do the same with an embedded Linux
>>>>distribution w/ drivers, X, and a game that takes up less than a gig of
>>>>HD space, can it?
>>>
>>>I don't see space as being the problem but rather the need to have
>>>drivers for all of the hardware that it can expect to encounter.
>>
>>
>> Really? I thought you said that the Opteron magically did away with
>> the need for hardware drivers. Which is it, David? You can't have
>> it both ways.
>
>When I read the Gamestorm story, my thought was that they are using the
>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>typically do on 32-bit systems.
One other question:
If the Opteron really does give such an *incredible* boost to 3D
rendering performance, how come banchmarks show it often rendering 3D
scenes *slower* than the 32bit Athlon?
http://www.tomshardware.com/cpu/20030422/opteron-23.html
http://www.extremetech.com/article2/0,3973,1217390,00.asp
More magic?
Marty
I had been trying to stay out of this, but this is just so painful, I
had to comment.
A Super Nintendo used a 3MHz custom-tailored graphics processor. You
can almost time it on a stopwatch. To perform the way this little dinky
processor performed required a 233MHz x86 CPU (running at 80% capacity
for graphics rendering, 20% for emulation of CPU opcodes). So let's say
186MHz generic x86 processor for graphics alone. That's 62x the
processing speed required by dedicated silicon.
I'll repeat that: 62x the processing speed. And that's for a graphics
chip thats most complex accelerated operation is rendering wedge-shaped
transparencies and rotating and scaling *2D* images. For more complex
operations like 3D and fog effects, the gap widens much faster between
the speed of dedicated silicon and power required from generic CPUs.
I say all that simply to add to the chorus that Johnson is an uninformed
blowhard, and general just a freaking idiot. :-)
David T. Johnson
Hmmm. I made the same comment in a post a couple of days ago.
>
>
>>>
>>>>Obviously, that would be impossible for hardware that is developed after
>>>>the CD is released and it would be difficult to be complete and
>>>>up-to-date for all of the hardware existing at the time of its release.
>>>>The game manufacturer would have to be constantly surveying the
>>>>hardware to be sure they were supporting both new and old hardware when
>>>>they released their games.
>>>>
>>>
>>>
>>>So why do you keep claiming that drivers aren't necessary? Make up
>>>your mind!
>>
>>Software to control hardware (i.e. drivers) will always be needed.
>>However, this software does not seem to be hardware-specific on the
>>Gamestorm/Opteron platform. As I mentioned, though, the details are not
>>yet available.
>>
>
>
> Yawn. You're an idiot if you think that drivers aren't hardware
> specific.
You are an idiot if you think that hardware drivers are always
hardware-specific. I'll spell it for you: S-T-A-N-D-A-R-D-S.
>
>
>
>>>
>>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>>Apparently, it is not difficult.
>>>>
>>>>Look what Gentoo says about that game on their website: "This CD
>>>>includes the full version of the America's Army game. The game runs on
>>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>>
>>>>The very restricted choice of video hardware supported by Gentoo
>>>>supports my point and restricts the number of systems which can play the
>>>>game.
>>>>
>>>
>>>
>>>Yes, it makes it clear that these tailored distributions may not
>>>support all the hardware out there. Which was raised previously and
>>>ignored by you.
>>
>>The Gentoo game distribution clearly specifies a very limited set of
>>hardware that it will run on. The Gamestorm/Opteron announcement does
>>not, hence my comment.
>>
>
>
> The Gamestorm press release said very little indeed. They did not, for
> example, specify that they had failed to create world peace. By your
> logic that means that all those nasty problems in the Middle East are
> now solved. Oh good.
Now you have strayed off-topic and are talking nonsense.
>
>
>>>
>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>
>>>>See above.
>>>>
>>>
>>>
>>[personal comment deleted]
>
>
[personal comment snipped]
>
>
>>>
>>>>>>>>No details about "Gamestorm" are available yet although a lot of people
>>>>>>>>are interested. It seems to have the capability of providing a
>>>>>>>>high-quality 3D game on Linux without using either DirectX or OpenGL and
>>>>>>>>appears to be able to run on any hardware. See below.
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>So that means somebody has to use a new SDK or learn a new API?
>>>>>>
>>>>>>This is unknown. My *guess* is that something very much like OpenGL is
>>>>>>used so the learning curve might not be steep.
>>>>>
>>>>>
>>>>>So close, yet so far. It *does* use OpenGL (and SDL). Ryan Gordon
>>>>>(http://www.icculus.org), who did the original 32-bit Linux port of
>>>>>"America's Army" also did the Opteron port. Once again, he used
>>>>>OpenGL and SDL.
>>>>
>>>>There is no mention of 64-bit x86-64 OpenGL libraries at opengl.org nor
>>>>have I seen a reference to anyone having developed these. Please
>>>>provide your URLs to information on the x86-64 opengl libraries here:
>>>>
>>>>[Scott Jone's URLs go here]
>>>>
>>>
>>>
>>>ROTFLMAO! You were the one claiming that the game was the brand new
>>>creation of SCI, written using their magical API's. You were wrong,
>>>David. Your constant attempts at deflection are lame.
>>
>>When I read the Gamestorm story, my thought was that they are using the
>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>typically do on 32-bit systems. If so, the system would essentially run
>>on any video hardware.
>>
>
>
> You thought wrong. They still need 3D boards to do the graphics
> grunt work.
You mistakenly seem to think that 'hardware' is always required.
yeah, right, and people will never need more than 640 kilobytes of
memory.
> and they would DIE if
> they did what you are claiming. If you believe otherwise it's just
> another example of how ignorant you are about How Things Work.
>
>
>>
>>>Or tell you what - why don't you try proving it instead of making yet
>>>more moronic speculations:
>>
>>There are no details that have been released, yet, on the Gamestorm
>>platform for the Opteron but I think it is very interesting.
>>
>
>
> I think it's very interesting that without any details at all you
> create entire fantasy worlds where the laws of physics do not apply,
No laws of physics are violated by software rendering. <laugh>
> and then embarrass yourself by pretending they are real in this ng.
> What is that about?
You are the guy posting about violations of the 'laws of physics.'
You're sounding unusually nutty today.
>
>
>>>[David T Johnsons URLs go here]
>>>
>
>
> Where's your evidence, David?
There are no details that have been released, yet, on the Gamestorm
platform for the Opteron but I think it is very interesting.
>
> And hey, why aren't you taking me up on that bet? LOL!
>
There is no reason to wager with you. You would probably be a piker,
anyway.
David T. Johnson
> On Sun, 05 Oct 2003 22:16:28 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
>
>
>>Scott Jones wrote:
>>
>>>On 2003-10-05, David T. Johnson <djoh...@isomedia.com> wrote:
>>>
>>>
>>>>>it cannot be an impossible task to do the same with an embedded Linux
>>>>>distribution w/ drivers, X, and a game that takes up less than a gig of
>>>>>HD space, can it?
>>>>
>>>>I don't see space as being the problem but rather the need to have
>>>>drivers for all of the hardware that it can expect to encounter.
>>>
>>>
>>>Really? I thought you said that the Opteron magically did away with
>>>the need for hardware drivers. Which is it, David? You can't have
>>>it both ways.
>>
>>When I read the Gamestorm story, my thought was that they are using the
>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>typically do on 32-bit systems. If so, the system would essentially run
>>on any video hardware. Software to control hardware (i.e. drivers) will
>>always be needed. However, this software does not seem to be
>>hardware-specific on the Gamestorm/Opteron platform. As I mentioned,
>>though, the details are not yet available.
>>
>
>
> You are cutting and pasting this over and over - and it has already
> been refuted.
Your comments hardly qualify as a refutation. <laugh>
> All you are doing is making it painfully obvious how
> bankrupt your claims are.
Painful for you, perhaps. <laugh>
>
>
>>>
>>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>>Apparently, it is not difficult.
>>>>
>>>>Look what Gentoo says about that game on their website: "This CD
>>>>includes the full version of the America's Army game. The game runs on
>>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>>
>>>>The very restricted choice of video hardware supported by Gentoo
>>>>supports my point and restricts the number of systems which can play the
>>>>game.
>>>
>>>
>>>The restricted choice of video hardware is due to the limited number of
>>>vendors with Linux drivers that support 3D acceleration in hardware.
>>>It's a short list: ATI and nVidia. No serious gamer in their right
>>>mind is going to slow down their machine by using software rendering.
>>
>>I agree that this is true with typical 32-bit systems. The
>>Gamestorm/Opteron platform, however, has apparently found a way to get
>>acceptable performance without using hardware-specific drivers.
>>
>
>
> Where does it say that? Why were Nvidia involved if there are no
> hardware or driver dependencies?
What do you think Nvidia does, anyway. Most of their work is
software-related. Even their 'hardware' is mostly a software effort.
>
>
>>
>>>
>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>
>>>>See above.
>>>
>>>
>>>Was there an answer up there? All I saw was a regurgitated quote
>>>from a website that neither supports nor denies your claim as to the
>>>supposed difficulty of the task.
>>
>>The very restricted choice of video hardware supported by Gentoo
>>supports my point and restricts the number of systems which can play the
>>game.
>>
>
>
> It doesn't support your point at all. Nvidia's presence in that
> press release completely vaporises your "point".
No, it doesn't. Rather, it supports it because Nvidia is obviously
involved and yet no specific hardware requirement for Nvidia hardware
was listed.
They don't seem to be using the 'accelerators' in the video hardware
because they don't list any particular hardware requirements and they
suggest that it will boot on any Opteron computer: "They can run the
game on a floor model at Best Buy, and leave the hard drive untouched,"
he says."
Other claims by SCI:
"cuts out legacy operating system overhead..."
"without the usual overhead from the legacy operating systems
raditionally used in the gaming industry, SCI claims."
"With the AMD Athlon 64 processor and GameStorm technology, AMD is able
to showcase a fully-integrated 64-bit environment that delivers
performance and realism to the most demanding gamers,"
"the company is now working on porting several more 32-bit and 64-bit
games to the GameStorm platform"
The last sentence is revealing because it suggests that 'porting' is
required to run other games on the Gamestorm platform which suggests
that Gamestorm does not run OpenGL games.
>
> It's *you* that's making all the grandiose claims based upon nothing
> but a vague press release. If you have any evidence to the contrary
> then *present it*
I am speculating, rather than making claims. There is no information
about the Gamestorm platform yet. You have been told that previously.
>
>
>>Also, the absence
>>of any specific video hardware requirements for the game seems
>>misleading at best.
>
>
> Misleading to you maybe, but nobody else seems confused by it.
It would be misleading to anyone who intended to use the game who lacked
the specific hardware that you allege is required.
>
>
>>It's interesting that you mention Mesa, since the
>>standard Mesa release supports only software rendering which is what I
>>have been suggesting might be the case for the Gamestorm thing.
>>
>
>
> LOL - wow, that thing will be a real dog!
Time will tell us.
>
>
>>>
>>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>>to be a part of the project.
>>>>
>>>>I don't see why not. Exactly the opposite would seem to be the case
>>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>>graphics implementation.
>>>
>>>
>>>So they would willing cut their own throats by simultaneously making
>>>their hardware irrelevant and assisting their competitors? Perhaps.
>>
>>No one can stay with the same approach forever. Things change very
>>quickly in the computer business. The whole idea of having a dedicated
>>video hardware with its own memory sticking out on the PCI/AGP bus
>>doesn't seem like a long-term thing to me for video.
>
>
> Really? So you think that a general purpose CPU can do things just as
> fast as dedicated silicon? ROTFL! Damn you're funny!
Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
than the video card processor. Speed, however, is affected by the
design, overhead from other activities, and the software. Gamestorm
makes no claim to be 'as fast as' a 32-bit platform or to be the
'fastest' or anything like that. They only suggest that their
performance is competitive with other game experiences.
>
>
>>The Opteron CPU is
>>much more powerful than the video chip
>
>
> Duh! The video chip wipes the floor with the Opteron for 3D
> rendering.
Reference?
> There is no comparison. Otherwise, software rendering
> would be just as fast as hardware rendering!
The ultimate speed is affected by a lot of things other than the video
card chip.
>
> id don't even include software renderers in their games since QuakeII
> because a general purpose CPU can't keep up. Are you saying that
> they are going to start again? ROTFLMAO!
We'll have to see what Gamestorm is doing.
>
>
>
>>and the Opteron system memory is
>>much faster and has much greater bandwidth than the video memory.
>
>
> DUH!
>
> 9800Pro = 21.8GB/s
> Athlon64 = 6.4GB/sec
The Opteron has the memory controller integrated with the Opteron
processor which means that the front side bus (interface to memory) runs
at the speed of the processor and has a front side bandwidth of 14.4
GB/sec.
The more significant thing, however, is that that the Opteron system
will typically have 1GB or more of memory (and a 1MB cache) while a
high-performance video card will only have perhaps 128 MB of memory. It
seems possible that a system with such a large high-speed memory could
therefore render more images in memory and maintain an acceptabe
framerate by working ahead. The main limit for such a hypothetical
system doing software rendering would probably be the speed of the AGP
bus to transfer the rendered images to the video memory. Gamestorm
hasn't claimed that their system is faster than all others but only has
suggested that its performance is competitive. It seems possible that
they could be doing some sort of software rendering although no
information has been provided.
>
[personal comment snipped]
>
>
>> Seems
>>like software rendering could ultimately provide better performance when
>>the software is developed, even if Gamestorm has not done that.
>
>
[personal comment snipped]
>
>
> Regards,
> David Sutherland
> (note **ANTI-SPAM** in reply field)
--
David T. Johnson
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Sun, 05 Oct 2003 22:16:28 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>>
>>>Scott Jones wrote:
>>>
>>>>On 2003-10-05, David T. Johnson <djoh...@isomedia.com> wrote:
>>>>
>>>>
>>>>>>it cannot be an impossible task to do the same with an embedded Linux
>>>>>>distribution w/ drivers, X, and a game that takes up less than a gig of
>>>>>>HD space, can it?
>>>>>
>>>>>I don't see space as being the problem but rather the need to have
>>>>>drivers for all of the hardware that it can expect to encounter.
>>>>
>>>>
>>>>Really? I thought you said that the Opteron magically did away with
>>>>the need for hardware drivers. Which is it, David? You can't have
>>>>it both ways.
>>>
>>>When I read the Gamestorm story, my thought was that they are using the
>>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>>typically do on 32-bit systems.
>>
>>
>> One other question:
>>
>> If the Opteron really does give such an *incredible* boost to 3D
>> rendering performance,
>
>No claim has been made that the Opteron gives an "incredible boost" to
>3D rendering performance.
That is a lie. You have claimed that the 3D gaming performance of the
Opteron - with no additional hardware - is going to kill DirectX.
This is patently stupid, but it *is* your claim.
You *did* create this very thread entitled "Death blow for Windows
DirectX", after all.
>The benchmarks you you provide below, however,
>are for the Opteron running 32-bit software rather than 64-bit software.
> The Gamestorm platform is apparently 64-bit.
>
Given that existing CPU's are most certainly not capable of rendering
3D with anything like the same speed or quality of hardware boards,
how else do you explain your claim that such hardware is now obsolete?
Can you provide *any* evidence that the speed of 64bit render software
is significantly (as in more than twice) faster than that of 32bit?
There are certainly plenty of 64bit benchmarks which show that you can
indeed get up to almost twice the throughput over 32bits for *some*
types of processing. But this is worlds away from the 20-30 fold
increase you would need in order to render complex scenes in realtime
for a game!
>> how come banchmarks show it often rendering 3D
>> scenes *slower* than the 32bit Athlon?
>>
>> http://www.tomshardware.com/cpu/20030422/opteron-23.html
>>
>> http://www.extremetech.com/article2/0,3973,1217390,00.asp
>>
>> More magic?
LOL - you can't answer, can you?
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Sun, 05 Oct 2003 22:16:28 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>>
>>>Scott Jones wrote:
>>>
>>>>On 2003-10-05, David T. Johnson <djoh...@isomedia.com> wrote:
>>>>
>>>>
>>>>>>it cannot be an impossible task to do the same with an embedded Linux
>>>>>>distribution w/ drivers, X, and a game that takes up less than a gig of
>>>>>>HD space, can it?
>>>>>
>>>>>I don't see space as being the problem but rather the need to have
>>>>>drivers for all of the hardware that it can expect to encounter.
>>>>
>>>>
>>>>Really? I thought you said that the Opteron magically did away with
>>>>the need for hardware drivers. Which is it, David? You can't have
>>>>it both ways.
>>>
>>>When I read the Gamestorm story, my thought was that they are using the
>>>64-bit Opteron CPU to do stuff that the video hardware would otherwise
>>>typically do on 32-bit systems. If so, the system would essentially run
>>>on any video hardware. Software to control hardware (i.e. drivers) will
>>>always be needed. However, this software does not seem to be
>>>hardware-specific on the Gamestorm/Opteron platform. As I mentioned,
>>>though, the details are not yet available.
>>>
>>
>>
>> You are cutting and pasting this over and over - and it has already
>> been refuted.
>
>Your comments hardly qualify as a refutation. <laugh>
>
People think you're an idiot. "<laugh>"
You just keep on proving them right.
>> All you are doing is making it painfully obvious how
>> bankrupt your claims are.
>
>Painful for you, perhaps. <laugh>
>
Do you have any idea how stupid you have made yourself look with this
latest thread? I'm embarrassed for you.
>>
>>
>>>>
>>>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>>>Apparently, it is not difficult.
>>>>>
>>>>>Look what Gentoo says about that game on their website: "This CD
>>>>>includes the full version of the America's Army game. The game runs on
>>>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>>>
>>>>>The very restricted choice of video hardware supported by Gentoo
>>>>>supports my point and restricts the number of systems which can play the
>>>>>game.
>>>>
>>>>
>>>>The restricted choice of video hardware is due to the limited number of
>>>>vendors with Linux drivers that support 3D acceleration in hardware.
>>>>It's a short list: ATI and nVidia. No serious gamer in their right
>>>>mind is going to slow down their machine by using software rendering.
>>>
>>>I agree that this is true with typical 32-bit systems. The
>>>Gamestorm/Opteron platform, however, has apparently found a way to get
>>>acceptable performance without using hardware-specific drivers.
>>>
>>
>>
>> Where does it say that? Why were Nvidia involved if there are no
>> hardware or driver dependencies?
>
>What do you think Nvidia does, anyway. Most of their work is
>software-related. Even their 'hardware' is mostly a software effort.
>
You are joking, right? ROTFLMAO@johnson.
Sure David, the silicon has got nothing to do with it, and all along
ATI and Nvidia have just been pulling everyone's leg.
You are utterly ridiculous. You are dumber than Tim Martin, and I
didn't think that was possible.
>>
>>
>>>
>>>>
>>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>>
>>>>>See above.
>>>>
>>>>
>>>>Was there an answer up there? All I saw was a regurgitated quote
>>>>from a website that neither supports nor denies your claim as to the
>>>>supposed difficulty of the task.
>>>
>>>The very restricted choice of video hardware supported by Gentoo
>>>supports my point and restricts the number of systems which can play the
>>>game.
>>>
>>
>>
>> It doesn't support your point at all. Nvidia's presence in that
>> press release completely vaporises your "point".
>
>No, it doesn't. Rather, it supports it because Nvidia is obviously
>involved and yet no specific hardware requirement for Nvidia hardware
>was listed.
>
Ah, you're back to claiming that the *absence* of information makes
all of your stupid claims true. They *also* didn't say that David T
Johnson is a complete asshole, with the technical knowledge of a newt
which *obviously* means that it must be true ;)
The fact of Nvidias prominent mention in the press release should just
be ignored because otherwise you end up looking like a complete fool.
Tough luck David - it's already the case.
You look like a complete jerk.
Why do Nvidia feature in the press release? AMD managed to make their
Opteron perform as fast as dedicated silicon and yet this fact has
*not* been released to the press? Sure David - your fantasy world is
splitting at the seams with inconsistencies.
>"They can run the
>game on a floor model at Best Buy, and leave the hard drive untouched,"
>he says."
>
All the PC's on the floor at Best Buy have video cards in them.
>Other claims by SCI:
>
>"cuts out legacy operating system overhead..."
>"without the usual overhead from the legacy operating systems
>raditionally used in the gaming industry, SCI claims."
>"With the AMD Athlon 64 processor and GameStorm technology, AMD is able
>to showcase a fully-integrated 64-bit environment that delivers
>performance and realism to the most demanding gamers,"
>"the company is now working on porting several more 32-bit and 64-bit
>games to the GameStorm platform"
>
I've already refuted these.
>The last sentence is revealing because it suggests that 'porting' is
>required to run other games on the Gamestorm platform which suggests
>that Gamestorm does not run OpenGL games.
>
Bullshit. Americas Army is an OPenGL game as you have already been
told.
That you ignore the evidence is *your* problem, not mine.
>>
>> It's *you* that's making all the grandiose claims based upon nothing
>> but a vague press release. If you have any evidence to the contrary
>> then *present it*
>
>I am speculating, rather than making claims. There is no information
>about the Gamestorm platform yet. You have been told that previously.
>
And yet you create a thread proclaiming DirectX dead?
Your "speculation" is the speculation of a clueless fool. Don't
expect to shoot your mouth off without being challenged on it.
>>
>>
>>>Also, the absence
>>>of any specific video hardware requirements for the game seems
>>>misleading at best.
>>
>>
>> Misleading to you maybe, but nobody else seems confused by it.
>
>It would be misleading to anyone who intended to use the game who lacked
>the specific hardware that you allege is required.
>
Which hardware is that? A 3D video card?
>
>>
>>
>>>It's interesting that you mention Mesa, since the
>>>standard Mesa release supports only software rendering which is what I
>>>have been suggesting might be the case for the Gamestorm thing.
>>>
>>
>>
>> LOL - wow, that thing will be a real dog!
>
>Time will tell us.
>
Rational thought and all the evidence has already told us.
>>
>>
>>>>
>>>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>>>to be a part of the project.
>>>>>
>>>>>I don't see why not. Exactly the opposite would seem to be the case
>>>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>>>graphics implementation.
>>>>
>>>>
>>>>So they would willing cut their own throats by simultaneously making
>>>>their hardware irrelevant and assisting their competitors? Perhaps.
>>>
>>>No one can stay with the same approach forever. Things change very
>>>quickly in the computer business. The whole idea of having a dedicated
>>>video hardware with its own memory sticking out on the PCI/AGP bus
>>>doesn't seem like a long-term thing to me for video.
>>
>>
>> Really? So you think that a general purpose CPU can do things just as
>> fast as dedicated silicon? ROTFL! Damn you're funny!
>
>Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
>than the video card processor.
LOL!!!! Not when it comes to rendering 3D graphics!
God, you are *such* a loser :)
>Speed, however, is affected by the
>design, overhead from other activities, and the software. Gamestorm
>makes no claim to be 'as fast as' a 32-bit platform or to be the
>'fastest' or anything like that. They only suggest that their
>performance is competitive with other game experiences.
>
So long as you have an accelerated graphics solution, sure. If you
try to run it in software emulation mode then prepare to be *very*
disappointed.
>>
>>
>>>The Opteron CPU is
>>>much more powerful than the video chip
>>
>>
>> Duh! The video chip wipes the floor with the Opteron for 3D
>> rendering.
>
>Reference?
>
See the render times from the 3D rendering benchmarks I've already
pointed you at.
>> There is no comparison. Otherwise, software rendering
>> would be just as fast as hardware rendering!
>
>The ultimate speed is affected by a lot of things other than the video
>card chip.
>
Of course - but that's irrelevant to your claim.
>>
>> id don't even include software renderers in their games since QuakeII
>> because a general purpose CPU can't keep up. Are you saying that
>> they are going to start again? ROTFLMAO!
>
>We'll have to see what Gamestorm is doing.
>
They have recompiled an OpenGL game and bundled it with Linux. That's
what they are doing.
>>
>>
>>
>>>and the Opteron system memory is
>>>much faster and has much greater bandwidth than the video memory.
>>
>>
>> DUH!
>>
>> 9800Pro = 21.8GB/s
>> Athlon64 = 6.4GB/sec
>
>The Opteron has the memory controller integrated with the Opteron
>processor which means that the front side bus (interface to memory) runs
>at the speed of the processor and has a front side bandwidth of 14.4
>GB/sec.
>
a) you are wrong about the speed
b) it's still MUCH slower than the Radeon
c) reference?
Your claim above is utterly false, David. You even admit it with
your own made-up figures! ROTFL!
>The more significant thing, however, is that that the Opteron system
>will typically have 1GB or more of memory (and a 1MB cache) while a
>high-performance video card will only have perhaps 128 MB of memory.
Blah blah blah - you don't *need* a gig of memory to render a scene,
David. And your claim that the Opteron has faster memory is quite
simply FALSE.
> It
>seems possible that a system with such a large high-speed memory could
>therefore render more images in memory and maintain an acceptabe
>framerate by working ahead.
So now your CPU is rendering MULTIPLE frames simultaneously with
*less* bandwidth than a video card? All while it also handles the
game engine? Damn but you are quite the little circus sideshow,
aren't you?
You are so stupid it hurts.
>The main limit for such a hypothetical
>system doing software rendering would probably be the speed of the AGP
>bus to transfer the rendered images to the video memory. Gamestorm
>hasn't claimed that their system is faster than all others but only has
>suggested that its performance is competitive. It seems possible that
>they could be doing some sort of software rendering although no
>information has been provided.
>
You are utterly lost on this David. You are talking absolute
bollocks.
>>
>[personal comment snipped]
>>
>>
>>> Seems
>>>like software rendering could ultimately provide better performance when
>>>the software is developed, even if Gamestorm has not done that.
>>
>>
>[personal comment snipped]
You have no idea what you're talking about. Don't you ever get bored
of parading your stupidity?
When are you going to take me up on my bet? ;)
Menno Willemse
Irv Spalten wrote:
>
> Yes, but.... today, a lot of games live on different platforms and
> consoles. That is where the money is, universal usage. Single platform
> games don't do as well as multi-platform.
This type of game carries its own platform with it. The platform is not
Windows, or OS/2, or Linux. It is the PC.
> The 'real' question should be 'will' the other game producers abandon
> the DirectX support and shift.
There's no need for other game developers to join SCI. Even if SCI is
the only company to use this technology, it still enables them to reach
more users than Windows-only developers. If anything, it gives them an edge.
> Why would they if they already can
> produce the games that work on Win machines and now create a bootable CD
> that would work on the same machine?
That's not the point. The point is they can sell their game to people
who have NOT got Windows on their PCs, like me.
If SCI can produce a game that will run on any PC regardless of what is
on the hard disk, then you increase your market. You can sell to OS/2
users, Linux users AND Windows users instead of just to Windows users.
Also you remove the Microsoft dependency if you use Linux under the hood.
> Also, take into account the many libraries, tools, and routines these
> existing companies have for creating games on the existing platforms.
> Can these still be used or will time and money need to be spent to
> create the same for the new guy?
Well, the web article didn't say. Apparently, the game in question
exists for Windows as well so I would guess SCI could develop both
versions with the same tools. I don't know how much effort that takes.
> However, you are talking about a single company, not an industry. When
> the game industry embraces it, then you have something. Until then, it
> is just another 'attempt'.
I don't see it that way. I think that SCI has a nice new technology that
enables them to produce games and address markets that are currently
closed to Windows-only companies.
> You make it sound simple, easy, and cheap. The question 'why' would
> they? Yes, roll it out in a NEW release that they can sell, makes sense.
> Go back and update a 2 year old game with hardly any new sales, doesn't
> make sense, unless it is just a DD, not a programming change.
You only need to make the new device do the old tricks as well as the
new once. For most cases, a new device driver would be enough, I think.
I'm assuming here that SCI don't redesign their entire device driver
paradigm whenever a new card comes out. No matter what new tricks a
video card can do, it still needs to draw triangles at some point. The
hardest part would be when a new card "forgets" an old feature.
It wouldn't actually be much of an investment either. You have the
filetree of the application. All you need to do is add the new device
driver, run mkisofs and drop the resulting file on your server. The
exciting new features of the later hardware wouldn't be used by the old
game anyway. Just for laughs, I started up Stellar Frontier. It's years
old and still being played.
> Hmm, any platform will give you problems? One 'feature' of DirectX or
> OpenGL is the non-need to do the hardware things, thus eliminating both
> serious coding work and making the program see the h/w as a transparent
> item, no need to enumerate or write code for specific h/w.
Given the fact that under the hood is Linux, device drivers are usually
produced by the Linux crowd. Which is not to say that SCI couldn't lend
a hand with device driver development. Seems only fair.
> Royalties? Like what? DirectX doesn't have royalties, does it?
Maybe the payment takes some other form. Produce for Windows, pay Microsoft.
> Spurious changes, I assume you mean DirectX version changes causing
> incompatibilies?
For instance. Or the user changing the screen resolution, or otherwise
mucking about with Windows internals essential to the program. It could
also be the other way round. I've actually seen someone muck up his
system by installing a game. I've installed Black & White on a Windows
2000 computer (not my own, I hasten to add) and it simply didn't work.
Installing several games on your Windows box may not be impossible, but
it's not a smooth ride either.
> I've got a bookcase of at least 20 to 30 games. Each has been loaded at
> one point onto a computer and the CD's put away (except for those
> blasted ones, like MS FS 2004, that require a CD to play [there is a
> crack to make this a non-issue]). Seems more handy to me than needed a
> different CD each time I want to play a game.
I wonders... Didn't Linux have some way of booting from a file on
harddisk? This would mean that you'd install the Great Big File
somewhere on your disk, and boot from there. Sounds doable to me. It's
the way I run Knoppix with VPC at the moment.
> I don't 'do' consoles, but they require the game usually to be inserted
> before you can play it, no? I guess this will turn a PC into a game
> console?
Essentially, yes. Though I think there would be ways round having to
insert the CD.
> I do occasionally RUN something in the background and play a game. Handy
> feature that is usefull. I like it that way too.
I admit that's one disadvantage of this system. When you have the game
controlling the whole PC, you can't have Windows doing the same.
Emulators are too slow.
Cheers/2,
Menno
David Sutherland
<djoh...@isomedia.com> wrote:
What, that you are unable to face up to your mistakes? I must have
missed it. Glad you agree, though.
>>
>>
>>>>
>>>>>Obviously, that would be impossible for hardware that is developed after
>>>>>the CD is released and it would be difficult to be complete and
>>>>>up-to-date for all of the hardware existing at the time of its release.
>>>>>The game manufacturer would have to be constantly surveying the
>>>>>hardware to be sure they were supporting both new and old hardware when
>>>>>they released their games.
>>>>>
>>>>
>>>>
>>>>So why do you keep claiming that drivers aren't necessary? Make up
>>>>your mind!
>>>
>>>Software to control hardware (i.e. drivers) will always be needed.
>>>However, this software does not seem to be hardware-specific on the
>>>Gamestorm/Opteron platform. As I mentioned, though, the details are not
>>>yet available.
>>>
>>
>>
>> Yawn. You're an idiot if you think that drivers aren't hardware
>> specific.
>
>You are an idiot if you think that hardware drivers are always
>hardware-specific. I'll spell it for you: S-T-A-N-D-A-R-D-S.
>
I'll spell it out for you: P-E-R-F-O-R-M-A-N-C-E
We *are* talking about video drivers, aren't we? Try playing your
game with a VGA driver - LOL!
>>
>>
>>
>>>>
>>>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>>>Apparently, it is not difficult.
>>>>>
>>>>>Look what Gentoo says about that game on their website: "This CD
>>>>>includes the full version of the America's Army game. The game runs on
>>>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>>>
>>>>>The very restricted choice of video hardware supported by Gentoo
>>>>>supports my point and restricts the number of systems which can play the
>>>>>game.
>>>>>
>>>>
>>>>
>>>>Yes, it makes it clear that these tailored distributions may not
>>>>support all the hardware out there. Which was raised previously and
>>>>ignored by you.
>>>
>>>The Gentoo game distribution clearly specifies a very limited set of
>>>hardware that it will run on. The Gamestorm/Opteron announcement does
>>>not, hence my comment.
>>>
>>
>>
>> The Gamestorm press release said very little indeed. They did not, for
>> example, specify that they had failed to create world peace. By your
>> logic that means that all those nasty problems in the Middle East are
>> now solved. Oh good.
>
>Now you have strayed off-topic and are talking nonsense.
>
I'm illustrating the emptiness of your "logic." Try to keep up.
>>
>>
>>>>
>>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>>
>>>>>See above.
>>>>>
>>>>
>>>>
>>>[personal comment deleted]
>>
>>
>[personal comment snipped]
You're an idiot, and you are full of shit. It's not personal David,
it's a statement of fact.
>>
>>
>>>>
Unless you are happy with a truly awful image and abysmal frame-rates,
then yes - hardware is required. Unless of course you are looking to
play Tetris! Americas Army is a FPS, and will run like a dog without
hardware acceleration, if it runs at all.
So given enough time a moped will be as fast as a Ferrari? Wow!
>> and they would DIE if
>> they did what you are claiming. If you believe otherwise it's just
>> another example of how ignorant you are about How Things Work.
>>
>>
>>>
>>>>Or tell you what - why don't you try proving it instead of making yet
>>>>more moronic speculations:
>>>
>>>There are no details that have been released, yet, on the Gamestorm
>>>platform for the Opteron but I think it is very interesting.
>>>
>>
>>
>> I think it's very interesting that without any details at all you
>> create entire fantasy worlds where the laws of physics do not apply,
>
>No laws of physics are violated by software rendering. <laugh>
>
Is *that* what you understood me to have said? Guess your RCP (aka
deliberate avoidance of the argument) is starting to kick in.
>> and then embarrass yourself by pretending they are real in this ng.
>> What is that about?
>
>You are the guy posting about violations of the 'laws of physics.'
>You're sounding unusually nutty today.
>
Apparently you are having trouble following the conversation. Ain't
that a surprise.
If you want "nutty", check out the stupid claims of the guy who
started this thread. He thinks that video acceleration is obsolete,
that the Opteron needs no drivers to address any hardware now or in
the future, and that a company called SCI has developed a brand-new
platform independent competitor to DirectX and OpenGL - all in the
space of a couple of months! Initially he also claimed that they
wrote Americas Army, but you know - he's kind of gone silent on that
point :)
R.O.T.F.L.M.A.O.@.U!!!
>>
>>
>>>>[David T Johnsons URLs go here]
>>>>
>>
>>
>> Where's your evidence, David?
>
>There are no details that have been released, yet, on the Gamestorm
>platform for the Opteron but I think it is very interesting.
>
In other words you have no evidence. There is, however, plenty of
evidence that shows your "speculation" is ill-informed, where it is
not outright stupid.
>>
>> And hey, why aren't you taking me up on that bet? LOL!
>>
>
>There is no reason to wager with you. You would probably be a piker,
>anyway.
I'm quite serious, David. I bet you $100 that the SCI distribution
ships with hardware-specific drivers for video and sound. Put up, or
shut up.
The real reason you don't want to accept is because you *know* that
you are full shit, but your field-mouse sized testicles won't let you
admit it. Poor little David! You are a gutless coward who likes to
talk big but doesn't even have the strength of his own convictions...
pathetic.
David Sutherland
wrote:
>Hello World,
>
>Irv Spalten wrote:
>>
>> Yes, but.... today, a lot of games live on different platforms and
>> consoles. That is where the money is, universal usage. Single platform
>> games don't do as well as multi-platform.
>
>This type of game carries its own platform with it. The platform is not
>Windows, or OS/2, or Linux. It is the PC.
>
>> The 'real' question should be 'will' the other game producers abandon
>> the DirectX support and shift.
>
>There's no need for other game developers to join SCI. Even if SCI is
>the only company to use this technology, it still enables them to reach
>more users than Windows-only developers. If anything, it gives them an edge.
>
The "technology" in question is OpenGL. SCI did not create it. All
they've done is bundle an existing game on a bootable CD. If you
wanted to, you could roll your own.
[snip]
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
>
>Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
>than the video card processor. Speed, however, is affected by the
>design, overhead from other activities, and the software. Gamestorm
>makes no claim to be 'as fast as' a 32-bit platform or to be the
>'fastest' or anything like that. They only suggest that their
>performance is competitive with other game experiences.
Given that your claims revolve around the performance of "Americas
Army", this quote direct for the developers website is really quite
amusing:
http://www.americasarmy.com/faq_linux.php?p=3&t=8#faq4
Q: Is there a software only mode?
A: No. Modern 3D graphics cards are so powerful that even the fastest
CPU would not be able to take their place with a software mode. Since
America's Army and the new Unreal Engine take full advantage of the
processing power provided by these video cards, we cannot support
software rendering.
Hilarious, don't you think! :)
Here's some more for chuckles:
Q: I have Intel or Intel3D integrated graphics , can I run the game?
A: No, you can't. Please read the minimum system requirements.
(Last Updated: 2003-08-21)
Wondering what those might be? Knew you'd ask!
http://www.americasarmy.com/faq_linux.php?p=3&t=4
Q: What are the minimum system requirements for the Linux v1.9 Client?
A:
Pentium3 766 (or equivalent)
128MB RAM
glibc 2.1 or newer.
Hardware accelerated OpenGL drivers (Nvidia's and XiG's are known)
Q: I have a Kyro, KyroII, or Hercules 'prophet' branded video card,
can I run the game?
A: No, you can't. Please read the minimum system requirements. While
the kyro chipsets are recent and have the neccessary amount of video
ram, they do not have some of the transform and lighting features
required to show the high quality rendering that this game uses.
==
Face facts, David - your claims are complete nonsense. Even the
people whose product you are trying to use as evidence directly
contradict everything you say.
Fact - the game needs an OpenGL accelerated card.
Fact - the Opteron (or any other CPU) cannot perform the same
functions as dedicated 3D hardware in an acceptable manner.
Fact - if your card isn't fast enough, you can't play (so much for
hardware independence!)
Fact - you have been talking crap.
Irv Spalten
Menno Willemse wrote:
> Hello World,
>
> Irv Spalten wrote:
> >
> > Yes, but.... today, a lot of games live on different platforms and
> > consoles. That is where the money is, universal usage. Single platform
> > games don't do as well as multi-platform.
>
> This type of game carries its own platform with it. The platform is not
> Windows, or OS/2, or Linux. It is the PC.
I was mopre refuting dumbo's allegations that DirectX was dead. What
we've both said *is* true however.
From a GAME DEVELOPER'S viewpoint, which would you rather write to? A
single interface that has not been well accepted yet, a single platform,
or, an API set with tools to support multiple platforms (write once,
work on many).
Check this out,
http://www.gamespot.com/pc/action/freedomthebattleforli/index.html,
ships on PC, XBOX, Paystation 2, and GameCube. 4 birds with one set of
source I bet.
Not to mention that they have the tools to do it.
Why would THEY switch? If the tools were there to create the version
from thier source, they probably would. Without it, costly and an uphill
challenge.
>
> > The 'real' question should be 'will' the other game producers abandon
> > the DirectX support and shift.
>
> There's no need for other game developers to join SCI. Even if SCI is
> the only company to use this technology, it still enables them to reach
> more users than Windows-only developers. If anything, it gives them an edge.
Well, not all developer's are 'Windows Only'. Many have the same titles
on PC's (Windows) and gameboxes. These are usually from a single source.
Big seller's.
Besides, it is the TOPIC OF THE GAME that determines if it sells well.
There has to be a learning curve to use it too. Why switch? For a few
more customers? What if they can't get equal performance/features as
they don't know how? What is the true incentive?
>
> > Why would they if they already can
> > produce the games that work on Win machines and now create a bootable CD
> > that would work on the same machine?
>
> That's not the point. The point is they can sell their game to people
> who have NOT got Windows on their PCs, like me.
True, but if you 'really' wanted to play a game, you'd probably get a
platform that ran the game? True, self-contained booting would make it
easier for you. 'In the beginning', people bought PC's to run Visicalc,
no?
One can deduce either you really don't care to play any game out there
or you are unwilling to put down a game playing OS platform. If the
first, well why would one SCI puts out interest you? I'm sure 'versions'
of the game exist today (shooters, arcade, adventure, simulators, etc.)
that are reasonable facsimiles. If the latter, then no game has you
interested enough to do so (or you don't have the h/w or space).
> If SCI can produce a game that will run on any PC regardless of what is
> on the hard disk, then you increase your market.
BIG if! Questions should be :
1) Can the game OUTPERFORM other platforms.
2) Can it work on enough h/w to make it acceptable to a
large majority of people out there.
3) Does the game offer end-users a reason to buy it?
4) Can existing developers move to the platform
effortlessly?
Like I said, some developers write single source and then use tools to
create platform versions. This method would be the reverse, single CD,
multiple platforms. Probably cheaper for a developer, but with they
1) Reach the same audiences with the single CD (XBOX, GameCube,
Playstation, PC)?
2) Have the same capability on all the h/w?
3) Have no steep learning curve?
4) see cost savings with this method?
If all the above are YES, then maybe DirectX could be challenged as the
vehicle to write s/w. Otherwise, if all are NO, why would anyone do it,
at least EXISTING big s/w game companies?
> You can sell to OS/2
> users, Linux users AND Windows users instead of just to Windows users.
Not just Win users here, there are the consoles from the same source.
> Also you remove the Microsoft dependency if you use Linux under the hood.
Why if it works for the game developer? They have all learned to live
with it.
Questions, assuming that it can be done, that is write a game one would
have produced for WinXX only utilizing DirectX, would there be the same
performance and would the COST to the end-user be the same? Developers
might not have to pay MS anything to produce a game for WinXX (other
than purchase some MS compilers and SDK's probably), but what would be
the cost of the Bootable CD with Linux on it be as an adder?
> > Also, take into account the many libraries, tools, and routines these
> > existing companies have for creating games on the existing platforms.
> > Can these still be used or will time and money need to be spent to
> > create the same for the new guy?
>
> Well, the web article didn't say. Apparently, the game in question
> exists for Windows as well so I would guess SCI could develop both
> versions with the same tools. I don't know how much effort that takes.
If those tools exist, then it should be easy and well accepted. So the
COST of the program in Linux and WinXX versions would/could be the
deciding factor.
>
> > However, you are talking about a single company, not an industry. When
> > the game industry embraces it, then you have something. Until then, it
> > is just another 'attempt'.
>
> I don't see it that way. I think that SCI has a nice new technology that
> enables them to produce games and address markets that are currently
> closed to Windows-only companies.
Beta and OS/2 were nice new technologies. Adaption is what counts.
> > You make it sound simple, easy, and cheap. The question 'why' would
> > they? Yes, roll it out in a NEW release that they can sell, makes sense.
> > Go back and update a 2 year old game with hardly any new sales, doesn't
> > make sense, unless it is just a DD, not a programming change.
>
> You only need to make the new device do the old tricks as well as the
> new once. For most cases, a new device driver would be enough, I think.
> I'm assuming here that SCI don't redesign their entire device driver
> paradigm whenever a new card comes out. No matter what new tricks a
> video card can do, it still needs to draw triangles at some point. The
> hardest part would be when a new card "forgets" an old feature.
>
> It wouldn't actually be much of an investment either. You have the
> filetree of the application. All you need to do is add the new device
> driver, run mkisofs and drop the resulting file on your server. The
> exciting new features of the later hardware wouldn't be used by the old
> game anyway. Just for laughs, I started up Stellar Frontier. It's years
> old and still being played.
It still takes manpower and costs to the developer to do this. Oh, I
STILL play the OLD DOS Empire game, one of my favorites. Works on WinXP
and eCS too... This is an early 80's game no less. Still topical and I
prefer it to other 'war' games.
>
> > Hmm, any platform will give you problems? One 'feature' of DirectX or
> > OpenGL is the non-need to do the hardware things, thus eliminating both
> > serious coding work and making the program see the h/w as a transparent
> > item, no need to enumerate or write code for specific h/w.
>
> Given the fact that under the hood is Linux, device drivers are usually
> produced by the Linux crowd. Which is not to say that SCI couldn't lend
> a hand with device driver development. Seems only fair.
>
> > Royalties? Like what? DirectX doesn't have royalties, does it?
>
> Maybe the payment takes some other form. Produce for Windows, pay Microsoft.
Not to my knowledge, other than the MS tools to create programs. Heck,
MS will even publish YOUR game or buy your company <G>...
> > Spurious changes, I assume you mean DirectX version changes causing
> > incompatibilies?
>
> For instance. Or the user changing the screen resolution, or otherwise
> mucking about with Windows internals essential to the program. It could
> also be the other way round. I've actually seen someone muck up his
> system by installing a game. I've installed Black & White on a Windows
> 2000 computer (not my own, I hasten to add) and it simply didn't work.
> Installing several games on your Windows box may not be impossible, but
> it's not a smooth ride either.
I've seen people BACK LEVEL DirectX or overwrite SYSTEM DLL's and muck
up a system.
I've had many programs mess up my desktop by not resetting them to my
original resolution too. These are in a minority, and generally freebies
downloaded from the web however.
> > I've got a bookcase of at least 20 to 30 games. Each has been loaded at
> > one point onto a computer and the CD's put away (except for those
> > blasted ones, like MS FS 2004, that require a CD to play [there is a
> > crack to make this a non-issue]). Seems more handy to me than needed a
> > different CD each time I want to play a game.
>
> I wonders... Didn't Linux have some way of booting from a file on
> harddisk? This would mean that you'd install the Great Big File
> somewhere on your disk, and boot from there. Sounds doable to me. It's
> the way I run Knoppix with VPC at the moment.
Possibly, but many images would be large. Whenever I get a game, the
first thing I do is SEARCH for a crack to remove the need to keep a CD
in to play the game. For instance, the NEW MS 2004 Flight Simulator,
within DAYS of its release, the crack was available. I don't like stacks
of CD's needing to be within reach to play a game, sorry.
> > I don't 'do' consoles, but they require the game usually to be inserted
> > before you can play it, no? I guess this will turn a PC into a game
> > console?
I don't do them either. Yup, it would. Again, the GAME has to be the
draw. For different people, different game styles will be the draw. Just
like the original PC, people bought the PC to run VisiCalc. Make a game
on that format people WANT and they will buy it.
> Essentially, yes. Though I think there would be ways round having to
> insert the CD.
>
> > I do occasionally RUN something in the background and play a game. Handy
> > feature that is usefull. I like it that way too.
>
> I admit that's one disadvantage of this system. When you have the game
> controlling the whole PC, you can't have Windows doing the same.
> Emulators are too slow.
>
>
> Cheers/2,
> Menno
>
Irv
David T. Johnson
>
>
>
>
>>> All you are doing is making it painfully obvious how
>>>bankrupt your claims are.
>>
>>Painful for you, perhaps. <laugh>
>>
>
>
>
>>>
>>>>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>>>>Apparently, it is not difficult.
>>>>>>
>>>>>>Look what Gentoo says about that game on their website: "This CD
>>>>>>includes the full version of the America's Army game. The game runs on
>>>>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>>>>
>>>>>>The very restricted choice of video hardware supported by Gentoo
>>>>>>supports my point and restricts the number of systems which can play the
>>>>>>game.
>>>>>
>>>>>
>>>>>The restricted choice of video hardware is due to the limited number of
>>>>>vendors with Linux drivers that support 3D acceleration in hardware.
>>>>>It's a short list: ATI and nVidia. No serious gamer in their right
>>>>>mind is going to slow down their machine by using software rendering.
>>>>
>>>>I agree that this is true with typical 32-bit systems. The
>>>>Gamestorm/Opteron platform, however, has apparently found a way to get
>>>>acceptable performance without using hardware-specific drivers.
>>>>
>>>
>>>
>>>Where does it say that? Why were Nvidia involved if there are no
>>>hardware or driver dependencies?
>>
>>What do you think Nvidia does, anyway. Most of their work is
>>software-related. Even their 'hardware' is mostly a software effort.
>>
>
>
>
> Sure David, the silicon has got nothing to do with it, and all along
> ATI and Nvidia have just been pulling everyone's leg.
Incorrect.
>
[snip of ad hominem]
>
>
>>>
>>>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>>>
>>>>>>See above.
>>>>>
>>>>>
>>>>>Was there an answer up there? All I saw was a regurgitated quote
>>>>
>>>>>from a website that neither supports nor denies your claim as to the
>>>>
>>>>>supposed difficulty of the task.
>>>>
>>>>The very restricted choice of video hardware supported by Gentoo
>>>>supports my point and restricts the number of systems which can play the
>>>>game.
>>>>
>>>
>>>
>>>It doesn't support your point at all. Nvidia's presence in that
>>>press release completely vaporises your "point".
>>
>>No, it doesn't. Rather, it supports it because Nvidia is obviously
>>involved and yet no specific hardware requirement for Nvidia hardware
>>was listed.
>>
>
>
[snip of ad hominems]
Gamestorm makes no claim to be 'as fast as' a 32-bit platform or to be
the 'fastest' or anything like that. They only suggest that their
performance is competitive with other game experiences.
>
>
>>"They can run the
>>game on a floor model at Best Buy, and leave the hard drive untouched,"
>>he says."
>>
>
>
> All the PC's on the floor at Best Buy have video cards in them.
>
>
>>Other claims by SCI:
>>
>>"cuts out legacy operating system overhead..."
>>"without the usual overhead from the legacy operating systems
>>raditionally used in the gaming industry, SCI claims."
>>"With the AMD Athlon 64 processor and GameStorm technology, AMD is able
>>to showcase a fully-integrated 64-bit environment that delivers
>>performance and realism to the most demanding gamers,"
>>"the company is now working on porting several more 32-bit and 64-bit
>>games to the GameStorm platform"
>>
>
>
> I've already refuted these.
You are not capable of refuting the claims made by SCI about their
Gamestorm platform because 1) you obviously don't know very much about
computer graphics and 2) you don't know anything about the Gamestorm
platform.
>
>
>
>>The last sentence is revealing because it suggests that 'porting' is
>>required to run other games on the Gamestorm platform which suggests
>>that Gamestorm does not run OpenGL games.
>>
>
>
> Bullshit. Americas Army is an OPenGL game as you have already been
> told.
The 32-bit version is an OpenGL game. We don't know anything about the
new 64-bit 'Gamestorm' version yet.
>
> That you ignore the evidence is *your* problem, not mine.
You are ignoring the facts and the discussion points.
>
>
>
>>> It's *you* that's making all the grandiose claims based upon nothing
>>>but a vague press release. If you have any evidence to the contrary
>>>then *present it*
>>
>>I am speculating, rather than making claims. There is no information
>>about the Gamestorm platform yet. You have been told that previously.
>>
>
>
> And yet you create a thread proclaiming DirectX dead?
Correct.
>
> Your "speculation" is the speculation of a clueless fool. Don't
> expect to shoot your mouth off without being challenged on it.
If these kinds of speculations are not to be permitted by you here in a
usenet newsgroup, then exactly where would you permit them? <laugh> My
speculation is probably pretty good but we'll have to wait and see what
comes out about the Gamestorm platform in the coming weeks.
>
>
>>>
>>>>Also, the absence
>>>>of any specific video hardware requirements for the game seems
>>>>misleading at best.
>>>
>>>
>>>Misleading to you maybe, but nobody else seems confused by it.
>>
>>It would be misleading to anyone who intended to use the game who lacked
>>the specific hardware that you allege is required.
>>
>
>
> Which hardware is that? A 3D video card?
Yes, you have alleged that the 64-bit America's Army game will not run
with acceptable performance unless specific 3D video hardware is
present. The Gamestorm SCI announcement, OTOH, makes no mention of
specific hardware requirements.
>
>
>>>
>>>>It's interesting that you mention Mesa, since the
>>>>standard Mesa release supports only software rendering which is what I
>>>>have been suggesting might be the case for the Gamestorm thing.
>>>>
>>>
>>>
>>>LOL - wow, that thing will be a real dog!
>>
>>Time will tell us.
>>
>
>
> Rational thought and all the evidence has already told us.
Ahhh yes. The same 'evidence' no doubt that your
similarly-free-thinking ancestors used as a justification for their need
to burn those 'witches' at the stake.
>
>
>>>
>>>>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>>>>to be a part of the project.
>>>>>>
>>>>>>I don't see why not. Exactly the opposite would seem to be the case
>>>>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>>>>graphics implementation.
>>>>>
>>>>>
>>>>>So they would willing cut their own throats by simultaneously making
>>>>>their hardware irrelevant and assisting their competitors? Perhaps.
>>>>
>>>>No one can stay with the same approach forever. Things change very
>>>>quickly in the computer business. The whole idea of having a dedicated
>>>>video hardware with its own memory sticking out on the PCI/AGP bus
>>>>doesn't seem like a long-term thing to me for video.
>>>
>>>
>>>Really? So you think that a general purpose CPU can do things just as
>>>fast as dedicated silicon? ROTFL! Damn you're funny!
>>
>>Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
>>than the video card processor.
>
>
> LOL!!!! Not when it comes to rendering 3D graphics!
The NV34 GPU in my video card has a clock speed of 325 Mhz which is not
exactly lightning-speed anymore.
[snip of ad hominem]
>
>>Speed, however, is affected by the
>>design, overhead from other activities, and the software. Gamestorm
>>makes no claim to be 'as fast as' a 32-bit platform or to be the
>>'fastest' or anything like that. They only suggest that their
>>performance is competitive with other game experiences.
>>
>
>
> So long as you have an accelerated graphics solution, sure. If you
> try to run it in software emulation mode then prepare to be *very*
> disappointed.
We will all find out how disappointed we will be. <laugh> In the
meantime, DirectX is looking a little green around its gills.
>
>
>>>
>>>>The Opteron CPU is
>>>>much more powerful than the video chip
>>>
>>>
>>>Duh! The video chip wipes the floor with the Opteron for 3D
>>>rendering.
>>
>>Reference?
>>
>
>
> See the render times from the 3D rendering benchmarks I've already
> pointed you at.
Those are 32-bit applications as I told you previously and they do not
make any comparison of performance between the video card GPU and the
Opteron CPU.
>
>
>
>>> There is no comparison. Otherwise, software rendering
>>>would be just as fast as hardware rendering!
>>
>>The ultimate speed is affected by a lot of things other than the video
>>card chip.
>>
>
>
> Of course - but that's irrelevant to your claim.
No, it isn't. The Gamestorm platform may have acceptable performance
without the need for hardware-specific drivers.
>
>
>>>id don't even include software renderers in their games since QuakeII
>>>because a general purpose CPU can't keep up. Are you saying that
>>>they are going to start again? ROTFLMAO!
>>
>>We'll have to see what Gamestorm is doing.
>>
>
>
> They have recompiled an OpenGL game and bundled it with Linux. That's
> what they are doing.
SCI also claims to have 'ported' the the game to their Gamestorm
platform. They further claim that it is now a 64-bit game, has
acceptable performance, and runs on Linux.
>
>
>>>
>>>
>>>>and the Opteron system memory is
>>>>much faster and has much greater bandwidth than the video memory.
>>>
>>>
>>>DUH!
>>>
>>>9800Pro = 21.8GB/s
>>>Athlon64 = 6.4GB/sec
>>
>>The Opteron has the memory controller integrated with the Opteron
>>processor which means that the front side bus (interface to memory) runs
>>at the speed of the processor and has a front side bandwidth of 14.4
>>GB/sec.
>>
>
>
> a) you are wrong about the speed
No.
> b) it's still MUCH slower than the Radeon
But not slower than most current video cards which have memory bandwidth
in the 10-15 GB/sec range.
> c) reference?
Google is your friend.
>
> Your claim above is utterly false, David. You even admit it with
> your own made-up figures! ROTFL!
Incorrect.
>
>
>>The more significant thing, however, is that that the Opteron system
>>will typically have 1GB or more of memory (and a 1MB cache) while a
>>high-performance video card will only have perhaps 128 MB of memory.
>
>
> Blah blah blah - you don't *need* a gig of memory to render a scene,
> David.
More memory allows many more things to be done, in particular much more
detail can be provided at a higher framerate. That is likely to be one
of the features of the Gamestorm 64-bit platform. The point is that
there is never too much memory.
> And your claim that the Opteron has faster memory is quite
> simply FALSE.
Incorrect.
>
>
>
>>It
>>seems possible that a system with such a large high-speed memory could
>>therefore render more images in memory and maintain an acceptabe
>>framerate by working ahead.
>
>
> So now your CPU is rendering MULTIPLE frames simultaneously with
> *less* bandwidth than a video card? All while it also handles the
> game engine?
'Integrated' graphics setups on current 32-bit systems use the system
memory for video memory. You did know that didn't you? A 64-bit CPU
could certainly provide better video performance than a 32-bit CPU if
64-bit software was used.
>
[ad hominem snipped]
>
>
>>The main limit for such a hypothetical
>>system doing software rendering would probably be the speed of the AGP
>>bus to transfer the rendered images to the video memory. Gamestorm
>>hasn't claimed that their system is faster than all others but only has
>>suggested that its performance is competitive. It seems possible that
>>they could be doing some sort of software rendering although no
>>information has been provided.
>>
>
>
[ad hominem snipped]
>
>
>>[personal comment snipped]
>>
>>>
>>>>Seems
>>>>like software rendering could ultimately provide better performance when
>>>>the software is developed, even if Gamestorm has not done that.
>>>
>>>
>>[personal comment snipped]
>
>
[ad hominem snipped]
>
> When are you going to take me up on my bet? ;)
>
When pigs fly? <laugh>
David T. Johnson
You point to the identical comment that I had made repeatedly and
accused me of 'wiggling'. <laugh>
>
>
>>>
>>>>>>Obviously, that would be impossible for hardware that is developed after
>>>>>>the CD is released and it would be difficult to be complete and
>>>>>>up-to-date for all of the hardware existing at the time of its release.
>>>>>>The game manufacturer would have to be constantly surveying the
>>>>>>hardware to be sure they were supporting both new and old hardware when
>>>>>>they released their games.
>>>>>>
>>>>>
>>>>>
>>>>>So why do you keep claiming that drivers aren't necessary? Make up
>>>>>your mind!
>>>>
>>>>Software to control hardware (i.e. drivers) will always be needed.
>>>>However, this software does not seem to be hardware-specific on the
>>>>Gamestorm/Opteron platform. As I mentioned, though, the details are not
>>>>yet available.
>>>>
>>>
>>>
>>>Yawn. You're an idiot if you think that drivers aren't hardware
>>>specific.
>>
>>You are an idiot if you think that hardware drivers are always
>>hardware-specific. I'll spell it for you: S-T-A-N-D-A-R-D-S.
>>
>
>
> I'll spell it out for you: P-E-R-F-O-R-M-A-N-C-E
>
> We *are* talking about video drivers, aren't we? Try playing your
> game with a VGA driver - LOL!
The VBE 3.0 standard could probably be used with all current video cards
with acceptable performance.
>
>
>
>>>
>>>
>>>>>>>How about the already-mentioned Gentoo Games, which produced the
>>>>>>>"America's Army" GameCD for Linux-32 (with, incidentally, the help of
>>>>>>>Super Computer, Inc)? An OS, X, drivers and game in a 525MiB ISO.
>>>>>>>Apparently, it is not difficult.
>>>>>>
>>>>>>Look what Gentoo says about that game on their website: "This CD
>>>>>>includes the full version of the America's Army game. The game runs on
>>>>>>recent Intel Pentium and AMD Athlon-based machines, and requires an
>>>>>>NVIDIA or recent ATI (Radeon 8500 or higher) graphics card."
>>>>>>
>>>>>>The very restricted choice of video hardware supported by Gentoo
>>>>>>supports my point and restricts the number of systems which can play the
>>>>>>game.
>>>>>>
>>>>>
>>>>>
>>>>>Yes, it makes it clear that these tailored distributions may not
>>>>>support all the hardware out there. Which was raised previously and
>>>>>ignored by you.
>>>>
>>>>The Gentoo game distribution clearly specifies a very limited set of
>>>>hardware that it will run on. The Gamestorm/Opteron announcement does
>>>>not, hence my comment.
>>>>
>>>
>>>
>>>The Gamestorm press release said very little indeed. They did not, for
>>>example, specify that they had failed to create world peace. By your
>>>logic that means that all those nasty problems in the Middle East are
>>>now solved. Oh good.
>>
>>Now you have strayed off-topic and are talking nonsense.
>>
>
>
> I'm illustrating the emptiness of your "logic." Try to keep up.
It sounded as if you were talking about the Middle East and world peace.
<laugh>
>
>
>>>
>>>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>>>
>>>>>>See above.
>>>>>>
>>>>>
>>>>>
>>>>[personal comment deleted]
>>>
>>>
>>[personal comment snipped]
>
>
[ad hominem snipped]
You seem to think that a quality image and acceptable frame rates are
not possible without a 3D card when the 64-bit Gamestorm platform is
used. Since no details on the Gamestorm platform are yet available, we
don't know what it is capable of.
> Unless of course you are looking to
> play Tetris!
My old DOS computer will play Doom and Castle Wolfenstein (both early 3D
games) without the need for accelerated 3D video cards or even a
powerful CPU. Both of these games are much more sophisticated that your
Tetris example. <laugh>
> Americas Army is a FPS, and will run like a dog without
> hardware acceleration, if it runs at all.
Time will tell...
Irrelevant.
>
>
>>>and they would DIE if
>>>they did what you are claiming. If you believe otherwise it's just
>>>another example of how ignorant you are about How Things Work.
>>>
>>>
>>>
>>>>>Or tell you what - why don't you try proving it instead of making yet
>>>>>more moronic speculations:
>>>>
>>>>There are no details that have been released, yet, on the Gamestorm
>>>>platform for the Opteron but I think it is very interesting.
>>>>
>>>
>>>
>>>I think it's very interesting that without any details at all you
>>>create entire fantasy worlds where the laws of physics do not apply,
>>
>>No laws of physics are violated by software rendering. <laugh>
>>
>
> Is *that* what you understood me to have said? Guess your RCP (aka
> deliberate avoidance of the argument) is starting to kick in.
David S: "...without any details at all you create entire fantasy
worlds where the laws of physics do not apply" <chuckle>
>
>
>>>and then embarrass yourself by pretending they are real in this ng.
>>>What is that about?
>>
>>You are the guy posting about violations of the 'laws of physics.'
>>You're sounding unusually nutty today.
>>
>
>
> Apparently you are having trouble following the conversation. Ain't
> that a surprise.
See above.
>
> If you want "nutty", check out the stupid claims of the guy who
> started this thread. He thinks that video acceleration is obsolete,
Incorrect. See above.
> that the Opteron needs no drivers to address any hardware now or in
> the future,
Incorrect. You have continued to make this false statement. I was
referring to 'hardware-specific' drivers which are probably on the way
out, thanks to new video standards.
> and that a company called SCI has developed a brand-new
> platform independent competitor to DirectX and OpenGL - all in the
> space of a couple of months! Initially he also claimed that they
> wrote Americas Army,
Incorrect. The authorship of AA has not been discussed.
> but you know - he's kind of gone silent on that
> point :)
This is a new claim for you.
>
> R.O.T.F.L.M.A.O.@.U!!!
>
>
>
>>>
>>>>>[David T Johnsons URLs go here]
>>>>>
>>>
>>>
>>>Where's your evidence, David?
>>
>>There are no details that have been released, yet, on the Gamestorm
>>platform for the Opteron but I think it is very interesting.
>>
>
>
> In other words you have no evidence. There is, however, plenty of
> evidence that shows your "speculation" is ill-informed, where it is
> not outright stupid.
Incorrect.
>
>
>>>And hey, why aren't you taking me up on that bet? LOL!
>>>
>>
>>There is no reason to wager with you. You would probably be a piker,
>>anyway.
>
>
> I'm quite serious, David. I bet you $100 that the SCI distribution
> ships with hardware-specific drivers for video and sound. Put up, or
> shut up.
There is no reason to wager with you.
[ad hominem snipped]
David T. Johnson
David T. Johnson
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Mon, 06 Oct 2003 08:40:25 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>> [snip]
>>
>>
>>>Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
>>>than the video card processor. Speed, however, is affected by the
>>>design, overhead from other activities, and the software. Gamestorm
>>>makes no claim to be 'as fast as' a 32-bit platform or to be the
>>>'fastest' or anything like that. They only suggest that their
>>>performance is competitive with other game experiences.
>>
>>
>> Given that your claims revolve around the performance of "Americas
>> Army", this quote direct for the developers website is really quite
>> amusing:
>>
>> http://www.americasarmy.com/faq_linux.php?p=3&t=8#faq4
>>
>> Q: Is there a software only mode?
>> A: No. Modern 3D graphics cards are so powerful that even the fastest
>> CPU would not be able to take their place with a software mode. Since
>> America's Army and the new Unreal Engine take full advantage of the
>> processing power provided by these video cards, we cannot support
>> software rendering.
>>
>> Hilarious, don't you think! :)
>
>This is referring to the current 32-bit version and not the new 64-bit
>Gamestorm version.
>
How is that relevant? They are not referring to the game engine
David, but the *CPU*. That comment was updated on the 21st August of
this year - when the Opteron was an available product. And the game
developers say that even the *fastest* CPU can't hope to match the
power of a hardware solution.
If you don't want to believe the people who know that engine better
than anyone else then it's obvious that you *choose* not to believe
them. That isn't surprising, but it does make clear just how low you
are prepared to sink rather than admit that you might be wrong.
>>
>> Here's some more for chuckles:
>>
>> Q: I have Intel or Intel3D integrated graphics , can I run the game?
>> A: No, you can't. Please read the minimum system requirements.
>> (Last Updated: 2003-08-21)
>>
>> Wondering what those might be? Knew you'd ask!
>>
>> http://www.americasarmy.com/faq_linux.php?p=3&t=4
>> Q: What are the minimum system requirements for the Linux v1.9 Client?
>> A:
>> Pentium3 766 (or equivalent)
>> 128MB RAM
>> glibc 2.1 or newer.
>> Hardware accelerated OpenGL drivers (Nvidia's and XiG's are known)
>
>This is referring to the current 32-bit version and not the new 64-bit
>Gamestorm version. The Gamestorm 64-bit version specifically states
>that it requires the 64-bit Opteron CPU.
>
How is that relevant? How many times faster do you think the Opteron
is in 64-bit mode? Two? Two hundred? Where are the benchmarks!
>
>>
>> Q: I have a Kyro, KyroII, or Hercules 'prophet' branded video card,
>> can I run the game?
>> A: No, you can't. Please read the minimum system requirements. While
>> the kyro chipsets are recent and have the neccessary amount of video
>> ram, they do not have some of the transform and lighting features
>> required to show the high quality rendering that this game uses.
>> ==
>>
>> Face facts, David - your claims are complete nonsense. Even the
>> people whose product you are trying to use as evidence directly
>> contradict everything you say.
>>
>> Fact - the game needs an OpenGL accelerated card.
>
>That is not a fact, nor even a reasonable supposition.
>
See the official website, David. Your personal speculations do not
outweigh it's assertions of fact. *THEY* know the game better than
you do.
>> Fact - the Opteron (or any other CPU) cannot perform the same
>> functions as dedicated 3D hardware in an acceptable manner.
>
>Also not a fact. The Opteron CPU is much more powerful than a typical
>graphics card GPU.
>
This is nonsense. If you really believe this then you'll be willing
to take my money over it, right?
US$100 says that the Opteron cannot render 3D games faster (or even
within 20%) the speed of the Radeon9800. Put up or shut up.
>> Fact - if your card isn't fast enough, you can't play (so much for
>> hardware independence!)
>
>Gamestorm did not list any specific video hardware requirements for
>their new game. You have been told this previously.
>
>> Fact - you have been talking crap.
>>
>
>That is certainly true for you.
LOL - have you noticed the scorn that people have been pouring on you
over this?
God you're pathetic.
David Sutherland
<djoh...@isomedia.com> wrote:
And?
>>
>>
>>>>
>>>>>>>Obviously, that would be impossible for hardware that is developed after
>>>>>>>the CD is released and it would be difficult to be complete and
>>>>>>>up-to-date for all of the hardware existing at the time of its release.
>>>>>>>The game manufacturer would have to be constantly surveying the
>>>>>>>hardware to be sure they were supporting both new and old hardware when
>>>>>>>they released their games.
>>>>>>>
>>>>>>
>>>>>>
>>>>>>So why do you keep claiming that drivers aren't necessary? Make up
>>>>>>your mind!
>>>>>
>>>>>Software to control hardware (i.e. drivers) will always be needed.
>>>>>However, this software does not seem to be hardware-specific on the
>>>>>Gamestorm/Opteron platform. As I mentioned, though, the details are not
>>>>>yet available.
>>>>>
>>>>
>>>>
>>>>Yawn. You're an idiot if you think that drivers aren't hardware
>>>>specific.
>>>
>>>You are an idiot if you think that hardware drivers are always
>>>hardware-specific. I'll spell it for you: S-T-A-N-D-A-R-D-S.
>>>
>>
>>
>> I'll spell it out for you: P-E-R-F-O-R-M-A-N-C-E
>>
>> We *are* talking about video drivers, aren't we? Try playing your
>> game with a VGA driver - LOL!
>
>The VBE 3.0 standard could probably be used with all current video cards
>with acceptable performance.
>
Define "acceptable." 40+fps at 1024x768 with bi or trilinear
filtering is acceptable to most, although many prefer a higher
framerate.
It sounded as if you were having trouble following the conversation
"<laugh>"
>>
>>
>>>>
>>>>>>>>Or is difficult=="DTJ doesn't know how to do it"?
>>>>>>>
>>>>>>>See above.
>>>>>>>
>>>>>>
>>>>>>
>>>>>[personal comment deleted]
>>>>
>>>>
>>>[personal comment snipped]
>>
>>
>[ad hominem snipped]
You're an idiot, and you are full of shit. It's not personal David,
it's a statement of fact.
>>
>>
>>>>
We know what the Opteron is capable of, and we know what 64bit
software is capable of. The benchmarks are out there. Ipso facto we
know what 64bit software on a 64bit Opteron is capable of.
>
>> Unless of course you are looking to
>> play Tetris!
>
>My old DOS computer will play Doom and Castle Wolfenstein (both early 3D
>games) without the need for accelerated 3D video cards or even a
>powerful CPU. Both of these games are much more sophisticated that your
>Tetris example. <laugh>
>
I'm really starting to wonder if you've even *seen* a modern 3D game.
Castle Wolfenstein? ROTFLMAO! That was released in 1981 - I would
sure hope your PC can run it!
Maybe you meant Wolfenstein3D? If so - check out the screenshot here:
http://members.chello.at/theodor.lauppert/games/wolf3d.htm
Wow, such cutting edge graphics! You are a joke!
>> Americas Army is a FPS, and will run like a dog without
>> hardware acceleration, if it runs at all.
>
>Time will tell...
>
The developers have stated it as a matter of public record. We don't
need to guess.
Why? Because you don't like it? LOL!
>>
>>
>>>>and they would DIE if
>>>>they did what you are claiming. If you believe otherwise it's just
>>>>another example of how ignorant you are about How Things Work.
>>>>
>>>>
>>>>
>>>>>>Or tell you what - why don't you try proving it instead of making yet
>>>>>>more moronic speculations:
>>>>>
>>>>>There are no details that have been released, yet, on the Gamestorm
>>>>>platform for the Opteron but I think it is very interesting.
>>>>>
>>>>
>>>>
>>>>I think it's very interesting that without any details at all you
>>>>create entire fantasy worlds where the laws of physics do not apply,
>>>
>>>No laws of physics are violated by software rendering. <laugh>
>>>
>>
>> Is *that* what you understood me to have said? Guess your RCP (aka
>> deliberate avoidance of the argument) is starting to kick in.
>
>David S: "...without any details at all you create entire fantasy
>worlds where the laws of physics do not apply" <chuckle>
>
Claiming that an Opteron which may have to perform many hundreds of
calculations to achieve a particular result is as fast or faster than
dedicated hardware which can do it in one pass *is* ignoring the rules
of physics.
Claiming that doubling the number of bits renders hardware drivers
obsolete is ignoring the rules of physics.
Which of these are you having trouble following?
>>
>>
>>>>and then embarrass yourself by pretending they are real in this ng.
>>>>What is that about?
>>>
>>>You are the guy posting about violations of the 'laws of physics.'
>>>You're sounding unusually nutty today.
>>>
>>
>>
>> Apparently you are having trouble following the conversation. Ain't
>> that a surprise.
>
>See above.
>
Where?
>>
>> If you want "nutty", check out the stupid claims of the guy who
>> started this thread. He thinks that video acceleration is obsolete,
>
>Incorrect. See above.
>
Yawn. You claimed that the 64bit Opteron does not need
hardware-specific drivers and can perform 3D ops as fast as dedicated
graphics hardware.
>> that the Opteron needs no drivers to address any hardware now or in
>> the future,
>
>Incorrect. You have continued to make this false statement. I was
>referring to 'hardware-specific' drivers which are probably on the way
>out, thanks to new video standards.
>
Define "hardware-specific." Your usage appears to be your usual
self-serving evasive bullshit.
>
>> and that a company called SCI has developed a brand-new
>> platform independent competitor to DirectX and OpenGL - all in the
>> space of a couple of months! Initially he also claimed that they
>> wrote Americas Army,
>
>Incorrect. The authorship of AA has not been discussed.
>
You claimed that SCI created an entirely new platform-agnostic API to
replace both OpenGL and DirectX and that AA had been written using
this.
Forgetting so soon, David? "<laugh">
>> but you know - he's kind of gone silent on that
>> point :)
>
>This is a new claim for you.
>
"It is the new 'gamestorm' 64-bit platform that
apparently allows the bootable CD to provide a high-quality gaming
experience without the need for the OpenGL and video drivers for
specific hardware." - DTJ
"Or, to put it another way, if the Gamestorm/America's army thing is
nothing but a recompiled game on Linux running with the XFree 86
Nivdia drivers, what have they been doing for nine months and why are
they claiming it is anything different than the earlier 32-bit game?"
- DTJ
"The last sentence is revealing because it suggests that 'porting' is
required to run other games on the Gamestorm platform which suggests
that Gamestorm does not run OpenGL games." - DTJ
So if it's not an OpenGL/DirectX game then it's not the same thing as
that running on the Unreal engine. Who wrote this new version, David?
>>
>> R.O.T.F.L.M.A.O.@.U!!!
>>
>>
>>
>>>>
>>>>>>[David T Johnsons URLs go here]
>>>>>>
>>>>
>>>>
>>>>Where's your evidence, David?
>>>
>>>There are no details that have been released, yet, on the Gamestorm
>>>platform for the Opteron but I think it is very interesting.
>>>
>>
>>
>> In other words you have no evidence. There is, however, plenty of
>> evidence that shows your "speculation" is ill-informed, where it is
>> not outright stupid.
>
>Incorrect.
>
Oh, how very cutting. If that's all you've got left in the tank then
you are toast :)
>>
>>
>>>>And hey, why aren't you taking me up on that bet? LOL!
>>>>
>>>
>>>There is no reason to wager with you. You would probably be a piker,
>>>anyway.
>>
>>
>> I'm quite serious, David. I bet you $100 that the SCI distribution
>> ships with hardware-specific drivers for video and sound. Put up, or
>> shut up.
>
>There is no reason to wager with you.
>
Cluck cluck cluck. Watch out Chicken Little, the sky is falling!
>[ad hominem snipped]
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
>
>You are not capable of refuting the claims made by SCI about their
>Gamestorm platform because 1) you obviously don't know very much about
>computer graphics
ROTFL! This from the guy claiming that dedicated 3D hardware is no
faster than a general purpose CPU! <shrieks with mirth>
>and 2) you don't know anything about the Gamestorm
>platform.
>
You certainly don't. You don't even know what the Opteron is capable
of.
>>
>>
>>
>>>The last sentence is revealing because it suggests that 'porting' is
>>>required to run other games on the Gamestorm platform which suggests
>>>that Gamestorm does not run OpenGL games.
>>>
>>
>>
>> Bullshit. Americas Army is an OPenGL game as you have already been
>> told.
>
>The 32-bit version is an OpenGL game. We don't know anything about the
>new 64-bit 'Gamestorm' version yet.
>
It is also an OpenGL game as it's built on the same engine. It was
done by the same folks who did the original Linux port. No doubt
*they've* got it wrong too - LOL!
http://icculus.org/
http://www.linux-gamers.net/modules/news/article.php?storyid=32
>>
>> That you ignore the evidence is *your* problem, not mine.
>
>You are ignoring the facts and the discussion points.
>
I'm still waiting for you to present any facts.
Here's a summary of your "points":
"somehow doing something different"
"in some way, as a guess. I am just guessing, though."
"They seem to be doing something"
"which does not seem to be"
"some sort of virtual hardware implementation or something"
ROTFLMFAO! Do you have the big red shoes to go with that clown suit?
>>
>>
>>
>>>> It's *you* that's making all the grandiose claims based upon nothing
>>>>but a vague press release. If you have any evidence to the contrary
>>>>then *present it*
>>>
>>>I am speculating, rather than making claims. There is no information
>>>about the Gamestorm platform yet. You have been told that previously.
>>>
>>
>>
>> And yet you create a thread proclaiming DirectX dead?
>
>Correct.
>
Why?
>>
>> Your "speculation" is the speculation of a clueless fool. Don't
>> expect to shoot your mouth off without being challenged on it.
>
>If these kinds of speculations are not to be permitted by you here in a
>usenet newsgroup, then exactly where would you permit them? <laugh> My
>speculation is probably pretty good but we'll have to wait and see what
>comes out about the Gamestorm platform in the coming weeks.
>
You can make the speculation, but expect to be challenged on it. If
you don't want to look foolish then I suggest you try arguing facts
instead.
>>
>>
>>>>
>>>>>Also, the absence
>>>>>of any specific video hardware requirements for the game seems
>>>>>misleading at best.
>>>>
>>>>
>>>>Misleading to you maybe, but nobody else seems confused by it.
>>>
>>>It would be misleading to anyone who intended to use the game who lacked
>>>the specific hardware that you allege is required.
>>>
>>
>>
>> Which hardware is that? A 3D video card?
>
>Yes, you have alleged that the 64-bit America's Army game will not run
>with acceptable performance unless specific 3D video hardware is
>present. The Gamestorm SCI announcement, OTOH, makes no mention of
>specific hardware requirements.
>
Which doesn't mean that there aren't any. The Americas Army website
makes it quite clear that there *are* hardware requirements.
That's what's known as a fact David.
>>
>>
>>>>
>>>>>It's interesting that you mention Mesa, since the
>>>>>standard Mesa release supports only software rendering which is what I
>>>>>have been suggesting might be the case for the Gamestorm thing.
>>>>>
>>>>
>>>>
>>>>LOL - wow, that thing will be a real dog!
>>>
>>>Time will tell us.
>>>
>>
>>
>> Rational thought and all the evidence has already told us.
>
>Ahhh yes. The same 'evidence' no doubt that your
>similarly-free-thinking ancestors used as a justification for their need
>to burn those 'witches' at the stake.
>
What are you talking about? Are you claiming that the Americas Army
website, and all the websites which have published benchmarks are just
making things up in order to discredit you?
We have *facts* to work with David - that you don't like them doesn't
make this a witch-hunt.
>>
>>
>>>>
>>>>>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>>>>>to be a part of the project.
>>>>>>>
>>>>>>>I don't see why not. Exactly the opposite would seem to be the case
>>>>>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>>>>>graphics implementation.
>>>>>>
>>>>>>
>>>>>>So they would willing cut their own throats by simultaneously making
>>>>>>their hardware irrelevant and assisting their competitors? Perhaps.
>>>>>
>>>>>No one can stay with the same approach forever. Things change very
>>>>>quickly in the computer business. The whole idea of having a dedicated
>>>>>video hardware with its own memory sticking out on the PCI/AGP bus
>>>>>doesn't seem like a long-term thing to me for video.
>>>>
>>>>
>>>>Really? So you think that a general purpose CPU can do things just as
>>>>fast as dedicated silicon? ROTFL! Damn you're funny!
>>>
>>>Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
>>>than the video card processor.
>>
>>
>> LOL!!!! Not when it comes to rendering 3D graphics!
>
>The NV34 GPU in my video card has a clock speed of 325 Mhz which is not
>exactly lightning-speed anymore.
>
So clock-speed is the only thing that matters?
>[snip of ad hominem]
>>
>>>Speed, however, is affected by the
>>>design, overhead from other activities, and the software. Gamestorm
>>>makes no claim to be 'as fast as' a 32-bit platform or to be the
>>>'fastest' or anything like that. They only suggest that their
>>>performance is competitive with other game experiences.
>>>
>>
>>
>> So long as you have an accelerated graphics solution, sure. If you
>> try to run it in software emulation mode then prepare to be *very*
>> disappointed.
>
>We will all find out how disappointed we will be. <laugh> In the
>meantime, DirectX is looking a little green around its gills.
>
Why? Because you say so?
>>
>>
>>>>
>>>>>The Opteron CPU is
>>>>>much more powerful than the video chip
>>>>
>>>>
>>>>Duh! The video chip wipes the floor with the Opteron for 3D
>>>>rendering.
>>>
>>>Reference?
>>>
>>
>>
>> See the render times from the 3D rendering benchmarks I've already
>> pointed you at.
>
>Those are 32-bit applications as I told you previously and they do not
>make any comparison of performance between the video card GPU and the
>Opteron CPU.
>
Because such a comparison is absurd. The fastest result for the
Opteron was 45 seconds to render *ONE* frame.
>>
>>
>>
>>>> There is no comparison. Otherwise, software rendering
>>>>would be just as fast as hardware rendering!
>>>
>>>The ultimate speed is affected by a lot of things other than the video
>>>card chip.
>>>
>>
>>
>> Of course - but that's irrelevant to your claim.
>
>No, it isn't. The Gamestorm platform may have acceptable performance
>without the need for hardware-specific drivers.
>
Might it? How? Magic pixie dust?
>>
>>
>>>>id don't even include software renderers in their games since QuakeII
>>>>because a general purpose CPU can't keep up. Are you saying that
>>>>they are going to start again? ROTFLMAO!
>>>
>>>We'll have to see what Gamestorm is doing.
>>>
>>
>>
>> They have recompiled an OpenGL game and bundled it with Linux. That's
>> what they are doing.
>
>SCI also claims to have 'ported' the the game to their Gamestorm
>platform. They further claim that it is now a 64-bit game, has
>acceptable performance, and runs on Linux.
>
It already has acceptable performance, and already runs on Linux. So
long as you have a 3D accelerator card.
>>
>>
>>>>
>>>>
>>>>>and the Opteron system memory is
>>>>>much faster and has much greater bandwidth than the video memory.
>>>>
>>>>
>>>>DUH!
>>>>
>>>>9800Pro = 21.8GB/s
>>>>Athlon64 = 6.4GB/sec
>>>
>>>The Opteron has the memory controller integrated with the Opteron
>>>processor which means that the front side bus (interface to memory) runs
>>>at the speed of the processor and has a front side bandwidth of 14.4
>>>GB/sec.
>>>
>>
>>
>> a) you are wrong about the speed
>
>No.
>
Yes.
http://www.harddata.com/Systems/Opteron/
The CPU can only handle 5.4 GB/s. As you add more memory controllers
(max three) you can get close to your figure - but the cost is
extreme.
The Athlon64 increases that to 6.4GB/s per controller.
>> b) it's still MUCH slower than the Radeon
>
>But not slower than most current video cards which have memory bandwidth
>in the 10-15 GB/sec range.
>
It is still MUCH slower than the Radeon which *IS* a current video
card.
>> c) reference?
>
>Google is your friend.
>
Try it sometime.
>>
>> Your claim above is utterly false, David. You even admit it with
>> your own made-up figures! ROTFL!
>
>Incorrect.
>
"the Opteron system memory is much faster and has much greater
bandwidth than the video memory." - DTJ
9800Pro = 21.8GB/s
Three way Opteron system = 14.4GB/sec
You were wrong, no matter how you try and favour the Opteron.
>>
>>
>>>The more significant thing, however, is that that the Opteron system
>>>will typically have 1GB or more of memory (and a 1MB cache) while a
>>>high-performance video card will only have perhaps 128 MB of memory.
>>
>>
>> Blah blah blah - you don't *need* a gig of memory to render a scene,
>> David.
>
>More memory allows many more things to be done, in particular much more
>detail can be provided at a higher framerate.
Explain how this works. In order to increase framerate it's the
*speed* that is the deciding factor once you have enough RAM to hold
all your textures.
> That is likely to be one
>of the features of the Gamestorm 64-bit platform.
Is it? Reference?
>The point is that
>there is never too much memory.
>
Sure - but there is such a thing as having enough for the task.
>
>> And your claim that the Opteron has faster memory is quite
>> simply FALSE.
>
>Incorrect.
>
Liar.
9800Pro = 21.8GB/s
Three way Opteron system = 14.4GB/sec
Do you not *care* that you are proving your dishonesty?
>>
>>
>>
>>>It
>>>seems possible that a system with such a large high-speed memory could
>>>therefore render more images in memory and maintain an acceptabe
>>>framerate by working ahead.
>>
>>
>> So now your CPU is rendering MULTIPLE frames simultaneously with
>> *less* bandwidth than a video card? All while it also handles the
>> game engine?
>
>'Integrated' graphics setups on current 32-bit systems use the system
>memory for video memory. You did know that didn't you?
Who was talking about integrated solutions? Right above you are
talking about video cards only having 128MB RAM. Quit trying to
change the topic when you know you've lost this one, loser.
>A 64-bit CPU
>could certainly provide better video performance than a 32-bit CPU if
>64-bit software was used.
>
That is irrelevant to the question of whether it is faster than a
dedicated hardware device.
>>
>[ad hominem snipped]
Because it makes you look foolish.
>>
>>
>>>The main limit for such a hypothetical
>>>system doing software rendering would probably be the speed of the AGP
>>>bus to transfer the rendered images to the video memory. Gamestorm
>>>hasn't claimed that their system is faster than all others but only has
>>>suggested that its performance is competitive. It seems possible that
>>>they could be doing some sort of software rendering although no
>>>information has been provided.
>>>
>>
>>
>[ad hominem snipped]
>>
>>
>>>[personal comment snipped]
>>>
>>>>
>>>>>Seems
>>>>>like software rendering could ultimately provide better performance when
>>>>>the software is developed, even if Gamestorm has not done that.
>>>>
>>>>
>>>[personal comment snipped]
>>
>>
>[ad hominem snipped]
>>
>> When are you going to take me up on my bet? ;)
>>
>When pigs fly? <laugh>
Cluck cluck cluck. You are chickenshit.
David Sutherland
<djoh...@isomedia.com> wrote:
So what *are* you claiming then, David? That the Opteron is *not*
much faster for 3D graphics than 32bit CPUs?
That blows your whole argument out of the water.
>> You have claimed that the 3D gaming performance of the
>> Opteron - with no additional hardware - is going to kill DirectX.
>
>No, I did not.
Look at your claims in this thread, David, and then wipe the drool off
your chin.
> I claimed that the Gamestorm platform inflicted a mortal
>wound on DirectX due to the ability of the Gamestorm platform to boot
>the OS and game off of a CD with no installation on the hard disk
>required. DirectX is a miserable kludgy piece of software whose only
>purpose for even existing is to allow high performance games to run
>while Windows is also booted. DirectX (and Direct3D) was supposed to
>give them 'direct' access to the hardware to improve performance. If
>people are running games from a CD with Linux rather than booting from
>the hard drive, they will not be using DirectX and DirectX is dead. Get it?
>
No. People are not going to ditch Windows just so they can have the
joy of losing their familiar computing environment every time they
want to play a game - especially when that game is already available
for Windows.
>
>> This is patently stupid, but it *is* your claim.
>
>Actually, I think it's pretty good.
>
Your capacity for self-congratulation has not gone un-noticed.
>>
>> You *did* create this very thread entitled "Death blow for Windows
>> DirectX", after all.
>
>One of the few correct things you have managed to write among your mound
>of gibberish.
>
I understand why you are so angry about being proved wrong with facts,
David. It's okay :)
>>
>>
>>>The benchmarks you you provide below, however,
>>>are for the Opteron running 32-bit software rather than 64-bit software.
>>> The Gamestorm platform is apparently 64-bit.
>>>
>>
>>
>> Given that existing CPU's are most certainly not capable of rendering
>> 3D with anything like the same speed or quality of hardware boards,
>
>I don't accept that as a given.
You don't accept a lot of things - that doesn't make the facts of the
matter any less true.
>There is nothing special about hardware
>boards that cannot be done with the CPU.
LOL - sure you can just *emulate* all the custom single-pass rendering
hardware!
>The clock speed of the NV34
>GPU in my video card is only 325 Mhz which is not earth-shattering and
>pretty much the only thing these video processors do is floating point
>calculations.
>
Oh. My. God. You really *are* stupid.
>> how else do you explain your claim that such hardware is now obsolete?
>
>I certainly haven't claimed that all video hardware is obsolete. I have
>suggested that in its current form, it is not likely to last. I am also
>suggesting that standards will make hardware-specific drivers
>unnecessary. Video hardware, in some form, will always be needed,
>although RAMDAC chips will probably be gone fairly soon. Also, I am
>speculating, rather than making claims. There is no information about
>the Gamestorm platform yet. You have been told that previously.
>
If there is no information, on what basis are you making all your
claims?
>>
>> Can you provide *any* evidence that the speed of 64bit render software
>> is significantly (as in more than twice) faster than that of 32bit?
>
>There is no information about the Gamestorm platform performance yet.
>You have been told that previously.
>
If there is no information, on what basis are you making all your
claims?
>
>> There are certainly plenty of 64bit benchmarks which show that you can
>> indeed get up to almost twice the throughput over 32bits for *some*
>> types of processing. But this is worlds away from the 20-30 fold
>> increase you would need in order to render complex scenes in realtime
>> for a game!
>
>You don't have a clue as to what the performance requirements would be.
>
You can't make up your mind if there *are* any performance increases.
>>
>>
>>
>>>>how come banchmarks show it often rendering 3D
>>>>scenes *slower* than the 32bit Athlon?
>>>>
>>>>http://www.tomshardware.com/cpu/20030422/opteron-23.html
>>>>
>>>>http://www.extremetech.com/article2/0,3973,1217390,00.asp
>>>>
>>>>More magic?
>>
>>
>> LOL - you can't answer, can you?
Nope, didn't think so.
MMI
<snip>
> > Given that existing CPU's are most certainly not capable of rendering
> > 3D with anything like the same speed or quality of hardware boards,
>
> I don't accept that as a given. There is nothing special about hardware
> boards that cannot be done with the CPU. The clock speed of the NV34
> GPU in my video card is only 325 Mhz which is not earth-shattering and
> pretty much the only thing these video processors do is floating point
> calculations.
<snip>
Well, I tried good ol' Duke3D for DOS (which uses NO HW 3D
acceleration) on my 1333MHz Athlon Thunderbird with 800x600 and in
some cases it was DAMN SLOW. Videocard in my machine was Matrox G400
4xAGP. Such an old game, which should FLY on CPU like mine. HW board
is HW board, no matter how many MHZs it has. It will beat almost any
general-purpose CPU of its time on 3D operations.
Cheers,
Martin
David T. Johnson
Running a game on the 64-bit "Gamestorm" platform running on 64-bit
Linux is obviously different than running an OpenGL game on 32-bit
Linux, if we are to believe the Gamestorm announcement. There have not
yet been any subsequent details so we don't know the specifics. You are
pointing to hardware and software requirements for running the 32-bit
OpenGL version of America's Army on Linux that obviously do not apply to
the Gamestorm platform and are therefore irrelevant. For example, the
requirements you list say that the game will run acceptably on a 733 Mhz
PIII while the Gamestorm platform *requires* a 64-bit Opteron.
>
> If you don't want to believe the people who know that engine better
> than anyone else then it's obvious that you *choose* not to believe
> them. That isn't surprising, but it does make clear just how low you
> are prepared to sink rather than admit that you might be wrong.
We don't even know what the "Gamestorm" 'engine' is yet and we do not
know what the video hardware requirements will be for Gamestorm. If we
did, they would have been listed in this thread by now. Your pathetic
attempt to list the 32-bit OpenGL hardware requirements and claim that
they are also the hardware requirements for Gamestorm is an example of
your complete lack of understanding of the technical aspects.
>
>
>>>Here's some more for chuckles:
>>>
>>>Q: I have Intel or Intel3D integrated graphics , can I run the game?
>>>A: No, you can't. Please read the minimum system requirements.
>>> (Last Updated: 2003-08-21)
>>>
>>>Wondering what those might be? Knew you'd ask!
>>>
>>>http://www.americasarmy.com/faq_linux.php?p=3&t=4
>>>Q: What are the minimum system requirements for the Linux v1.9 Client?
>>>A:
>>>Pentium3 766 (or equivalent)
>>>128MB RAM
>>>glibc 2.1 or newer.
>>>Hardware accelerated OpenGL drivers (Nvidia's and XiG's are known)
>>
>>This is referring to the current 32-bit version and not the new 64-bit
>>Gamestorm version. The Gamestorm 64-bit version specifically states
>>that it requires the 64-bit Opteron CPU.
>>
>
>
> How is that relevant? How many times faster do you think the Opteron
> is in 64-bit mode? Two? Two hundred? Where are the benchmarks!
No Gamestorm benchmarks are available nor is specific information on the
Gamestorm platform available yet. You have been told that several times
previously.
>
>
>>>
>>>Q: I have a Kyro, KyroII, or Hercules 'prophet' branded video card,
>>>can I run the game?
>>>A: No, you can't. Please read the minimum system requirements. While
>>>the kyro chipsets are recent and have the neccessary amount of video
>>>ram, they do not have some of the transform and lighting features
>>>required to show the high quality rendering that this game uses.
>>>==
>>>
>>>Face facts, David - your claims are complete nonsense. Even the
>>>people whose product you are trying to use as evidence directly
>>>contradict everything you say.
>>>
>>>Fact - the game needs an OpenGL accelerated card.
>>
>>That is not a fact, nor even a reasonable supposition.
>>
>
>
> See the official website, David. Your personal speculations do not
> outweigh it's assertions of fact. *THEY* know the game better than
> you do.
They have not provided any specific information yet on the 64-bit
Gamestorm port of the game. You have been told that several times
previously. One thing we can be certain of is that the capabilities of
the 64-bit port will not be the same as the 32-bit version. At a
minimum, I would speculate that it will have more detail in the frames
than it does in the 32-bit version, making the game more realistic.
>
>
>
>>>Fact - the Opteron (or any other CPU) cannot perform the same
>>>functions as dedicated 3D hardware in an acceptable manner.
>>
>>Also not a fact. The Opteron CPU is much more powerful than a typical
>>graphics card GPU.
>>
>
>
> This is nonsense. If you really believe this then you'll be willing
> to take my money over it, right?
>
> US$100 says that the Opteron cannot render 3D games faster (or even
> within 20%) the speed of the Radeon9800. Put up or shut up.
Neither the Opteron CPU nor a graphics card GPU 'render 3D games.' They
do floating point calculations and the Opteron is much more powerful
that a typical graphics card GPU.
>
>
>
>>>Fact - if your card isn't fast enough, you can't play (so much for
>>>hardware independence!)
>>
>>Gamestorm did not list any specific video hardware requirements for
>>their new game. You have been told this previously.
>>
>>
>>>Fact - you have been talking crap.
>>>
>>
>>That is certainly true for you.
>
>
>
[snip of ad hominem]
--
David T. Johnson
I would speculate that the 64-bit Gamestorm games will generally have
more detail in the frames (i.e. many more polygons) due to the huge
amount of memory that will be available and will also have a better
frame rate that the 32-bit OpenGL games. We don't know what the video
hardware requirements since those requirements were conspicuously absent
from the announcement. We do know that the Gamestorm platform will
allow the OS and the game to be booted directly from the CD with no
'installation' required on the hard disk. Sweet. ...and goodbye to
DirectX...see the title of the thread. <laugh>
[ad hominem snipped]
No, we cannot use Opteron benchmarks and un-related 64-bit software
benchmarks to predict the performance of the Gamestorm platform. The
fact that you think this is a reasonable thing to do is one more example
of lack on knowledge.
>
>
>>> Unless of course you are looking to
>>>play Tetris!
>>
>>My old DOS computer will play Doom and Castle Wolfenstein (both early 3D
>>games) without the need for accelerated 3D video cards or even a
>>powerful CPU. Both of these games are much more sophisticated that your
>>Tetris example. <laugh>
>>
>
> I'm really starting to wonder if you've even *seen* a modern 3D game.
> Castle Wolfenstein? ROTFLMAO! That was released in 1981
Well, the first IBM PC was released in 1981 but there was no "Castle
Wolfenstein" in 1981. Microsoft "Decathalon" perhaps? <laugh>
> - I would
> sure hope your PC can run it!
>
> Maybe you meant Wolfenstein3D? If so - check out the screenshot here:
>
> http://members.chello.at/theodor.lauppert/games/wolf3d.htm
>
> Wow, such cutting edge graphics!
I was referring to Doom and "Wolfenstein 3D" which was released in 1992.
The 3D graphics on these were "cutting edge" in 1992. Both
Wolfenstein and Doom will run acceptably on an old DOS computer without
any "3D acceleration" or powerful CPU.
[ad hominem snipped]
>
>
>>> Americas Army is a FPS, and will run like a dog without
>>>hardware acceleration, if it runs at all.
>>
>>Time will tell...
>>
>
>
> The developers have stated it as a matter of public record.
Incorrect. No performance information nor hardware requirements have
yet been provided for the 64-bit version of America's Army or the
Gamestorm platform. You have been told this several times previously.
> We don't
> need to guess.
In the absence of hard information, we can only speculate.
It is irrelevant for several reasons that should be obvious to you.
>
>
>>>
>>>>>and they would DIE if
>>>>>they did what you are claiming. If you believe otherwise it's just
>>>>>another example of how ignorant you are about How Things Work.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>>>Or tell you what - why don't you try proving it instead of making yet
>>>>>>>more moronic speculations:
>>>>>>
>>>>>>There are no details that have been released, yet, on the Gamestorm
>>>>>>platform for the Opteron but I think it is very interesting.
>>>>>>
>>>>>
>>>>>
>>>>>I think it's very interesting that without any details at all you
>>>>>create entire fantasy worlds where the laws of physics do not apply,
>>>>
>>>>No laws of physics are violated by software rendering. <laugh>
>>>>
>>>Is *that* what you understood me to have said? Guess your RCP (aka
>>>deliberate avoidance of the argument) is starting to kick in.
>>
>>David S: "...without any details at all you create entire fantasy
>>worlds where the laws of physics do not apply" <chuckle>
>>
>
> Claiming that an Opteron which may have to perform many hundreds of
> calculations to achieve a particular result is as fast or faster than
> dedicated hardware which can do it in one pass *is* ignoring the rules
> of physics.
Both the Opteron and the GPU do floating point calculations one step at
a time. Well, actually the Opteron does several on each clock tick
thanks to its triple-pipeline FPU but that is another issue. No laws of
physics have been violated.
>
> Claiming that doubling the number of bits renders hardware drivers
> obsolete is ignoring the rules of physics.
<laugh> Hardware-specific drivers may be obsoleted by improved CPU
performance which will allow video standards such as VBE 3.0 to provide
acceptable performance. No laws of physics would be violated.
>
> Which of these are you having trouble following?
>
>
>>>
>>>>>and then embarrass yourself by pretending they are real in this ng.
>>>>>What is that about?
>>>>
>>>>You are the guy posting about violations of the 'laws of physics.'
>>>>You're sounding unusually nutty today.
>>>>
>>>
>>>
>>>Apparently you are having trouble following the conversation. Ain't
>>>that a surprise.
>>
>>See above.
>>
>
>
> Where?
>
>
>>>If you want "nutty", check out the stupid claims of the guy who
>>>started this thread. He thinks that video acceleration is obsolete,
>>
>>Incorrect. See above.
>>
>
>
> Yawn. You claimed that the 64bit Opteron does not need
> hardware-specific drivers and can perform 3D ops as fast as dedicated
> graphics hardware.
Well, to be accurate, I have not made any claims about the speed of the
64-bit Gamestorm platform relative to the 32-bit Open GL games other
than to refer to the Gamestorm announcement which suggests that the
performance will be competitive.
>
>
>>>that the Opteron needs no drivers to address any hardware now or in
>>>the future,
>>
>>Incorrect. You have continued to make this false statement. I was
>>referring to 'hardware-specific' drivers which are probably on the way
>>out, thanks to new video standards.
>>
>
>
> Define "hardware-specific."
Specific to a particular piece of video hardware.
[ad hominem snipped]
>
>
>>>and that a company called SCI has developed a brand-new
>>>platform independent competitor to DirectX and OpenGL - all in the
>>>space of a couple of months! Initially he also claimed that they
>>>wrote Americas Army,
>>
>>Incorrect. The authorship of AA has not been discussed.
>>
>
>
> You claimed that SCI created an entirely new platform-agnostic API to
> replace both OpenGL and DirectX
Incorrect. I speculated that the 64-bit Gamestorm platform did not have
specific video hardware requirements since none were listed in the
announcement.
> and that AA had been written using
> this.
AA has obviously been ported to the 64-bit Gamestorm platform if we are
to believe the announcement. I have not discussed the authors of AA.
>
> Forgetting so soon, David? "<laugh">
See above.
>
>
>>>but you know - he's kind of gone silent on that
>>>point :)
>>
>>This is a new claim for you.
>>
>
>
> "It is the new 'gamestorm' 64-bit platform that
> apparently allows the bootable CD to provide a high-quality gaming
> experience without the need for the OpenGL and video drivers for
> specific hardware." - DTJ
> "Or, to put it another way, if the Gamestorm/America's army thing is
> nothing but a recompiled game on Linux running with the XFree 86
> Nivdia drivers, what have they been doing for nine months and why are
> they claiming it is anything different than the earlier 32-bit game?"
> - DTJ
> "The last sentence is revealing because it suggests that 'porting' is
> required to run other games on the Gamestorm platform which suggests
> that Gamestorm does not run OpenGL games." - DTJ
>
> So if it's not an OpenGL/DirectX game then it's not the same thing as
> that running on the Unreal engine. Who wrote this new version, David?
I have not speculated about that. I suspect that the 64-bit Gamestorm
port was done by SCI but I don't know nor do I care.
>
>
>>>R.O.T.F.L.M.A.O.@.U!!!
>>>
>>>
>>>
>>>
>>>>>>>[David T Johnsons URLs go here]
>>>>>>>
>>>>>
>>>>>
>>>>>Where's your evidence, David?
>>>>
>>>>There are no details that have been released, yet, on the Gamestorm
>>>>platform for the Opteron but I think it is very interesting.
>>>>
>>>
>>>
>>>In other words you have no evidence. There is, however, plenty of
>>>evidence that shows your "speculation" is ill-informed, where it is
>>>not outright stupid.
>>
>>Incorrect.
>>
>
>
[ad hominem snipped]
>
>
>>>
>>>>>And hey, why aren't you taking me up on that bet? LOL!
>>>>>
>>>>
>>>>There is no reason to wager with you. You would probably be a piker,
>>>>anyway.
>>>
>>>
>>>I'm quite serious, David. I bet you $100 that the SCI distribution
>>>ships with hardware-specific drivers for video and sound. Put up, or
>>>shut up.
>>
>>There is no reason to wager with you.
>>
>
>
[ad hominem snipped]
>
>
>>[ad hominem snipped]
David T. Johnson
> On Tue, 07 Oct 2003 15:48:17 -0700, "David T. Johnson"
> <djoh...@isomedia.com> wrote:
>
> [snip]
>
>
>>You are not capable of refuting the claims made by SCI about their
>>Gamestorm platform because 1) you obviously don't know very much about
>>computer graphics
>
>
> ROTFL! This from the guy claiming that dedicated 3D hardware is no
> faster than a general purpose CPU! <shrieks with mirth>
Actually, I compared the GPU with the CPU. The CPU is obviously much
more powerful. Don't know why you would find that humorous.
>
>>and 2) you don't know anything about the Gamestorm
>>platform.
>>
>
>
> You certainly don't. You don't even know what the Opteron is capable
> of.
I am very familiar with the Opteron because I am using one. We don't
know the specifics of the Gamestorm platform because they have not been
released yet. You have been told that several time previously.
>
>
>>>
>>>
>>>>The last sentence is revealing because it suggests that 'porting' is
>>>>required to run other games on the Gamestorm platform which suggests
>>>>that Gamestorm does not run OpenGL games.
>>>>
>>>
>>>
>>>Bullshit. Americas Army is an OPenGL game as you have already been
>>>told.
>>
>>The 32-bit version is an OpenGL game. We don't know anything about the
>>new 64-bit 'Gamestorm' version yet.
>>
>
>
> It is also an OpenGL game as it's built on the same engine.
We don't know any of the details about the 64-bit port of AA running on
the Gamestorm platform for Opteron. You have been told that several
times previously.
> It was
> done by the same folks who did the original Linux port.
I don't think we even know that, though you could actually be correct.
> No doubt
> *they've* got it wrong too - LOL!
>
> http://icculus.org/
> http://www.linux-gamers.net/modules/news/article.php?storyid=32
I see no mention of the 64-bit Gamestorm port of America's Army.
>
>
>>>That you ignore the evidence is *your* problem, not mine.
>>
>>You are ignoring the facts and the discussion points.
>>
>
>
> I'm still waiting for you to present any facts.
>
> Here's a summary of your "points":
>
> "somehow doing something different"
> "in some way, as a guess. I am just guessing, though."
> "They seem to be doing something"
> "which does not seem to be"
> "some sort of virtual hardware implementation or something"
>
[ad hominem snipped]
We don't know what the video hardware requirements are for the 64-bit
version of America's Army running on Gamestorm since those were not
listed. It might be as simple as "any VBE 3.0-compliant" video card."
>
> That's what's known as a fact David.
That is what is known as the "absence of facts" which is undoubtedly a
normal condition for your arguments, judging by what you say here.
>
>
>>>
>>>>>>It's interesting that you mention Mesa, since the
>>>>>>standard Mesa release supports only software rendering which is what I
>>>>>>have been suggesting might be the case for the Gamestorm thing.
>>>>>>
>>>>>
>>>>>
>>>>>LOL - wow, that thing will be a real dog!
>>>>
>>>>Time will tell us.
>>>>
>>>
>>>
>>>Rational thought and all the evidence has already told us.
>>
>>Ahhh yes. The same 'evidence' no doubt that your
>>similarly-free-thinking ancestors used as a justification for their need
>>to burn those 'witches' at the stake.
>>
>
>
> What are you talking about? Are you claiming that the Americas Army
> website, and all the websites which have published benchmarks are just
> making things up in order to discredit you?
No information has been provided about the 64-bit version of America's
Army running on Gamestorm for Opteron. You have been told that several
times previously.
>
> We have *facts* to work with David - that you don't like them doesn't
> make this a witch-hunt.
No information has been provided about the 64-bit version of America's
Army running on Gamestorm for Opteron. You have been told that several
times previously.
>
>
>>>
>>>>>>>>>No, you are ignorant. If the Opteron had somehow magically done away
>>>>>>>>>with the need for video drivers, there would have no reason for NVIDIA
>>>>>>>>>to be a part of the project.
>>>>>>>>
>>>>>>>>I don't see why not. Exactly the opposite would seem to be the case
>>>>>>>>IMO. Nvidia could provide valuable assistance in developing a software
>>>>>>>>graphics implementation.
>>>>>>>
>>>>>>>
>>>>>>>So they would willing cut their own throats by simultaneously making
>>>>>>>their hardware irrelevant and assisting their competitors? Perhaps.
>>>>>>
>>>>>>No one can stay with the same approach forever. Things change very
>>>>>>quickly in the computer business. The whole idea of having a dedicated
>>>>>>video hardware with its own memory sticking out on the PCI/AGP bus
>>>>>>doesn't seem like a long-term thing to me for video.
>>>>>
>>>>>
>>>>>Really? So you think that a general purpose CPU can do things just as
>>>>>fast as dedicated silicon? ROTFL! Damn you're funny!
>>>>
>>>>Yes, a 'general purpose' CPU like Opteron is *much* more powerful than
>>>>than the video card processor.
>>>
>>>
>>>LOL!!!! Not when it comes to rendering 3D graphics!
>>
>>The NV34 GPU in my video card has a clock speed of 325 Mhz which is not
>>exactly lightning-speed anymore.
>>
>
>
> So clock-speed is the only thing that matters?
No, of course not, but it is one measure of comparison between processors.
>
>
>>[snip of ad hominem]
>>
>>>>Speed, however, is affected by the
>>>>design, overhead from other activities, and the software. Gamestorm
>>>>makes no claim to be 'as fast as' a 32-bit platform or to be the
>>>>'fastest' or anything like that. They only suggest that their
>>>>performance is competitive with other game experiences.
>>>>
>>>
>>>
>>>So long as you have an accelerated graphics solution, sure. If you
>>>try to run it in software emulation mode then prepare to be *very*
>>>disappointed.
>>
>>We will all find out how disappointed we will be. <laugh> In the
>>meantime, DirectX is looking a little green around its gills.
>>
>
>
> Why? Because you say so?
>
>
>>>
>>>>>>The Opteron CPU is
>>>>>>much more powerful than the video chip
>>>>>
>>>>>
>>>>>Duh! The video chip wipes the floor with the Opteron for 3D
>>>>>rendering.
>>>>
>>>>Reference?
>>>>
>>>
>>>
>>>See the render times from the 3D rendering benchmarks I've already
>>>pointed you at.
>>
>>Those are 32-bit applications as I told you previously and they do not
>>make any comparison of performance between the video card GPU and the
>>Opteron CPU.
>>
>
>
> Because such a comparison is absurd. The fastest result for the
> Opteron was 45 seconds to render *ONE* frame.
No information has been provided about the 64-bit version of America's
Army running on Gamestorm for Opteron. You have been told that several
times previously.
>
>
>
>>>
>>>
>>>>> There is no comparison. Otherwise, software rendering
>>>>>would be just as fast as hardware rendering!
>>>>
>>>>The ultimate speed is affected by a lot of things other than the video
>>>>card chip.
>>>>
>>>
>>>
>>>Of course - but that's irrelevant to your claim.
>>
>>No, it isn't. The Gamestorm platform may have acceptable performance
>>without the need for hardware-specific drivers.
>>
>
>
> Might it? How? Magic pixie dust?
No information has been provided about the 64-bit version of America's
Army running on Gamestorm for Opteron. You have been told that several
times previously.
>
>
>>>
>>>>>id don't even include software renderers in their games since QuakeII
>>>>>because a general purpose CPU can't keep up. Are you saying that
>>>>>they are going to start again? ROTFLMAO!
>>>>
>>>>We'll have to see what Gamestorm is doing.
>>>>
>>>
>>>
>>>They have recompiled an OpenGL game and bundled it with Linux. That's
>>>what they are doing.
>>
>>SCI also claims to have 'ported' the the game to their Gamestorm
>>platform. They further claim that it is now a 64-bit game, has
>>acceptable performance, and runs on Linux.
>>
>
>
> It already has acceptable performance, and already runs on Linux. So
> long as you have a 3D accelerator card.
I would speculate that the 64-bit version will provide more detail in
the frames and perhaps an improved frame rate at a higher screen
resolution. No information has been provided about the 64-bit version of
America's Army running on Gamestorm for Opteron. You have been told
that several times previously.
>
>
>>>
>>>>>
>>>>>>and the Opteron system memory is
>>>>>>much faster and has much greater bandwidth than the video memory.
>>>>>
>>>>>
>>>>>DUH!
>>>>>
>>>>>9800Pro = 21.8GB/s
>>>>>Athlon64 = 6.4GB/sec
>>>>
>>>>The Opteron has the memory controller integrated with the Opteron
>>>>processor which means that the front side bus (interface to memory) runs
>>>>at the speed of the processor and has a front side bandwidth of 14.4
>>>>GB/sec.
>>>>
>>>
>>>
>>>a) you are wrong about the speed
>>
>>No.
>>
>
>
> Yes.
>
> http://www.harddata.com/Systems/Opteron/
>
> The CPU can only handle 5.4 GB/s. As you add more memory controllers
> (max three) you can get close to your figure - but the cost is
> extreme.
>
> The Athlon64 increases that to 6.4GB/s per controller.
Google is your friend.
>
>
>
>>>b) it's still MUCH slower than the Radeon
>>
>>But not slower than most current video cards which have memory bandwidth
>>in the 10-15 GB/sec range.
>>
>
>
> It is still MUCH slower than the Radeon which *IS* a current video
> card.
>
>
>>>c) reference?
>>
>>Google is your friend.
>>
>
>
> Try it sometime.
>
>
>>>Your claim above is utterly false, David. You even admit it with
>>>your own made-up figures! ROTFL!
>>
>>Incorrect.
>>
>
>
> "the Opteron system memory is much faster and has much greater
> bandwidth than the video memory." - DTJ
>
> 9800Pro = 21.8GB/s
> Three way Opteron system = 14.4GB/sec
>
> You were wrong, no matter how you try and favour the Opteron.
I don't 'favor' the Opteron but I am interested in what its 64-bit
capabilities might bring to a 64-bit version of games such as the
Gamestorm platform. The Opteron has the memory controller integrated
with the Opteron processor which means that the front side bus
(interface to memory) runs at the speed of the processor and has a front
side bandwidth of 14.4 GB/sec at a processor speed of 1.8 Ghz. The
bandwidth increases to 16 GB/sec at a processor speed of 2.0 Ghz. You
are apparently referring to the 'inter-processor' bandwidth which is 6.4
GB/sec for the Opteron.
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8796_8799,00.html
>
>
>>>
>>>>The more significant thing, however, is that that the Opteron system
>>>>will typically have 1GB or more of memory (and a 1MB cache) while a
>>>>high-performance video card will only have perhaps 128 MB of memory.
>>>
>>>
>>>Blah blah blah - you don't *need* a gig of memory to render a scene,
>>>David.
>>
>>More memory allows many more things to be done, in particular much more
>>detail can be provided at a higher framerate.
>
>
> Explain how this works. In order to increase framerate it's the
> *speed* that is the deciding factor once you have enough RAM to hold
> all your textures.
An Opteron can address a quite large amount of memory at very good
speed. This will make new capabilities available for 64-bit versions of
games.
>
>
>>That is likely to be one
>>of the features of the Gamestorm 64-bit platform.
>
>
> Is it? Reference?
No information has been provided about the 64-bit version of America's
Army running on Gamestorm for Opteron. You have been told that several
times previously.
>
>
>>The point is that
>>there is never too much memory.
>>
>
>
> Sure - but there is such a thing as having enough for the task.
>
>
>>> And your claim that the Opteron has faster memory is quite
>>>simply FALSE.
>>
>>Incorrect.
>>
>
>
> Liar.
>
> 9800Pro = 21.8GB/s
> Three way Opteron system = 14.4GB/sec
>
> Do you not *care* that you are proving your dishonesty?
See above.
>
>
>>>
>>>
>>>>It
>>>>seems possible that a system with such a large high-speed memory could
>>>>therefore render more images in memory and maintain an acceptabe
>>>>framerate by working ahead.
>>>
>>>
>>>So now your CPU is rendering MULTIPLE frames simultaneously with
>>>*less* bandwidth than a video card? All while it also handles the
>>>game engine?
>>
>>'Integrated' graphics setups on current 32-bit systems use the system
>>memory for video memory. You did know that didn't you?
>
>
> Who was talking about integrated solutions? Right above you are
> talking about video cards only having 128MB RAM. Quit trying to
> change the topic when you know you've lost this one, loser.
<laugh> You were pointing to memory bandwidth as an advantage of video
hardware so I pointed to the memory used by integrated graphics setups
on some systems. Their bandwidth is much less than an Opteron and yet
they are able to run some 3D games with acceptable performance, although
they are nowhere near as good as a dedicated video card (on a 32-bit
system).
>
>
>
>>A 64-bit CPU
>>could certainly provide better video performance than a 32-bit CPU if
>>64-bit software was used.
>>
>
>
> That is irrelevant to the question of whether it is faster than a
> dedicated hardware device.
>
>
>>[ad hominem snipped]
>
>
[ad hominem snipped]
>
>
>>>
>>>>The main limit for such a hypothetical
>>>>system doing software rendering would probably be the speed of the AGP
>>>>bus to transfer the rendered images to the video memory. Gamestorm
>>>>hasn't claimed that their system is faster than all others but only has
>>>>suggested that its performance is competitive. It seems possible that
>>>>they could be doing some sort of software rendering although no
>>>>information has been provided.
>>>>
>>>
>>>
>>[ad hominem snipped]
>>
>>>
>>>>[personal comment snipped]
>>>>
>>>>
>>>>>>Seems
>>>>>>like software rendering could ultimately provide better performance when
>>>>>>the software is developed, even if Gamestorm has not done that.
>>>>>
>>>>>
>>>>[personal comment snipped]
>>>
>>>
>>[ad hominem snipped]
>>
>>>When are you going to take me up on my bet? ;)
>>>
>>
>>When pigs fly? <laugh>
>
[ad hominem snipped]
David T. Johnson
I have pointed to the Gamestorm/SCI announcement which suggests that
performance will be competitive with current 32-bit systems. No claim
has been made by myself or others that there will be an "incredible boost."
>
> That blows your whole argument out of the water.
Incorrect.
>
>
>>> You have claimed that the 3D gaming performance of the
>>>Opteron - with no additional hardware - is going to kill DirectX.
>>
>>No, I did not.
>
>
[ad hominem snipped]
>
>
>>I claimed that the Gamestorm platform inflicted a mortal
>>wound on DirectX due to the ability of the Gamestorm platform to boot
>>the OS and game off of a CD with no installation on the hard disk
>>required. DirectX is a miserable kludgy piece of software whose only
>>purpose for even existing is to allow high performance games to run
>>while Windows is also booted. DirectX (and Direct3D) was supposed to
>>give them 'direct' access to the hardware to improve performance. If
>>people are running games from a CD with Linux rather than booting from
>>the hard drive, they will not be using DirectX and DirectX is dead. Get it?
>>
>
>
> No. People are not going to ditch Windows just so they can have the
> joy of losing their familiar computing environment every time they
> want to play a game - especially when that game is already available
> for Windows.
People will ditch Windows for running games in a heartbeat if they can
run games without the need to 'install' on their hard drive or deal with
all of the DirectX and .DLL nonsense of the Windows platform. I can't
think of a worse platform for running high performance games than
Windows. Windows would *never* be used for game except that it is the
OS that is used by something like 99 percent of all computers. Face it.
DirectX is dead if Gamestorm can do what they claim and the only
person who will cry about its demise will be a few people like yourself.
>
>
>
>>>This is patently stupid, but it *is* your claim.
>>
>>Actually, I think it's pretty good.
>>
>
>
> Your capacity for self-congratulation has not gone un-noticed.
>
>
>>>You *did* create this very thread entitled "Death blow for Windows
>>>DirectX", after all.
>>
>>One of the few correct things you have managed to write among your mound
>>of gibberish.
>>
>
>
> I understand why you are so angry about being proved wrong with facts,
> David. It's okay :)
You have not proved anything yet with any of your 'facts.'
>
>
>>>
>>>>The benchmarks you you provide below, however,
>>>>are for the Opteron running 32-bit software rather than 64-bit software.
>>>>The Gamestorm platform is apparently 64-bit.
>>>>
>>>
>>>
>>>Given that existing CPU's are most certainly not capable of rendering
>>>3D with anything like the same speed or quality of hardware boards,
>>
>>I don't accept that as a given.
>
>
> You don't accept a lot of things - that doesn't make the facts of the
> matter any less true.
>
>
>>There is nothing special about hardware
>>boards that cannot be done with the CPU.
>
>
> LOL - sure you can just *emulate* all the custom single-pass rendering
> hardware!
yes, with a performance penalty, depending on the implementation.
>
>
>>The clock speed of the NV34
>>GPU in my video card is only 325 Mhz which is not earth-shattering and
>>pretty much the only thing these video processors do is floating point
>>calculations.
>>
>
>
[ad hominem snipped]
>
>
>>>how else do you explain your claim that such hardware is now obsolete?
>>
>>I certainly haven't claimed that all video hardware is obsolete. I have
>>suggested that in its current form, it is not likely to last. I am also
>>suggesting that standards will make hardware-specific drivers
>>unnecessary. Video hardware, in some form, will always be needed,
>>although RAMDAC chips will probably be gone fairly soon. Also, I am
>>speculating, rather than making claims. There is no information about
>>the Gamestorm platform yet. You have been told that previously.
>>
>
>
> If there is no information, on what basis are you making all your
> claims?
I have not made any claims. I have speculated about the lack of a need
for a hardware-specific video driver with the Gamestorm platform. Just
for giggles and grins, where is the 64-bit driver for your ATI card?
[Sutherland's URL goes here]
>
>
>>>Can you provide *any* evidence that the speed of 64bit render software
>>>is significantly (as in more than twice) faster than that of 32bit?
>>
>>There is no information about the Gamestorm platform performance yet.
>>You have been told that previously.
>>
>
>
> If there is no information, on what basis are you making all your
> claims?
I have not made any claims. I have speculated about the lack of a need
for a hardware-specific video driver with the Gamestorm platform.
>
>
>>>There are certainly plenty of 64bit benchmarks which show that you can
>>>indeed get up to almost twice the throughput over 32bits for *some*
>>>types of processing. But this is worlds away from the 20-30 fold
>>>increase you would need in order to render complex scenes in realtime
>>>for a game!
>>
>>You don't have a clue as to what the performance requirements would be.
>>
>
>
> You can't make up your mind if there *are* any performance increases.
>
>
>>>
>>>
>>>>>how come banchmarks show it often rendering 3D
>>>>>scenes *slower* than the 32bit Athlon?
>>>>>
>>>>>http://www.tomshardware.com/cpu/20030422/opteron-23.html
>>>>>
>>>>>http://www.extremetech.com/article2/0,3973,1217390,00.asp
>>>>>
>>>>>More magic?
>>>
>>>
>>>LOL - you can't answer, can you?
>
>
>
> Nope, didn't think so.
--
David T. Johnson
Jack Troughton
>Might it? How? Magic pixie dust?
The way this conversation is going, I'd suspect angel dust.
--
-------------------------------------------------------------------
* Jack Troughton jake at consultron.ca *
* http://consultron.ca irc.ecomstation.ca *
* Kingston Ontario Canada news://news.consultron.ca *
-------------------------------------------------------------------
David Sutherland
<djoh...@isomedia.com> wrote:
32bit systems using dedicated graphics solutions.
> No claim
>has been made by myself or others that there will be an "incredible boost."
>
Given that existing CPU's are completely unable to provide anything
close to the 3D performance of a dedicated hardware solution the only
way they could become comparable is by receiving an incredible boost
in performance.
You even went on to claim that this boost will come without any need
for hardware drivers! ROTFL!
>>
>> That blows your whole argument out of the water.
>
>Incorrect.
>
Nobody else believes you, and all the evidence shows how misinformed
you are.
>>
>>
>>>> You have claimed that the 3D gaming performance of the
>>>>Opteron - with no additional hardware - is going to kill DirectX.
>>>
>>>No, I did not.
>>
>>
>[ad hominem snipped]
>>
>>
>>>I claimed that the Gamestorm platform inflicted a mortal
>>>wound on DirectX due to the ability of the Gamestorm platform to boot
>>>the OS and game off of a CD with no installation on the hard disk
>>>required. DirectX is a miserable kludgy piece of software whose only
>>>purpose for even existing is to allow high performance games to run
>>>while Windows is also booted. DirectX (and Direct3D) was supposed to
>>>give them 'direct' access to the hardware to improve performance. If
>>>people are running games from a CD with Linux rather than booting from
>>>the hard drive, they will not be using DirectX and DirectX is dead. Get it?
>>>
>>
>>
>> No. People are not going to ditch Windows just so they can have the
>> joy of losing their familiar computing environment every time they
>> want to play a game - especially when that game is already available
>> for Windows.
>
>People will ditch Windows for running games in a heartbeat if they can
>run games without the need to 'install' on their hard drive or deal with
>all of the DirectX and .DLL nonsense of the Windows platform.
They can do that now. It's called a console. And still people choose
Windows.
> I can't
>think of a worse platform for running high performance games than
>Windows. Windows would *never* be used for game except that it is the
>OS that is used by something like 99 percent of all computers.
What about the consoles? What about Linux?
>Face it.
> DirectX is dead if Gamestorm can do what they claim and the only
>person who will cry about its demise will be a few people like yourself.
>
Trouble is that they haven't claimed the things that you have.
>>
>>
>>
>>>>This is patently stupid, but it *is* your claim.
>>>
>>>Actually, I think it's pretty good.
>>>
>>
>>
>> Your capacity for self-congratulation has not gone un-noticed.
>>
>>
>>>>You *did* create this very thread entitled "Death blow for Windows
>>>>DirectX", after all.
>>>
>>>One of the few correct things you have managed to write among your mound
>>>of gibberish.
>>>
>>
>>
>> I understand why you are so angry about being proved wrong with facts,
>> David. It's okay :)
>
>You have not proved anything yet with any of your 'facts.'
>
Which evidence that I've presented do you think is false?
>>
>>
>>>>
>>>>>The benchmarks you you provide below, however,
>>>>>are for the Opteron running 32-bit software rather than 64-bit software.
>>>>>The Gamestorm platform is apparently 64-bit.
>>>>>
>>>>
>>>>
>>>>Given that existing CPU's are most certainly not capable of rendering
>>>>3D with anything like the same speed or quality of hardware boards,
>>>
>>>I don't accept that as a given.
>>
>>
>> You don't accept a lot of things - that doesn't make the facts of the
>> matter any less true.
>>
>>
>>>There is nothing special about hardware
>>>boards that cannot be done with the CPU.
>>
>>
>> LOL - sure you can just *emulate* all the custom single-pass rendering
>> hardware!
>
>yes, with a performance penalty, depending on the implementation.
>
A *huge* performance penalty.
>>
>>
>>>The clock speed of the NV34
>>>GPU in my video card is only 325 Mhz which is not earth-shattering and
>>>pretty much the only thing these video processors do is floating point
>>>calculations.
>>>
>>
>>
>[ad hominem snipped]
>>
>>
>>>>how else do you explain your claim that such hardware is now obsolete?
>>>
>>>I certainly haven't claimed that all video hardware is obsolete. I have
>>>suggested that in its current form, it is not likely to last. I am also
>>>suggesting that standards will make hardware-specific drivers
>>>unnecessary. Video hardware, in some form, will always be needed,
>>>although RAMDAC chips will probably be gone fairly soon. Also, I am
>>>speculating, rather than making claims. There is no information about
>>>the Gamestorm platform yet. You have been told that previously.
>>>
>>
>>
>> If there is no information, on what basis are you making all your
>> claims?
>
>I have not made any claims. I have speculated
Ad nauseum...without any facts to back you up. All your speculation
is based, by your own admission, on what SCI *did not say*.
>about the lack of a need
>for a hardware-specific video driver with the Gamestorm platform. Just
>for giggles and grins, where is the 64-bit driver for your ATI card?
http://mirror.ati.com/companyinfo/press/2003/4638.html
Specifically for the Opteron :) So much for it rendering drivers
obsolete! LOL!!!
>
>[Sutherland's URL goes here]
>
I fail to see the relevance to the discussion, but see above ;)
You *could* try researching this stuff before making an idiot of
yourself, you know.
>>
>>
>>>>Can you provide *any* evidence that the speed of 64bit render software
>>>>is significantly (as in more than twice) faster than that of 32bit?
>>>
>>>There is no information about the Gamestorm platform performance yet.
>>>You have been told that previously.
>>>
>>
>>
>> If there is no information, on what basis are you making all your
>> claims?
>
>I have not made any claims. I have speculated about the lack of a need
>for a hardware-specific video driver with the Gamestorm platform.
Based upon what the press release *didn't* say. And from that you
built your whole crumbling fantasy world, and spelt out to everybody
just how ignorant you are of the technology. Good work! :)
>>
>>
>>>>There are certainly plenty of 64bit benchmarks which show that you can
>>>>indeed get up to almost twice the throughput over 32bits for *some*
>>>>types of processing. But this is worlds away from the 20-30 fold
>>>>increase you would need in order to render complex scenes in realtime
>>>>for a game!
>>>
>>>You don't have a clue as to what the performance requirements would be.
>>>
>>
>>
>> You can't make up your mind if there *are* any performance increases.
>>
>>
>>>>
>>>>
>>>>>>how come banchmarks show it often rendering 3D
>>>>>>scenes *slower* than the 32bit Athlon?
>>>>>>
>>>>>>http://www.tomshardware.com/cpu/20030422/opteron-23.html
>>>>>>
>>>>>>http://www.extremetech.com/article2/0,3973,1217390,00.asp
>>>>>>
>>>>>>More magic?
>>>>
>>>>
>>>>LOL - you can't answer, can you?
>>
>>
>>
>> Nope, didn't think so.
Still can't refute the evidence? :)
David Sutherland
<jake!sp...@consultron.ca> wrote:
>On Wed, 8 Oct 2003 08:48:49 UTC, David Sutherland <sutherda@**ANTI-SPAM**btconnect.com> wrote:
>
>>Might it? How? Magic pixie dust?
>
>The way this conversation is going, I'd suspect angel dust.
He does seem to live in cloud cuckoo land. He's now repeating himself
and disowning his own comments, so I think he's almost done :)
Jason Bowen
David Sutherland <sutherda@**ANTI-SPAM**btconnect.com> wrote:
>
>How is that relevant? They are not referring to the game engine
>David, but the *CPU*. That comment was updated on the 21st August of
>this year - when the Opteron was an available product. And the game
>developers say that even the *fastest* CPU can't hope to match the
>power of a hardware solution.
>
It's obvious from David's postings( along with those of Eddie ReTards)
that some people really shouldn't attempt to talk about things of which
they really have no understanding. David's feeble attempts to make
arguments with regards to software development and his understanding of
how software/hardware fit together really shows him to be a fool. I don't
think he understands why it is that a GPU is faster than the same
operations being taken over by software being executed on a general
purpose CPU.
David Sutherland
<djoh...@isomedia.com> wrote:
Which part of the gamestorm announcement *specifically* are you
referring to? Where do they say that the game no longer requires
hardware support? Where do they say that it no longer uses OpenGL?
The official website states quite categorically that the game requires
both of these things.
>There have not
>yet been any subsequent details so we don't know the specifics.
Correction - there have been no specifics *at all* other than those
from the official website.
>You are
>pointing to hardware and software requirements for running the 32-bit
>OpenGL version of America's Army on Linux that obviously do not apply to
>the Gamestorm platform and are therefore irrelevant.
Bullshit - the folks who did the original Linux port did the 64bit
work as well. They have *not* rewritten a brand new engine from
scratch. Go look at their website.
>For example, the
>requirements you list say that the game will run acceptably on a 733 Mhz
>PIII while the Gamestorm platform *requires* a 64-bit Opteron.
DUH - do you think maybe it's got something to do with the OS and game
being 64bit? ROTFLMAO!
>
>>
>> If you don't want to believe the people who know that engine better
>> than anyone else then it's obvious that you *choose* not to believe
>> them. That isn't surprising, but it does make clear just how low you
>> are prepared to sink rather than admit that you might be wrong.
>
>We don't even know what the "Gamestorm" 'engine' is yet
Why do you assume that gamestorm refers to a game engine at all?
Where is it stated in the press release?
>and we do not
>know what the video hardware requirements will be for Gamestorm.
We know what the video hardware requirements are for Americas Army.
> If we
>did, they would have been listed in this thread by now.
They already have.
>Your pathetic
>attempt to list the 32-bit OpenGL hardware requirements and claim that
>they are also the hardware requirements for Gamestorm is an example of
>your complete lack of understanding of the technical aspects.
>
LOL - you really are a very funny little man. What "technical
aspects" have you managed to prove, David? You can't even figure out
the 21 is greater than 14! Loser! ;)
>>
>>
>>>>Here's some more for chuckles:
>>>>
>>>>Q: I have Intel or Intel3D integrated graphics , can I run the game?
>>>>A: No, you can't. Please read the minimum system requirements.
>>>> (Last Updated: 2003-08-21)
>>>>
>>>>Wondering what those might be? Knew you'd ask!
>>>>
>>>>http://www.americasarmy.com/faq_linux.php?p=3&t=4
>>>>Q: What are the minimum system requirements for the Linux v1.9 Client?
>>>>A:
>>>>Pentium3 766 (or equivalent)
>>>>128MB RAM
>>>>glibc 2.1 or newer.
>>>>Hardware accelerated OpenGL drivers (Nvidia's and XiG's are known)
>>>
>>>This is referring to the current 32-bit version and not the new 64-bit
>>>Gamestorm version. The Gamestorm 64-bit version specifically states
>>>that it requires the 64-bit Opteron CPU.
>>>
>>
>>
>> How is that relevant? How many times faster do you think the Opteron
>> is in 64-bit mode? Two? Two hundred? Where are the benchmarks!
>
>No Gamestorm benchmarks are available nor is specific information on the
>Gamestorm platform available yet. You have been told that several times
>previously.
>
Answer the question. If the Opteron can render one 3D frame in 45
seconds for a benchmark, then in order to render at the required 45
frames *per second* then you are looking at a 1000-fold increase in
calculation speed! Even if we were to reduce the scene complexity to
just 10% of the original it's a 100-fold increase. Point out the
benchmark result that supports such an absurd conclusion. The Opteron
is *NOT* 100x faster than it's predecessors!
>>
>>
>>>>
>>>>Q: I have a Kyro, KyroII, or Hercules 'prophet' branded video card,
>>>>can I run the game?
>>>>A: No, you can't. Please read the minimum system requirements. While
>>>>the kyro chipsets are recent and have the neccessary amount of video
>>>>ram, they do not have some of the transform and lighting features
>>>>required to show the high quality rendering that this game uses.
>>>>==
>>>>
>>>>Face facts, David - your claims are complete nonsense. Even the
>>>>people whose product you are trying to use as evidence directly
>>>>contradict everything you say.
>>>>
>>>>Fact - the game needs an OpenGL accelerated card.
>>>
>>>That is not a fact, nor even a reasonable supposition.
>>>
>>
>>
>> See the official website, David. Your personal speculations do not
>> outweigh it's assertions of fact. *THEY* know the game better than
>> you do.
>
>They have not provided any specific information yet on the 64-bit
>Gamestorm port of the game.
The only way they could do the same graphics without acceleration is
with an incredible increase in CPU power and memory bandwidth. No
benchmark supports such a claim.
>You have been told that several times
>previously. One thing we can be certain of is that the capabilities of
>the 64-bit port will not be the same as the 32-bit version.
Why? What capabilities do you think the 64bit version of Americas
Army will gain?
>At a
>minimum, I would speculate that it will have more detail in the frames
>than it does in the 32-bit version, making the game more realistic.
Why? Have they redone all the textures? Remodeled all the game
elements? Nope, they haven't done any of those things.
You are just pissing into the wind :)
>
>>
>>
>>
>>>>Fact - the Opteron (or any other CPU) cannot perform the same
>>>>functions as dedicated 3D hardware in an acceptable manner.
>>>
>>>Also not a fact. The Opteron CPU is much more powerful than a typical
>>>graphics card GPU.
>>>
>>
>>
>> This is nonsense. If you really believe this then you'll be willing
>> to take my money over it, right?
>>
>> US$100 says that the Opteron cannot render 3D games faster (or even
>> within 20%) the speed of the Radeon9800. Put up or shut up.
>
>Neither the Opteron CPU nor a graphics card GPU 'render 3D games.' They
>do floating point calculations and the Opteron is much more powerful
>that a typical graphics card GPU.
>
Now you are simply wallowing in your semantics trough in order to
avoid the consequences of your claims.
If you really believe that an Opteron without dedicated 3D hardware
can match 32bit platforms *with* a Radeon or Geforce then accept my
wager. Otherwise the only conclusion is that you do not think you
will win. And that would make you a liar :)
>>
>>
>>
>>>>Fact - if your card isn't fast enough, you can't play (so much for
>>>>hardware independence!)
>>>
>>>Gamestorm did not list any specific video hardware requirements for
>>>their new game. You have been told this previously.
>>>
>>>
>>>>Fact - you have been talking crap.
>>>>
>>>
>>>That is certainly true for you.
>>
>>
>>
>[snip of ad hominem]
Cluck cluck cluck.
David Sutherland
<djoh...@isomedia.com> wrote:
>David Sutherland wrote:
>> On Tue, 07 Oct 2003 15:48:17 -0700, "David T. Johnson"
>> <djoh...@isomedia.com> wrote:
>>
>> [snip]
>>
>>
>>>You are not capable of refuting the claims made by SCI about their
>>>Gamestorm platform because 1) you obviously don't know very much about
>>>computer graphics
>>
>>
>> ROTFL! This from the guy claiming that dedicated 3D hardware is no
>> faster than a general purpose CPU! <shrieks with mirth>
>
>Actually, I compared the GPU with the CPU. The CPU is obviously much
>more powerful. Don't know why you would find that humorous.
>
Claiming that dedicated silicon is slower than a general purpose CPU
is *very* funny. Do you know how Nvidia and ATI build their GPU's?
They emulate them on CPU's - with *much* *MUCH* slower performance.
Maybe you should go and tell them what they are doing wrong!
>>
>>>and 2) you don't know anything about the Gamestorm
>>>platform.
>>>
>>
>>
>> You certainly don't. You don't even know what the Opteron is capable
>> of.
>
>I am very familiar with the Opteron because I am using one.
In 32bit legacy mode.
> We don't
>know the specifics of the Gamestorm platform because they have not been
>released yet. You have been told that several time previously.
>
That hasn't stopped you making some very stupid claims about it.
>>
>>
>>>>
>>>>
>>>>>The last sentence is revealing because it suggests that 'porting' is
>>>>>required to run other games on the Gamestorm platform which suggests
>>>>>that Gamestorm does not run OpenGL games.
>>>>>
>>>>
>>>>
>>>>Bullshit. Americas Army is an OPenGL game as you have already been
>>>>told.
>>>
>>>The 32-bit version is an OpenGL game. We don't know anything about the
>>>new 64-bit 'Gamestorm' version yet.
>>>
>>
>>
>> It is also an OpenGL game as it's built on the same engine.
>
>We don't know any of the details about the 64-bit port of AA running on
>the Gamestorm platform for Opteron. You have been told that several
>times previously.
>
Since when does a "port" use a brand new engine? If it's been written
from the ground up it's no longer a port! Get with the program,
stupid!
>
>> It was
>> done by the same folks who did the original Linux port.
>
>I don't think we even know that, though you could actually be correct.
>
I so often am :)
>> No doubt
>> *they've* got it wrong too - LOL!
>>
>> http://icculus.org/
>> http://www.linux-gamers.net/modules/news/article.php?storyid=32
>
>I see no mention of the 64-bit Gamestorm port of America's Army.
>
From the website above:
"Our german partner site holarse.net reports that Super Computer Inc.
and icculus have developed a 64-bit linux distribution named
GameStorm, which has a 64-bit version of Americas Army integrated.
This happened in cooperation with AMD, nVidia and the US Army (for
which icculus works in order). You could read it out from icculus'
Finger, when he got back from San Franzisco, too.
Just got back from San Francisco, where we showed UT2004 running on
Win64 with DirectX9 and ArmyOps running on Linux/x86-64 from a
bootable CD.
Let me emphasize that Americas Army 64-bit is only running on AMD64
systems and only under Linux! It isn't much known about the publishing
of the 64-bit version of Americas Army, and we are still waiting for
an official announcement from Super Computer Inc. So stay tuned."
As this is the ONLY article on that page and it explicitly talks about
SCI and the 64bit AA, it's pretty damn obvious that you are either
blind or a liar. And I know you're not blind.
Do you think people are more likely to believe you when you lie so
outrageously?
I can't see any point responding further - you are now in complete
denial, refusing even to see text on a page if it disagrees with your
fantasies.
You've refused to endorse your claims by accepting my wager, so it's
clear that you don't even believe them yourself - you just have some
kind of pathological issue with admitting that you jumped to
conclusions. You have my sympathy.
David Sutherland
<djoh...@isomedia.com> wrote:
[snip]
If you really believe any of your bullshit, then why are you too
chicken to accept my bet? It would be easy money for you, right? If
you prefer, we'll make it payable to the charity of the winners
choice. But no - you will never accept because you know the reality,
even if you are too gutless to admit it.
Keep ignoring the evidence David, and enjoy the fact they you've
proven yourself clueless on yet another branch of technology :)
David Sutherland
Bowen) wrote:
He has humiliated himself to a much greater degree than usual with
this topic. He's now at the point where he is claiming that 14 is
greater than 21, and a page that talks about nothing but the game in
question doesn't mention it at all! You've got to wonder what's going
on in his head.
Jack Troughton
Angel dust. I'm not kidding.
It's the same thing that makes the 64 bit registers magically turn
into 32 bit registers.
You know, the CR0 register vs the "LMA mode" register.
Dunno... somehow I find the reference (all 522 pages of it) more
believable than him.
Curtis Bass
> A company named SCI has developed a bootable CD that contains a game and
> a Linux OS that can be booted on any Opteron PC to turn it into a
> powerful gaming console.
Uh, David, you do realize that the Opteron was not mentioned anywhere
in the article you cited, don't you? You've spent days arguing how
the _Opteron_ may somehow magically do away with dedicated graphics
hardware (a moronic conclusion, btw), but the article doesn't even
talk about the Opteron -- it's the Athlon that's being discussed.
See for yourself:
> http://www.linuxdevices.com/news/NS3103154102.html
Also, I've got a general purpose question for you. You keep saying
that the announcement referenced above makes no mention of graphics
hardware requirements, and you use this fact to support a claim/guess
that this somehow means that there are no hardware requirements, and
that all rendering is done by the CPU. What makes you think that your
conclusion/claim/guess is any more valid than the counter-claim,
namely, that the h/w requirements are simply the same as (or at least
similar to) those of the 32-bit version? You are suffering from the
logical fallacy of assuming that a lack of evidence is evidence of
lack, which is absurd. For starters, there _is_ the evidence that
Sutherland has been posting, namely, the 32-bit requirements. Also,
there is the URL that Dave Sutherland posted that explicitly stated,
"Modern 3D graphics cards are so powerful that even the fastest CPU
would not be able to take their place with a software mode. " In
light of this, why do you keep repeating your baseless _guess_ that
"The Gamestorm platform may have acceptable performance without the
need for hardware-specific drivers"?
You are being two-faced, as well. You keep repeating your baseless
guess, but when someone asks you to provide evidence to support your
guess, you insipidly barf up, "No information has been provided about
the 64-bit version of America's Army running on Gamestorm for Opteron.
You have been told that several times previously." Well, if there is
"no information", David, then quit making your stupid statements about
how this "kills DirectX" and how there is "no need for hardware
drivers" any more. You see, "no information" works both ways. Thing
is, Sutherland _does_ have some information, namely, the
aforementioned URLs.
Your conclusion that DirectX is dead is Just Plain Stupid. And so is
your conclusion that this somehow benefits OS/2.
Curtis
David T. Johnson
specifically for that hardware to obtain acceptable performance.
>
>>No claim
>>has been made by myself or others that there will be an "incredible boost."
>>
>
>
> Given that existing CPU's are completely unable to provide anything
> close to the 3D performance of a dedicated hardware solution the only
> way they could become comparable is by receiving an incredible boost
> in performance.
Incorrect. The Opteron CPU is a 64-bit cpu with a hypertransport
interface to much more memory at greater speed than is possible with
32-bit systems. We don't know what the Gamestorm 64-bit platform is
capable of yet, as you have been told previously.
>
> You even went on to claim that this boost will come without any need
> for hardware drivers! ROTFL!
No specific hardware has yet been identified by SCI as required for the
Gamestorm platform. I have pointed out that they may be using a
non-hardware-specific standard such as VBE 3.0 (supported by all current
video hardware) to interface with the video hardware, thereby
eliminating the need for hardware-specific drivers.
>
>
>
>>>That blows your whole argument out of the water.
>>
>>Incorrect.
>>
>
>
[snip of ad hominem]
>
>
>>>
>>>>>You have claimed that the 3D gaming performance of the
>>>>>Opteron - with no additional hardware - is going to kill DirectX.
>>>>
>>>>No, I did not.
>>>
>>>
>>[ad hominem snipped]
>>
>>>
>>>>I claimed that the Gamestorm platform inflicted a mortal
>>>>wound on DirectX due to the ability of the Gamestorm platform to boot
>>>>the OS and game off of a CD with no installation on the hard disk
>>>>required. DirectX is a miserable kludgy piece of software whose only
>>>>purpose for even existing is to allow high performance games to run
>>>>while Windows is also booted. DirectX (and Direct3D) was supposed to
>>>>give them 'direct' access to the hardware to improve performance. If
>>>>people are running games from a CD with Linux rather than booting from
>>>>the hard drive, they will not be using DirectX and DirectX is dead. Get it?
>>>>
>>>
>>>
>>>No. People are not going to ditch Windows just so they can have the
>>>joy of losing their familiar computing environment every time they
>>>want to play a game - especially when that game is already available
>>>for Windows.
>>
>>People will ditch Windows for running games in a heartbeat if they can
>>run games without the need to 'install' on their hard drive or deal with
>>all of the DirectX and .DLL nonsense of the Windows platform.
>
>
> They can do that now. It's called a console. And still people choose
> Windows.
A PC offers some advantages over a console gaming platform.
Specifically, it has more ways to interface with the system (joysticks,
keyboard, mouse), it has a much more powerful CPU, much more memory,
usually a better display screen, etc.
>
>
>>I can't
>>think of a worse platform for running high performance games than
>>Windows. Windows would *never* be used for game except that it is the
>>OS that is used by something like 99 percent of all computers.
>
>
> What about the consoles? What about Linux?
Both consoles and Linux are much better platforms for running games than
any version of Windows.
>
>
>>Face it.
>> DirectX is dead if Gamestorm can do what they claim and the only
>>person who will cry about its demise will be a few people like yourself.
>>
>
>
> Trouble is that they haven't claimed the things that you have.
Give them a couple of years of sales success and the claims will appear.
<laugh>
>
>
>>>
>>>
>>>>>This is patently stupid, but it *is* your claim.
>>>>
>>>>Actually, I think it's pretty good.
>>>>
>>>
>>>
>>>Your capacity for self-congratulation has not gone un-noticed.
>>>
>>>
>>>
>>>>>You *did* create this very thread entitled "Death blow for Windows
>>>>>DirectX", after all.
>>>>
>>>>One of the few correct things you have managed to write among your mound
>>>>of gibberish.
>>>>
>>>
>>>
>>>I understand why you are so angry about being proved wrong with facts,
>>>David. It's okay :)
>>
>>You have not proved anything yet with any of your 'facts.'
>>
>
>
> Which evidence that I've presented do you think is false?
Well, for starters, you claimed that the hardware requirements for
32-bit opengl games running on Linux were also the hardware requirements
for the Gamestorm platform. You also claimed that GPUs were much more
powerful than a "general purpose" CPU.
That is an announcement saying they will support the Opteron with
*future* drivers. There is no link to any driver, though, that you can
download and use *today*.
>
>
>>[Sutherland's URL goes here]
>>
>
>
> I fail to see the relevance to the discussion, but see above ;)
>
[ad hominem snipped]
>
>
>
>>>
>>>>>Can you provide *any* evidence that the speed of 64bit render software
>>>>>is significantly (as in more than twice) faster than that of 32bit?
>>>>
>>>>There is no information about the Gamestorm platform performance yet.
>>>>You have been told that previously.
>>>>
>>>
>>>
>>>If there is no information, on what basis are you making all your
>>>claims?
>>
>>I have not made any claims. I have speculated about the lack of a need
>>for a hardware-specific video driver with the Gamestorm platform.
>
>
> Based upon what the press release *didn't* say.
My speculation was reasonable, given the announced capabilities (boots
the OS and game without need for hard drive install, requires Opteron
system), and the known capabilities of Opteron (64-bit, hypertransport).
[ad hominem snipped]
>
>
>>>
>>>>>There are certainly plenty of 64bit benchmarks which show that you can
>>>>>indeed get up to almost twice the throughput over 32bits for *some*
>>>>>types of processing. But this is worlds away from the 20-30 fold
>>>>>increase you would need in order to render complex scenes in realtime
>>>>>for a game!
>>>>
>>>>You don't have a clue as to what the performance requirements would be.
>>>>
>>>
>>>
>>>You can't make up your mind if there *are* any performance increases.
>>>
>>>
>>>
>>>>>
>>>>>>>how come banchmarks show it often rendering 3D
>>>>>>>scenes *slower* than the 32bit Athlon?
>>>>>>>
>>>>>>>http://www.tomshardware.com/cpu/20030422/opteron-23.html
>>>>>>>
>>>>>>>http://www.extremetech.com/article2/0,3973,1217390,00.asp
>>>>>>>
>>>>>>>More magic?
>>>>>
>>>>>
>>>>>LOL - you can't answer, can you?
>>>
>>>
>>>
>>>Nope, didn't think so.
>
>
>
> Still can't refute the evidence? :)
--
David T. Johnson
The only specific hardware requirement listed is that an Opteron system
be used.
> Where do they say that it no longer uses OpenGL?
They did not say that OpenGL was required. Also, there are no 64-bit
hardware-specific OpenGL drivers yet for ATI or Nvidia video hardware
that I am aware of. Those would be required to run a 64-bit OpenGL game
unless software rendering was used.
>
> The official website states quite categorically that the game requires
> both of these things.
Only for the 32-bit version. The requirements of the Gamestorm port are
not yet known. You have been told that previously.
>
>
>>There have not
>>yet been any subsequent details so we don't know the specifics.
>
>
> Correction - there have been no specifics *at all* other than those
> from the official website.
The original SCI announcement did provide a few specifics, such as that
Gamestorm would only run on an Opteron system and that it was using
64-bit Linux.
>
>
>>You are
>>pointing to hardware and software requirements for running the 32-bit
>>OpenGL version of America's Army on Linux that obviously do not apply to
>>the Gamestorm platform and are therefore irrelevant.
>
>
>
> Bullshit - the folks who did the original Linux port did the 64bit
> work as well. They have *not* rewritten a brand new engine from
> scratch. Go look at their website.
The details of the Gamestorm 64-bit port of America's Army are not
known. You have been told that previously. There are no details of the
Gamestorm platform or the 64-bit AA port on any website yet AFAIK.
>
>
>>For example, the
>>requirements you list say that the game will run acceptably on a 733 Mhz
>>PIII while the Gamestorm platform *requires* a 64-bit Opteron.
>
>
> DUH - do you think maybe it's got something to do with the OS and game
> being 64bit? ROTFLMAO!
<laugh> You concede, then, that the Gamestorm port is 64-bit and will
therefore not run on the 733 Mhz PIII that is listed as the minimum
hardware requirement. If that hardware requirement is wrong, there is
no reason to think that the other hardware requirements are not.
>
>
>>>If you don't want to believe the people who know that engine better
>>>than anyone else then it's obvious that you *choose* not to believe
>>>them. That isn't surprising, but it does make clear just how low you
>>>are prepared to sink rather than admit that you might be wrong.
>>
>>We don't even know what the "Gamestorm" 'engine' is yet
>
>
> Why do you assume that gamestorm refers to a game engine at all?
> Where is it stated in the press release?
The details of the Gamestorm platform are not yet known. You have been
told that previously.
>
>
>>and we do not
>>know what the video hardware requirements will be for Gamestorm.
>
>
> We know what the video hardware requirements are for Americas Army.
Only for the 32-bit Open GL version. We do not yet know what the
hardware requirements are for the Gamestorm 64-bit version are. If we
did, they would have been listed in this thread by now.
>
>
>>If we
>>did, they would have been listed in this thread by now.
>
>
> They already have.
Incorrect. You listed the 32-bit OpenGL version requirements and
attempted to claim that those were also the requirements for the 64-bit
Gamestorm version. The details of the Gamestorm platform are not yet
known. You have been told that previously.
>
>
>>Your pathetic
>>attempt to list the 32-bit OpenGL hardware requirements and claim that
>>they are also the hardware requirements for Gamestorm is an example of
>>your complete lack of understanding of the technical aspects.
>>
>
>
[ad hominem snipped]
>
>
>>>
>>>>>Here's some more for chuckles:
>>>>>
>>>>>Q: I have Intel or Intel3D integrated graphics , can I run the game?
>>>>>A: No, you can't. Please read the minimum system requirements.
>>>>> (Last Updated: 2003-08-21)
>>>>>
>>>>>Wondering what those might be? Knew you'd ask!
>>>>>
>>>>>http://www.americasarmy.com/faq_linux.php?p=3&t=4
>>>>>Q: What are the minimum system requirements for the Linux v1.9 Client?
>>>>>A:
>>>>>Pentium3 766 (or equivalent)
>>>>>128MB RAM
>>>>>glibc 2.1 or newer.
>>>>>Hardware accelerated OpenGL drivers (Nvidia's and XiG's are known)
>>>>
>>>>This is referring to the current 32-bit version and not the new 64-bit
>>>>Gamestorm version. The Gamestorm 64-bit version specifically states
>>>>that it requires the 64-bit Opteron CPU.
>>>>
>>>
>>>
>>>How is that relevant? How many times faster do you think the Opteron
>>>is in 64-bit mode? Two? Two hundred? Where are the benchmarks!
>>
>>No Gamestorm benchmarks are available nor is specific information on the
>>Gamestorm platform available yet. You have been told that several times
>>previously.
>>
>
>
> Answer the question. If the Opteron can render one 3D frame in 45
> seconds for a benchmark,
Incorrect assumption.
> then in order to render at the required 45
> frames *per second* then you are looking at a 1000-fold increase in
> calculation speed! Even if we were to reduce the scene complexity to
> just 10% of the original it's a 100-fold increase. Point out the
> benchmark result that supports such an absurd conclusion. The Opteron
> is *NOT* 100x faster than it's predecessors!
The details of the Gamestorm platform are not yet known. You have been
told that previously.
>
>
>>>
>>>>>
>>>>>Q: I have a Kyro, KyroII, or Hercules 'prophet' branded video card,
>>>>>can I run the game?
>>>>>A: No, you can't. Please read the minimum system requirements. While
>>>>>the kyro chipsets are recent and have the neccessary amount of video
>>>>>ram, they do not have some of the transform and lighting features
>>>>>required to show the high quality rendering that this game uses.
>>>>>==
>>>>>
>>>>>Face facts, David - your claims are complete nonsense. Even the
>>>>>people whose product you are trying to use as evidence directly
>>>>>contradict everything you say.
>>>>>
>>>>>Fact - the game needs an OpenGL accelerated card.
>>>>
>>>>That is not a fact, nor even a reasonable supposition.
>>>>
>>>
>>>
>>>See the official website, David. Your personal speculations do not
>>>outweigh it's assertions of fact. *THEY* know the game better than
>>>you do.
>>
>>They have not provided any specific information yet on the 64-bit
>>Gamestorm port of the game.
>
>
> The only way they could do the same graphics without acceleration is
> with an incredible increase in CPU power and memory bandwidth.
As I mentioned earlier, there are systems with integrated graphics in
which the system memory is also used as video memory. These systems
have a 32-bit CPU and much less memory bandwidth than an Opteron system
and yet are able to run some current 3D games with acceptable
performance. The details of the Gamestorm platform are not yet known.
You have been told that previously.
> No
> benchmark supports such a claim.
>
>
>>You have been told that several times
>>previously. One thing we can be certain of is that the capabilities of
>>the 64-bit port will not be the same as the 32-bit version.
>
>
>
> Why? What capabilities do you think the 64bit version of Americas
> Army will gain?
The details of the Gamestorm platform are not yet known. You have been
told that previously.
>
>
>>At a
>>minimum, I would speculate that it will have more detail in the frames
>>than it does in the 32-bit version, making the game more realistic.
>
>
> Why? Have they redone all the textures? Remodeled all the game
> elements? Nope, they haven't done any of those things.
Those increased capabilities would be made possible by the much larger
amount of memory that an Opteron can address at a relatively high
bandwidth and by the presence of other hardware capabilities when
running in Long Mode such as new floating point instructions and more
and larger registers.
>
[ad hominem snipped]
>
>
>>>
>>>
>>>>>Fact - the Opteron (or any other CPU) cannot perform the same
>>>>>functions as dedicated 3D hardware in an acceptable manner.
>>>>
>>>>Also not a fact. The Opteron CPU is much more powerful than a typical
>>>>graphics card GPU.
>>>>
>>>
>>>
>>>This is nonsense. If you really believe this then you'll be willing
>>>to take my money over it, right?
>>>
>>>US$100 says that the Opteron cannot render 3D games faster (or even
>>>within 20%) the speed of the Radeon9800. Put up or shut up.
>>
>>Neither the Opteron CPU nor a graphics card GPU 'render 3D games.' They
>>do floating point calculations and the Opteron is much more powerful
>>that a typical graphics card GPU.
>>
>
>
> Now you are simply wallowing in your semantics trough in order to
> avoid the consequences of your claims.
>
> If you really believe that an Opteron without dedicated 3D hardware
> can match 32bit platforms *with* a Radeon or Geforce then accept my
> wager.
I have not said that video hardware was not required. I have said that
hardware-specific drivers may not be required the video hardware, thanks
to standards such as VBE 3.0 and the improved performance of the
Opteron/Gamestorm platform.
[ad hominem snipped]
>
>
>
>>>
>>>
>>>>>Fact - if your card isn't fast enough, you can't play (so much for
>>>>>hardware independence!)
>>>>
>>>>Gamestorm did not list any specific video hardware requirements for
>>>>their new game. You have been told this previously.
>>>>
>>>>
>>>>
>>>>>Fact - you have been talking crap.
>>>>>
>>>>
>>>>That is certainly true for you.
>>>
>>>
>>>
>>[snip of ad hominem]
>
>
> Cluck cluck cluck.
>