OGL and D3D are BOTH losing the API battle
Michael I. Gold
mgr...@iastate.edu wrote:
>
> Just thinking the other day...
>
> All this talk about D3D vs. OpenGL over the past year or so seems more
> pointless than ever in retrospect. Both are currently losing, and so
> is the industry IMO.
>
> Look what we have now as the dominant API : Glide.
>
> It's bad for most customers because it seems Nvidia alone is
> outselling voodoo at this point,
It is ironic that so many developers were afraid to use OpenGL, fearing
insufficient driver coverage, yet many of these same developers felt
comfortable targetting the Glide-only market. :-)
Chris Blackwell
mgr...@iastate.edu wrote:
> So why is Glide so popular? Well, it seems to be easy to program and
> efficient. The real answer, though, appears to be 3dfx's developer
> support. These guys are throwing a ton of resources towards helping
Developer support is part of it, particularly since the SDK is available
free of charge, no "registering" (read "get your e-mail for spam")
required. I think, however, that one of the reasons that it is so popular
is that it is
a) Fast
b) Simple
c) Give EASY access to ALL of the features on the the 3dFX, something
neither D3D or OpenGL do.
For me, item c is the biggest reason I have been using Glide. I love the
3dFX. I've seen just about all 3d consumer PC accelerators, and so far
none of them are quite as nifty (to me at least) as the 3dFX.The number of
really cool effects and that can be done by using various blending and
texturing modes can really only be achieved via Glide, since they are
effects that are very specific to the 3dFX and wouldn't be too useful in a
general purpose 3d library.
Just a side note, I think, due to the quality and growing user support for
3dFX, that it has the potential to become the "Sound Blaster" of the 3d
world. As much as OpenGL and D3D are good concepts, there is a BIG place
in my heart for standardized hardware (as long as it does the job well)
and would have no problem if we went from this mish mash of 3d cards that
currently exists to a bunch of 3dFX register compatible cards.
Cheers,
Chris Blackwell
mgr...@iastate.edu
Just thinking the other day...
All this talk about D3D vs. OpenGL over the past year or so seems more
pointless than ever in retrospect. Both are currently losing, and so
is the industry IMO.
Look what we have now as the dominant API : Glide.
It's bad for most customers because it seems Nvidia alone is
outselling voodoo at this point, and NVidia + Rendition + NEC combined
almost certainly are.* Even if this is not the case, a large number
of consumers are being left out of the 3d accelerated game market.
Its bad for developers because they don't sell as many games as they
would with d3d or opengl support, and because glide-only games
alienate customers. Sure the consumer can buy the software-only
version but if you're like me and have had a taste of 3d acceleration,
you can't go back.
So why is Glide so popular? Well, it seems to be easy to program and
efficient. The real answer, though, appears to be 3dfx's developer
support. These guys are throwing a ton of resources towards helping
devs make Glide games. 3dfx will hold your hand every step of the way
and even put its own devs on your project to help.
MS and SGI would do well to learn from this. SGI may not currently
have the $$$ to throw around, but MS certainly does. Working this
closely with the developers would give both a chance to find out what
developers really need and want in an API, what is good in D3d and
OpenGL, what is bad, what works, and what doesn't. And we'll all end
up with a Farenheit being more in tune with what devs really need. At
the very least, we'd have more games being sold.
-Matt
-------------
* Zealots, don't even think of starting a chipset flame war. Thats
not what this message is about.
Brandon Van Every
Michael I. Gold <go...@berkelium.com.xyzzy> wrote in article
<34DC0457...@berkelium.com.xyzzy>...
> mgr...@iastate.edu wrote:
> >
> > Look what we have now as the dominant API : Glide.
> >
> > It's bad for most customers because it seems Nvidia alone is
> > outselling voodoo at this point,
>
> insufficient driver coverage, yet many of these same developers felt
> comfortable targetting the Glide-only market. :-)
Well, their fears about OpenGL have always been and continue to be grounded
in the real world of IHV support. Whereas for a time, GLIDE was indeed the
best API option. Everyone was playing the "write for my proprietary API"
game and GLIDE did indeed win. Now because of NVIDIA there's a hiccup in
some people's product cycles, I guess you could call that "ironic" but I'd
say it was to be expected.
Cheers,
--
Brandon J. Van Every <vane...@blarg.net> DEC Commodity Graphics
<http://www.blarg.net/~vanevery> NT Intel Alpha OpenGL D3D
-----------------------------------------------------------------------
Come to Seattle Vtalk, to discuss all aspects of online virtual worlds!
Bi-weekly meetings, <http://www.emeraldnet.net/~tobler/VTALK/Vtalk.htm>
Brandon Van Every
mgr...@iastate.edu wrote in article <34dc0cd7.5002332@news>...
> Just thinking the other day...
>
>
> efficient. The real answer, though, appears to be 3dfx's developer
> support. These guys are throwing a ton of resources towards helping
> devs make Glide games. 3dfx will hold your hand every step of the way
> and even put its own devs on your project to help.
It also seems that Glide fits the game developer's preferred model at this
stage of commodity 3d HW evolution. Lets developers get their hands dirty
with all the geometry pipeline stuff they want to optimize. Shit, *I*
would like to be in that position and I sure don't feel that way working on
lousy OpenGL or D3D drivers. :-)
Anyways, point is that once the commodity HW subsumes the geometry engine,
Glide won't be relevant anymore. Of course at the current trajectory,
it'll be gone long before then anyways.
Sean Timarco Baggaley
In article <34DC0457...@berkelium.com.xyzzy>, Michael I. Gold
<go...@berkelium.com.xyzzy> writes
>> pointless than ever in retrospect. Both are currently losing, and so
>> is the industry IMO.
>>
>>
>> It's bad for most customers because it seems Nvidia alone is
>> outselling voodoo at this point,
>
>It is ironic that so many developers were afraid to use OpenGL, fearing
>insufficient driver coverage, yet many of these same developers felt
>comfortable targetting the Glide-only market. :-)
Am I the only one to spot the similarities between 3DFX cards and the
early years of Gravis' UltraSound? Need I also mention that UltraSound
*still* isn't the number one selling card, for exactly the same reason
as the 3DFX? (Great hardware [initially]; minimal OEM support).
Or am I missing something?
--
Sean Timarco Baggaley
E&OE
<panic>
DO NOT USE REPLY-TO! I DO NOT RECEIVE E-MAILS TO THIS ACCOUNT.
</panic>
Please send mail to: stim...@hotmail.com
Bjorn
This is from a game developers perspective:
OpenGL is so similar to Glide that you can easily write a game that supports
both (my own engine has the render-backend abstracted so it took 2 hours to
make a back-end for OpenGL, given the one for Glide, they are really
similar).
OpenGL is really easy and much more powerful in the _development sense_ than
Glide. The biggest advantage is scalability (the same argument kills
Direct3D btw). I can develop my game on a top-of-the-line SGI workstation or
a killer PC with OpenGL support, then scale it easily down to an end-user
level PC when the game is ready - the opengl miniport already exist for the
Voodoo boards and Diamond announced a real OpenGL driver for the Monster-3D
II.
The bottom line is that all these API's are so similar that I couldn't care
less which one is winning or loosing, because the game itself will be
identical anyway, the only thing that changes is the backend mapping and
that doesn't take many days to rewrite. Even if Direct3D sucks, it really
isn't that much job to write a back-end for it, for example.
From a consumers perspective:
I want a fast board that can run all programs. I shouldn't be locked into
picking Voodoo-based boards just because the Glide version of games is much
faster than the Direct-3D version. Therefore you might argue that 3dfx is
hurting the industry by providing their proprietary API (which obviously
cannot be made a standard, since it's so hardware feature dependant).
3dfx should provide a highly optimized opengl miniport instead and
developers should learn that it is as fast as Glide, then the problem is
solved. Just look at idSoftware, they don't use glide and quake1/2 are great
games.
/Bjorn
mgr...@iastate.edu wrote in message <34dc0cd7.5002332@news>...
>All this talk about D3D vs. OpenGL over the past year or so seems more
>pointless than ever in retrospect. Both are currently losing, and so
>is the industry IMO.
>
>Look what we have now as the dominant API : Glide.
>
>It's bad for most customers because it seems Nvidia alone is
Martin
Really confusing, this 3D world, isn't it ?
First we have several different 3D cards, then several companies try
to make a standard and now we have multiple standards.
As for me, If I would make a game that supported 3D acceleration, I
think I'd choose 3dfx. We need 1 (one!) standard on this, just like
the soundblaster is right now.
---
Martin van der Plas
mvd...@wxs.nl
-One man's windows are another man's walls-
Thomas Womack
Martin wrote in message <34dc990...@news.wxs.nl>...
>Really confusing, this 3D world, isn't it ?
>First we have several different 3D cards, then several companies try
>to make a standard and now we have multiple standards.
>
>As for me, If I would make a game that supported 3D acceleration, I
>think I'd choose 3dfx. We need 1 (one!) standard on this, just like
>the soundblaster is right now.
I hate to point this out, but the Soundblaster hasn't been a standard for
some time. Most people buying PCs now end up with Ensoniq or Yamaha chipsets
on their motherboards, on which software is needed to emulate the archaic
Soundblaster.
The 3dfx card, had it become standard slightly more quickly, would have made
as much of a mess of the 3d card market as the Soundblaster made of the
sound card one. You *DON'T* want hardware standards, you want API
standards - and you want versioned API standards, so, when the
Reality-Monster-on-an-AGP3-card comes out, programs using Direct3d version 3
will still run. Microsoft got this one right.
Tom
Martin
>The 3dfx card, had it become standard slightly more quickly, would have made
>as much of a mess of the 3d card market as the Soundblaster made of the
>sound card one. You *DON'T* want hardware standards, you want API
>standards - and you want versioned API standards, so, when the
>Reality-Monster-on-an-AGP3-card comes out, programs using Direct3d version 3
>will still run. Microsoft got this one right.
>
>Tom
>
This would be great if every developer used only one API.
No, instead they all use a different one. And as a result the end-user
needs to know about all of them, install new versions every time its
updated and so on. I don't think this is very good.
Oh well..... you don't HAVE to use an API of you don't want to...
Michael I. Gold
Swordfish wrote:
>
> Speaking of Opengl is there any gaming card out there that has a FULL
> opengl implementation running on it. Many cards advertise opengl these
> days, but do they have full implementations?
The recently released beta ICD for the RIVA 128 is a full implementation
which passes the OpenGL conformance test and runs a number of
applications very well, but there are some bugs. The Permedia driver is
also a full implementation.
john devoy
Well, seeing as microsoft are changing the direct3d API to the opengl
standard I guess opengl will be the winner.
Thomas Womack
Martin wrote in message <34dcc8a0...@news.wxs.nl>...
>>The 3dfx card, had it become standard slightly more quickly, would have
made
>>as much of a mess of the 3d card market as the Soundblaster made of the
>>sound card one. You *DON'T* want hardware standards, you want API
>>standards - and you want versioned API standards, so, when the
>>Reality-Monster-on-an-AGP3-card comes out, programs using Direct3d version
3
>>will still run. Microsoft got this one right.
>>
>>Tom
>>
>
>This would be great if every developer used only one API.
>No, instead they all use a different one.
AFAIK there are three - 3dfx (early versions of Turok, Barrage which I'm not
sure ever appeared, possibly some versions of SOTE), OpenGL (Quake,
Hexen ][, Quake ][), and Direct3d (all the rest).
>And as a result the end-user
>needs to know about all of them, install new versions every time its
>updated and so on. I don't think this is very good.
I don't have a 3dfx, so all I need to do is install DirectX 5 when some game
tells me to (and DirectX 5 is usually supplied on the game's CD in that
case), and perhaps install OpenGL drivers when I first install Quake. It's
not too much of a problem, and a good installation routine for a commercial
game should probably do it transparently (apart from perhaps adding an extra
reboot).
Tom
Swordfish
Michael I. Gold <go...@berkelium.com.xyzzy> wrote in article > It is ironic
that so many developers were afraid to use OpenGL, fearing
> insufficient driver coverage, yet many of these same developers felt
> comfortable targetting the Glide-only market. :-)
>
I think they would rather have thier game look fantastic on one card
instead of fantastic on some cards, mediocre on others ect. At least thats
how I interpret there choices of sticking with glide.
Speaking of Opengl is there any gaming card out there that has a FULL
opengl implementation running on it. Many cards advertise opengl these
days, but do they have full implementations?
--
*************************************************************************
Remove NOSPAM from the return adress to respond
Geoffrey Clark
Drywall Productions
http://www.concentric.net/~Rufertto
UIN#5527558
*************************************************************************
Thatcher Ulrich
In article <PdAxCDA4...@dial.pipex.com>,
stbag...@solarflair.demon.co.uk says...
> In article <34DC0457...@berkelium.com.xyzzy>, Michael I. Gold
> <go...@berkelium.com.xyzzy> writes
> >mgr...@iastate.edu wrote:
> >>
> >> Just thinking the other day...
> >>
> >> pointless than ever in retrospect. Both are currently losing, and so
> >> is the industry IMO.
> >>
> >> Look what we have now as the dominant API : Glide.
> >>
> >> It's bad for most customers because it seems Nvidia alone is
> >> outselling voodoo at this point,
> >
> >insufficient driver coverage, yet many of these same developers felt
> >comfortable targetting the Glide-only market. :-)
>
> early years of Gravis' UltraSound? Need I also mention that UltraSound
> *still* isn't the number one selling card, for exactly the same reason
> as the 3DFX? (Great hardware [initially]; minimal OEM support).
>
> Or am I missing something?
You're missing the fact that the Gravis cards were buggy buggy BUGGY and
hard to program. And quality control was spotty. One of the worst
hardware decisions we made on the VRBike was to go with Gravis. (To be
fair, they were cheap and the sound quality was quite good.)
--
Thatcher Ulrich
http://world.std.com/~ulrich
Rob Rodgers
"Michael I. Gold" <go...@berkelium.com.xyzzy> wrote:
>Swordfish wrote:
>> Speaking of Opengl is there any gaming card out there that has a FULL
>> opengl implementation running on it. Many cards advertise opengl these
>> days, but do they have full implementations?
>
>which passes the OpenGL conformance test and runs a number of
>applications very well, but there are some bugs. The Permedia driver is
>also a full implementation.
If you're loose as that, you could define ANY card as having a full
implementation running on it. Just use the software libs.
I think, however, the poster meant, "is there any gaming card which
implements all of OpenGL?" (at least as far as the rasterization
stage) and the answer is no, and that includes the R128.
Russ Williams
Martin wrote in message <34dc990...@news.wxs.nl>...
>Really confusing, this 3D world, isn't it ?
>First we have several different 3D cards, then several companies
>try to make a standard and now we have multiple standards.
>
>As for me, If I would make a game that supported 3D
>acceleration, I think I'd choose 3dfx. We need 1 (one!) standard
>on this, just like the soundblaster is right now.
Noooooo! The Soundblaster is the worst thing to ever happen
to sound cards. It was a trivial upgrade to the AdLib standard
and due to the needs for backward compatability, sound cards
stagnated. Very little changed between the AdLib and the SB16
- just more channels and more bits. The wavetable on the AWE
cards was a slight improvement, but compared to the advances
in video cards over the same time, it's pathetic.
Now that we don't have to code for DOS and the shite de facto
standards, we can actually get some *good* sound cards. Ones
that do hardware mixing, for example. How about some nice
bus-mastering PCI sound cards?
---
Russ
Andy Buchanan
john devoy wrote in message <6bjbkn$5of$1...@news.enterprise.net>...
>Well, seeing as microsoft are changing the direct3d API to the opengl
>standard I guess opengl will be the winner.
>
Andy Buchanan
Martin wrote in message <34dc990...@news.wxs.nl>...
>Really confusing, this 3D world, isn't it ?
>First we have several different 3D cards, then several companies try
>to make a standard and now we have multiple standards.
>
>As for me, If I would make a game that supported 3D acceleration, I
>think I'd choose 3dfx. We need 1 (one!) standard on this, just like
>the soundblaster is right now.
Actually, the Soundblaster Standard is obsolete except in terms of
running legacy DOS programs. The current standard for soundcards,
is merely to have a windows driver that works. I predict this to be the
case for 3D cards too.
--
AndyB, Programmer
Electronic Arts UK Studio
mailto:abuc...@ea.com
Thomas Womack
Swordfish wrote in message <01bd3443$4f634180$3d5553ce@mortimes>...
>
>
>Michael I. Gold <go...@berkelium.com.xyzzy> wrote in article > It is ironic
>that so many developers were afraid to use OpenGL, fearing
>> insufficient driver coverage, yet many of these same developers felt
>> comfortable targetting the Glide-only market. :-)
>>
>
> I think they would rather have thier game look fantastic on one card
>instead of fantastic on some cards, mediocre on others ect. At least thats
>how I interpret there choices of sticking with glide.
But this is completely insane from a commercial standpoint. It seems
perilously close to some of the more spurious 'artistic integrity'
arguments - 'if he's not going to see my game with exactly the coloured
lighting and bilinear filtering I've specified, he shouldn't see it at all'.
Is anyone still producing Glide-only games?
Tom
Matt Craighead
Michael I. Gold wrote:
>
> Swordfish wrote:
> >
> > Speaking of Opengl is there any gaming card out there that has a FULL
> > opengl implementation running on it. Many cards advertise opengl these
> > days, but do they have full implementations?
>
> The recently released beta ICD for the RIVA 128 is a full implementation
> which passes the OpenGL conformance test and runs a number of
> applications very well, but there are some bugs. The Permedia driver is
> also a full implementation.
One of my friends with a Permedia 2 regularily complains about how
horrible the OpenGL drivers are...
--
Matt Craighead
Utumno developer: http://www.citilink.com/~craighea/utumno/
Burnnn
>Is anyone still producing Glide-only games?
Oh lord yes. Longbow 2 for one. There are tons more, but for some
reason, I can't think of any at the moment. :o)
----------
Burnnn
Michael Haar
umfe...@cc.umanitoba.ca (Burnnn) writes:
No, there is a D3D-patch for Longbow 2 (for RIVA128 cards).
Same with most other new Glide-"only"-games ...
Michael
Martin
>Noooooo! The Soundblaster is the worst thing to ever happen
>to sound cards. It was a trivial upgrade to the AdLib standard
>and due to the needs for backward compatability, sound cards
>stagnated. Very little changed between the AdLib and the SB16
> - just more channels and more bits. The wavetable on the AWE
>cards was a slight improvement, but compared to the advances
>in video cards over the same time, it's pathetic.
I didn't say the soundblaster was good - it sucks - I was referring to
the fact that it was a standard.
And about that AWE32.. - have you ever seen how to program this thing
? 909978028937 lines of code for initializing ONLY. Now that's a bad
card. Oh well, you get good sound quality in return....
>Now that we don't have to code for DOS and the shite de facto
>standards, we can actually get some *good* sound cards. Ones
>that do hardware mixing, for example.
Like the Gravis Ultrasound ? or AWE32 ?
> How about some nice
>bus-mastering PCI sound cards?
You can't use DMA with PCI cards, so I don't think that is a good
idea. (soundblaster compatibiliby etc, lots of DOS games that don't
work anymore) --- hey, an UPGRADE card would be nice, just like the
monster 3d is for 3dfx....
Anyways, check out the Diamond monster sound card -> I think this one
has what you want.
Russ Williams
Martin wrote in message <34e09a87...@news.wxs.nl>...
>>Noooooo! The Soundblaster is the worst thing to ever happen
>>to sound cards. It was a trivial upgrade to the AdLib standard
>>and due to the needs for backward compatability, sound cards
>>stagnated. Very little changed between the AdLib and the SB16
>> - just more channels and more bits. The wavetable on the AWE
>>cards was a slight improvement, but compared to the advances
>>in video cards over the same time, it's pathetic.
>
>I didn't say the soundblaster was good - it sucks - I was referring
>to the fact that it was a standard.
A bad standard is worse than none at all (which, BTW, is
impossible - there will always be a partial standard).
>And about that AWE32.. - have you ever seen how to program
>this thing ?
Nope. I don't intend to find out, either.
>909978028937 lines of code for initializing ONLY.
Wouldn't suprise me...
>Now that's a bad card. Oh well, you get good sound quality
>in return....
>
>
>>Now that we don't have to code for DOS and the shite de facto
>>standards, we can actually get some *good* sound cards. Ones
>>that do hardware mixing, for example.
>
>Like the Gravis Ultrasound ? or AWE32 ?
The original SB doesn't. That practically killed the GUS.
>>How about some nice bus-mastering PCI sound cards?
>
>You can't use DMA with PCI cards, so I don't think that is a
>good idea.
Who needs DMA? Bus-mastering cards are memory mapped
and have their own blitters. No silly 64k limits or segment
aligning or buffers in DOS memory... Hell, even a simple
memcpy is faster than DMA!
The PCI bus is more than fast enough to play samples direct
from RAM and that sort of card would probably have 4mb of
local RAM for sound fonts and cacheing.
>(soundblaster compatibiliby etc,
Irrelevant.
>lots of DOS games that don't work anymore)
Again, irrelevant. But cards like the Diamond Monster Sound
emulate SBs under Windows and you can always keep your
old SB...
> --- hey, an UPGRADE card would be nice, just like the
>monster 3d is for 3dfx....
Have a look about the Monster Sound...
>Anyways, check out the Diamond monster sound card -> I think
>this one has what you want.
Yup. It's what I was thinking of when I wrote that. It's a good start,
but it's only 1 card. I'm sure there'll be more and better still to
come...
---
Russ
David Springer
Michael I. Gold (go...@berkelium.com.xyzzy) wrote:
> Swordfish wrote:
> >
> > Speaking of Opengl is there any gaming card out there that has a FULL
> > opengl implementation running on it. Many cards advertise opengl these
> > days, but do they have full implementations?
>
> The recently released beta ICD for the RIVA 128 is a full implementation
> which passes the OpenGL conformance test and runs a number of
> applications very well, but there are some bugs. The Permedia driver is
> also a full implementation.
It's also getting 45+ fps in Quake 2. Well beyond Voodoo and flirting with
Voodoo 2. It's less than half the expected price of Voodoo 2, doesn't
require the purchase of a separate 2D acclerator, and it works just as
well in a window as it does full screen. And the way it scales up with
processor speed... can't wait to see if it keeps scaling linearly all
the way through 450mhz Pentium II.
Fantastic OGL driver work Michael. Awesome 2D/3D chip.
David Springer
Senior Engineer
Dimension Product Group
Dell Computer Corporation
David Springer
Andy Buchanan (o...@clara.net) wrote:
>
> Actually, the Soundblaster Standard is obsolete except in terms of
> running legacy DOS programs. The current standard for soundcards,
> is merely to have a windows driver that works. I predict this to be the
> case for 3D cards too.
>
Yup. By next year you won't see any major PC vendors shipping
a product with an ISA card in it. The year after that the ISA
slots will be gone too. Good riddance. Antiques belong in
museums.
David Springer
Dimension Engineering
Dell Computer Corporation
Burnnn
On 10 Feb 1998 11:37:41 GMT, "Michael Haar"
<Michae...@Informatik.Uni-Oldenburg.DE> wrote:
I know that, BUT D3D "patch" to a Glide-only game. The D3D *beta*
patch is unsupported, barely tested, only works with RIVA, and is
still quite buggy.
There are Glide only games out there. If you want to find them, go to
http://www.3dfx.com I'm too lazy to go find more examples. :o)
----------
Burnnn
Paul Miller
> It's also getting 45+ fps in Quake 2. Well beyond Voodoo and flirting with
> Voodoo 2. It's less than half the expected price of Voodoo 2, doesn't
The only thing wrong with it is the inadequate 4MB of RAM. That's the
only thing keeping me from buying one.
> Fantastic OGL driver work Michael. Awesome 2D/3D chip.
Got your check, huh David? ;^)
--
Paul Miller | st...@fxtech.com
Thatcher Ulrich
In article <34E23556...@fxtech.com>, st...@fxtech.com says...
> > It's also getting 45+ fps in Quake 2. Well beyond Voodoo and flirting with
> > Voodoo 2. It's less than half the expected price of Voodoo 2, doesn't
>
> The only thing wrong with it is the inadequate 4MB of RAM. That's the
> only thing keeping me from buying one.
I just bought a new Dell that came with a RIVA-based card. Overall
I'm happy. The performance is excellent, and it's cool that it works in
a window. I also took a peek at the card itself, which has about three
chips on it; obviously less expensive than a Voodoo card, which bodes
well for market penetration of decent hardware acceleration.
But... is it just me, or does it not have sub-pixel precision? Also,
what's the story on its texture filtering -- it appears as though it
selects MIP maps on a per primitive (not per-pixel) basis, which looks
pretty cheesy. These may seem like minor things, but they do affect the
quality of the imagery, and they're things that Voodoo does better.
David Springer
First off, thank you for buying a Dell. I'm on the motherboard
engineering team for Dimension. That's my baby you got there. :-)
Second, yes, the weak point of the Riva 128 is image quality when
compared to Voodoo. Perhaps Michael Gold can elaborate further. I
understand it has to do with texture compression and MIP mapping.
Third, the 4MB limitation is only a problem if you need to get to
1280x1024 or greater in true color, which is why we also offer
an 8MB card (Permedia 2) which does the higher 2D modes but is
lower in 3D performance. The reason 4MB isn't a problem for 3D
is AGP. You'll find that the Riva 128 can play games (try Jedi
Knight at 800x600) that a Monster 3D can't handle for lack of
memory. The Riva 128 will use system memory for texture
storage through the 256MB/SEC (approx) AGP bus. If you use
a D3D benchmark program (I forget which one lists amount of
texture memory) it will tell you the card has 32MB of texture
memory (assuming you have 64MB of system memory).
Fourth, NVIDIA is still making amazing improvements in both
D3D and OGL drivers. Look for a new one coming out soon on
the STB web page. NVIDIA feeds reference drivers to its
OEM board manufacturers and they then incorporate the improvements
into their branded drivers.
David Springer
Dell Computer Corporation
--
*************** IGAMES INTERNET GAME LOBBY ****************
* *
* NOW SUPPORTING MICROSOFT DirectPlay 3 LOBBY STANDARD ! *
* *
* A real-time game lobby for the internet with many *
* exciting games and thousands of players. Game *
* developers, players, and ISP's can try it out at: *
* *
****************** http://www.igames.com ******************
Paul Miller
> But... is it just me, or does it not have sub-pixel precision? Also,
> what's the story on its texture filtering -- it appears as though it
> selects MIP maps on a per primitive (not per-pixel) basis, which looks
> pretty cheesy. These may seem like minor things, but they do affect the
Yeah, this is the other major (for people who notice these details)
deficiency in its design - no per-pixel mip-mapping and sometimes you
can see texture compression artifacts. The performance is nice but
quality-wise it doesn't touch the Voodoo (again, for people who SEE the
details - a lot of people dont even notice).
--
Paul Miller | st...@fxtech.com