AAA for Amiga from ESCOM?? I say _yes_
Brian King
1. True High-Density floppy support, including 4 Meg floppies
2. Various screen-modes which support ANIMATION smoothly, similar to HAM8,
but with many optimizations and enhancements.
Number 2 is quite important, as everyone keeps moving towards multi-media.
Heck, if the next Amiga had a nice backplane (like PCI), had an 060, had RTG
(which would probably be necessary to use AAA anyways), and had either MPEG
in hardware or in software, then animation-specific video modes would be
AWESOME.
Heck, you can take a P90, stick a great PCI SVGA card in it, and you still
get sucked out showing anything bigger than 256 colour MPEGs at 640x480. An
Amiga with 060 and AAA would perform better.
Ah, well, we're just dreaming (like Dave Haynie was), and it'll probably
never materialize.
--
+-------------------------------------+-----------------------------------+
| Brian D. King - Computer Engineer | Fourth Dimension Software, author |
| PBX/ISDN Software Development Group | of LhA-GUI, Cert. Amiga Developer |
| Mitel Corporation, Kanata, Ontario | *** ki...@software.mitel.com *** |
AGENT GORDON COLE
:Don't forget some of the other key benefits of AAA:
:1. True High-Density floppy support, including 4 Meg floppies
Does it support those new 120meg floppies that Compaq is putting in all their
new machines? Apparently the drive they are using can also read and write
regular hi-density(1.44meg) disks too. I think something like that would be
the best to use in any new Amigas since it will probably become the "standard"
--
------------------------------------------------------------------------------
-- Brian (AGENT GORDON COLE) Atkins||Like Twin Peaks,Blade Runner,The Elric --
-- Saga,Dune,or cool links? http://wc62.residence.gatech.edu/coop/home.html --
------------------------------------------------------------------------------
Mr Gavin Moran
OK, does anyone know if ESCOM has bought the rights to AAA as part of
the C= assets deal, and if so do they have any plans to finish the development
of them?
I'm fully aware there was much debate over the merits of AAA, with most of
the opposition to it being that 'it doesn't have the capabilties of top end
PC gfx cards, so why bother?'. Well, there are lots of reasons why it would
be viable.
(1) The AAA chipset offers some astounding features for multimedia and games,
and even though it wouldn't have the raw power of say, a Diamond Stealth PCI
card with 4 megs, it would be more than enough for many people.
(2) The Amiga benefits from having a tight, integrated, TV friendly chipset.
You can do things even on an A500 with OCS that a P100 with PCI card can't
match. AAA in all Amigas from a low end A1200 style model to the high end
060 based machines would be a major plus point.
(3) AAA offers excellent multi-channel 16-bit sound too, again great for
multimedia and games.
(4) AGA really needs updating. Its good enough for A1200s, but with new super
consoles on the horizon the Amiga needs some more powerful chips to be
attractive at the low end. We need 030/28 and AAA at least for a cheap low
end, with 060 , AAA and the option of RTG on the high end machines.
(5) Lets face it, if you are a power user and you want PC gfx then just go
and do what you'd have to do if you were a PC owner - buy a suitable gfx
board with a 64 bit svga chipset. Its not like you _can't_ get such boards
for the Amiga but they do cost a bit more. If there was RTG support in the
OS then its irrelevent for high end users whether AAA would be better than
the competition or not.
(6) In many ways AAA is a better chipset than any of the high end PC cards
will ever be. Why? Because its integrated into the system and TV friendly.
The trend in PC and Workstations is for device independence and simply to
throw CPU power and expensive hardware at problems to solve it. AAA would
slaughter ANY currently available PC chipset for DTV and multimedia.
(7) AAA is already at an advanced stage of development. A little extra
money would yield a powerful chipset for relatively low cost. In the
meantime owners can get by if there is RTG in the Amiga OS for the
high end users.
I want to see the Amiga evolve and AAA with faster CPUs is the way to
go for now. Put in RTG for those who moan about the gfx cards for PCs
being 'better' so that users can add such cards to the Amiga if they want.
gavan
--
email: G.M...@ee.qub.ac.uk | 'There can be only one!'
or gmo...@nyx.cs.du.edu | - The Highlander
kevin patrick kretsch
> OK, does anyone know if ESCOM has bought the rights to AAA as part of
>the C= assets deal, and if so do they have any plans to finish the development
>of them?
That's anyones guess, but it looks like the RISC chip will be the next
new one. I'm not holding my breath for the AAA. :(
To my knowledge, the AAA chip would be so good for the Amiga in
terms of graphics and sound because of the existing Amiga architecture,
i.e. graphics and sound chips. (Gary and Paula?) Another strong point
for the AAA, which Gavin didn't mention above, is that we would keep the
existing Amiga OS, which afterall is why we love the Amiga so much.
PLUS, we won't need 8Mb to run the system, like most PCs do. IF the RISC
based technology is chosen, then we're looking at 8Mb minimum
requirement, an OS that while similar to the old one, will not be
completely compatible with old stuff. But, apparrently, translating the
old software to the new system will be easy. (though we would still have
translate it!)
We just have to wait and see. Besides, how many of us writing all
this wonderful rhetoric about the new chip sets actually know anything
about them, other than what we read in magazines?? I don't...
...so I guess we should disregard everything I have just written!
Kev. :)
--
Kevin P. Kretsch |*| Department of Physics,
Final Year Undergrad. |*| Trinity College, Dublin,
E-Mail: kkre...@alf2.tcd.ie |*| Ireland.
**Any opinions above are not those of T.C.D., they could be my own though...**
Jason Compton
: In <3osu0q$14...@info4.rus.uni-stuttgart.de> gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
: That's anyones guess, but it looks like the RISC chip will be the next
: new one. I'm not holding my breath for the AAA. :(
: To my knowledge, the AAA chip would be so good for the Amiga in
: terms of graphics and sound because of the existing Amiga architecture,
: i.e. graphics and sound chips. (Gary and Paula?) Another strong point
: for the AAA, which Gavin didn't mention above, is that we would keep the
: existing Amiga OS, which afterall is why we love the Amiga so much.
: PLUS, we won't need 8Mb to run the system, like most PCs do. IF the RISC
: based technology is chosen, then we're looking at 8Mb minimum
: requirement, an OS that while similar to the old one, will not be
: completely compatible with old stuff. But, apparrently, translating the
: old software to the new system will be easy. (though we would still have
: translate it!)
You guys have to understand...
3D-RISC is not really intended to drive a system itself, it's intended to
be a coprocessor that can drive a low cost game machine if
necessary/required. That's why more than one Commodore engineer has said
that "3D-RISC makes a better CD32, but not necessarily a better 4000" or
words to that effect.
--
Jason Compton jcom...@xnet.com
Editor-in-Chief, Amiga Report Magazine (708) 741-0689 FAX
Whatever you do, better start doing it right.
AR on Aminet - docs/mags/ar???.lha AR Mailing list - Mail me
AR on WWW - http://www.cs.cmu.edu/~mjw/Amiga/News/AR/index.html
Dave Haynie
> OK, does anyone know if ESCOM has bought the rights to AAA as part of
>the C= assets deal,
Certainly, AAA was part of the central core of Amiga technologies.
>and if so do they have any plans to finish the development of them?
I would kind of doubt it, at this point.
>(1) The AAA chipset offers some astounding features for multimedia
>and games,
It's questionable just how much you'd be paying for that, though. In
most things, a modern 64-bit PCI chipset will be faster and
cheaper. Next year's models will probably blow AAA away
completely. While you need four chips and VRAM for a 32-bit AAA system
(six chips for a 32/64-bit system), most SVGA solutions are single
chips. The next generation is going to more advanced DRAM
technologies, like RDRAM, SDRAM, etc. that offers VRAM performance at
DRAM prices. The next generation is incorporating support for 3D and
MPEG2 on-chip, things not even considered for AAA. Some are even
supporting video features, AAA's last remaining advantage.
The reasons for these things are simple -- AAA is old. It was started
back in 1988 or so. Keep in mind what PCs did back then. Though never
finished, AAA is probably the first 64-bit graphics chip set started,
if never finished. Certainly for any personal computer. These days, a
video chip generation is about one year. So although AAA started way
ahead, it didn't magically stay there. Even Commodore had more
advanced graphics projects, like Hombre, at the end.
>(2) The Amiga benefits from having a tight, integrated, TV friendly
>chipset.
A good portion of the limitations you find on PCs, with respect to TV,
isn't due to the graphics chips themselves. It's the implementation of
the graphics card, the PC's OS, etc. While they're not quite as good
as the Amiga for video just yet, at least in terms of supporting
direct NTSC, PAL, overscan, etc. they're getting there, and probably
will overcome these problems in the next generation. Some chips have
already; all those video-compatible add-ons for the Amiga, with the
exception of the Video Toaster, used off-the-shelf parts.
>AAA in all Amigas from a low end A1200 style model to the high end
>060 based machines would be a major plus point.
AAA, at least as it exists today (or could exist with finished
versions of today's chips) is too expensive for an A1200, and in fact
wouldn't physically fit in one. You're taking about four large chips
for the basic system, plus lots of support logic that has yet to be
designed (a four-channel clock synthesizer, general Gary-style glue,
smart buffering, etc).
>(3) AAA offers excellent multi-channel 16-bit sound too, again great for
>multimedia and games.
Eight channels of 16-bit audio would certainly be useful. However,
you'd be much better off with two channels of direct audio and a
fairly decent General MIDI compatible synthesizer system. Doing good
music synthesis on the main CPU itself is fairly CPU intensive.
>(4) AGA really needs updating. Its good enough for A1200s, but with new super
>consoles on the horizon the Amiga needs some more powerful chips to be
>attractive at the low end.
Most of the new games consoles have more powerful graphics engines
than AAA. Sure, they may not do 1280x1024 (then again, the 32-bit
AAA system is limited to 800x600 anyway), but they have killer 3D
stuff for games, 24-bit (or thereabouts) graphics, etc. And they're
coming out at around $400, CD-ROM drive included.
>Its not like you _can't_ get such boards for the Amiga but they do
>cost a bit more.
That's one problem we had planned to address in the future -- ports
from the PClone industry cost more for two reasons. First of all, the
chips used to plug directly into the ISA bus, you needed extra logic
to convert to Zorro signaling. That's still true, though maybe today
it's PCI built-in, being converted to Zorro III. Either way, PCI is a
good bus, finally, and should be supported in future Amiga
systems. Also, AmigaOS 4.0 graphics.library intended to support chunky
framebuffers right out of the box. Since PCI can tell you whether a
card is VGA compatible or not, based on its device type bits, this
would have made any PCI graphics card plug and play, out of the box,
on the next generation Amiga. You would want an RTG driver to take
advantage of blitters and other special features, and I'm sure these
would become available. But you wouldn't pay any more for cards than
PClone folks would.
The problem, other than cost, is market. Amiga hardware vendors might
have a million potential buyers, more or less, for any hardware
item. So they have to charge more. PCI would also give these guys a
larger market, since the same hardware would play on the PCs, as well
as Alpha workstations, next generation Macs, IBMs, etc.
>If there was RTG support in the OS
That should be ESCOM's top priority on the "new stuff" checklist.
>then its irrelevent for high end users whether AAA would be better than
>the competition or not.
No it isn't. The resources of any company, ESCOM or Commodore
included, are limited. If they spend them on building a new graphics
chip, they won't have them to spend on something else. PC graphics
chips really are cheap, they're supported by 3rd party RTG today. Why
not use them in low end Amigas, even. I looked a bit at PC graphics
chips about a year ago. The TSENG W32p, for example, was a 32-bit
chip, supporting 24-bit graphics, high resolutions, Amiga-like
32-bit blitter, etc. I'm typing on one now (on a PC). This part was
$25 in quantity, and at the time they were working on a more
integrated version that would go for $15. That's cheap enough for the
inside of an A1200-type system. AAA chips were to start at more like
$20-$25 each, so your basic AAA system is maybe $80 minimum, cost,
before any glue logic is added.
>AAA would slaughter ANY currently available PC chipset for DTV and
>multimedia.
Not really. Multimedia looks like crap on most PCs mainly because
you're running through a little tube called Windows. Windows cripples
the hardware on the PC, plain and simple. Many different video chip
companies told me the same thing last year. On the Amiga, no such
limitations would exist, and that's a fact -- the AmigaOS is the best
OS on the market for support of realtime multimedia
processing. Nothing else even comes close. And realistically, all you
need to do is get close; PCs and Macs have several times the CPU
horsepower these days.
>(7) AAA is already at an advanced stage of development. A little extra
>money would yield a powerful chipset for relatively low cost.
The estimate, back at Commodore, was $500,000 to finish AAA. Of
course, at the time, there were still plenty of the AAA designers
working at Commodore. I would guess you might even have to double that
to get new people up to speed on the existing designs. That also adds
lots of time.
Dave Haynie | ex-Commodore Engineering | See my first film
Sr. Systems Engineer | Class of '94 | "The Deathbed Vigil"
Scala Inc., US R&D | C= Failure n. See: Greed | in...@iam.com
"Caught a bolt of lightning, cursed the day he let it go" -Pearl Jam
Dave Haynie
>kevin patrick kretsch (kkre...@tcd.ie) wrote:
>: In <3osu0q$14...@info4.rus.uni-stuttgart.de> gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
>: That's anyones guess, but it looks like the RISC chip will be the next
>: new one. I'm not holding my breath for the AAA. :(
>You guys have to understand...
>3D-RISC is not really intended to drive a system itself, it's intended to
>be a coprocessor that can drive a low cost game machine if
>necessary/required.
At the time, at least, the game system was a primary concern. Ideally,
Hombre would have been competitive (or better) than the emerging crop
of 32/64-bit games systems, like Sony Playstation, Sega Saturn,
Nintendo Ultra64, etc. It was supposed to be very low cost to make,
and use standard interfaces to allow things to be added that weren't
built-in. It would also plug into the PCI bus in a graphics card
configuration.
>That's why more than one Commodore engineer has said that "3D-RISC
>makes a better CD32, but not necessarily a better 4000" or words to
>that effect.
Maybe not A4000, but it would have had a place in the next generation
Amigas, which were to use PCI. Ideally, we would have had both Amiga
compatible and Hombre cards, user's choice. Or one could go for an
off-the-shelf SVGA card instead. But it's not like AmigaOS was running
on the PA-RISC, at least at this level. Though I think lots of folks
would have been happy with a fully fledged general purpose processor
available to crunch graphics...
Jyrki Saarinen
> PLUS, we won't need 8Mb to run the system, like most PCs do. IF the RISC
> based technology is chosen, then we're looking at 8Mb minimum
Why this minimum limit of 8MB, if AmigaDOS could be ported to HP/PA?
--
Warped Amiga programmer - Stellar
Des Whewell
: In <3p0md9$a...@flood.xnet.com>, jcom...@flood.xnet.com (Jason Compton) writes:
: >kevin patrick kretsch (kkre...@tcd.ie) wrote:
: >: In <3osu0q$14...@info4.rus.uni-stuttgart.de> gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
: >: That's anyones guess, but it looks like the RISC chip will be the next
: >: new one. I'm not holding my breath for the AAA. :(
: >You guys have to understand...
: >3D-RISC is not really intended to drive a system itself, it's intended to
: >be a coprocessor that can drive a low cost game machine if
: >necessary/required.
: At the time, at least, the game system was a primary concern. Ideally,
: Hombre would have been competitive (or better) than the emerging crop
: of 32/64-bit games systems, like Sony Playstation, Sega Saturn,
: Nintendo Ultra64, etc. It was supposed to be very low cost to make,
: and use standard interfaces to allow things to be added that weren't
: built-in. It would also plug into the PCI bus in a graphics card
: configuration.
Your last couple of posts have made interesting reading but were mainly
framed in the past tense. What do *you* think Peter Kittel should be doing
now? If the Amiga graphics and video handling can no longer be regarded as a
niche, where should the next generation machines be aimed?
--
.----------------------------------------------------.
! Des Whewell: d...@oregon.demon.co.uk - Coventry, UK !
`----------------------------------------------------'
Gary Alan Peake
>>money would yield a powerful chipset for relatively low cost.
>
>The estimate, back at Commodore, was $500,000 to finish AAA. Of
>course, at the time, there were still plenty of the AAA designers
>working at Commodore. I would guess you might even have to double that
>to get new people up to speed on the existing designs. That also adds
>lots of time.
Dave, speaking of 'new' people, can you give us a rundown on where the old
Amiga guys are and what they are doing now?
Gary Peake
1:106/7512.2
40:202/12.1
gpe...@cup.portal.com
Dave Haynie
>Dave Haynie <dave....@scala.com> wrote:
>Hm. Perhaps I'm being naive, but in my admittedly limited experience,
>a little thing known as HAM-8 has made a big difference for doing
>multimedia on AGA machines.
Of course it did. And HAM10, PACKLUT, and PACKHY modes on AAA would
also be quite useful for this. But primarily in the interest of saving
memory on hard disk, not increasing animation speed. For example, take
a AA system running HAM8, 640x400 interlaced, so there's no cycle
stealing at all. You can push a theoretical maximum of about 7MB/s
into that Chip RAM. Real systems do a bit less, depending on the
circumstances.
Now consider a PCI-based SVGA card. To do the same thing in 16-bit
pixels (also fairly good for video), you would need 14MB/s. For 24-bit
pixels, 21MB/s. When I was looking into available PCI graphics chips,
I found quite a few that could sustain 50MB/s into memory, no problem.
Some that approached 100MB/s. So hardware-wise, there's nothing in
frame animation by brute force that's beaten by HAM8. Well, almost
nothing; some of the first PCI chipsets for the other side of things,
CPU to PCI bus, sucked, limiting transfers to around 25MB/s (PCI bus
is limited to 132MB/s at 33MHz and 32-bits wide).
>All of **THAT** is expensive, too, and for video, HAM-8 is probably
>nearly indistinguishable from 24-bit except possibly for professionals.
>(Hey, moving NTSC looks fine -- do a freeze-frame and often one is
>astonished at how *lousy* the picture is.)
>
>The *point*, of course, is that none of these PeeCee chipsets have
>anything even *remotely* resembling HAM, despite its obvious utility
>for multimedia applications.
Well, we're not talking about what exists today, anyway, since AAA
doesn't. Virtually every prominant graphics chip company for PCI has
announced their next generation chip sets will have video
decompression in hardware. Usually MPEG-2. That will, in no uncertain
terms, bet the hell out of HAM. Or PACK. Yes, it's more difficult to
work with, you need special software to create it. Then again, you
generally do with HAM too.
>There's a lot of chatter about MPEG, but it seems to be in its way
>another "beat it to death" tool, because it requires so much
>computational power to decode it in real time.
And yet, if you get a hardware decoder, built into the graphics chip
you were going to have anyway, there's little if any incremental cost
to supporting MPEG. Kind of like HAM -- it's cheap to have it
built-in, but expensive to add-on.
>To be honest, I don't know if a HAM-encoded videoclip would be longer
>or shorter than an MPEG clip
HAM would be significantly longer, and at lower quality, than MPEG. A
good example is VideoCD, which could run from a single speed CD-ROM.
That's only MPEG-1, display quality varies from slightly better to
slightly worse than standard VHS, depending on the MPEG encoding job
and your opinion of MPEG artifacts vs. plain old VHS videotape. Now
try the same thing with HAM8 video. You need to deliver 30fps, and you
only get 150kB/s to play with from your media device. Have fun...
>but I suspect you'd be able to play back the HAM clip convincingly on
>a much less expensive machine than the MPEG.
Another basis for comparison is a simple existance proof -- if HAM
were in any way comparable to MPEG for real video playback, would
Commodore have bothered to develop the MPEG module for CD32?
Jyrki Saarinen
> Now consider a PCI-based SVGA card. To do the same thing in 16-bit
> pixels (also fairly good for video), you would need 14MB/s. For 24-bit
> pixels, 21MB/s. When I was looking into available PCI graphics chips,
> I found quite a few that could sustain 50MB/s into memory, no problem.
> Some that approached 100MB/s. So hardware-wise, there's nothing in
> frame animation by brute force that's beaten by HAM8. Well, almost
> nothing; some of the first PCI chipsets for the other side of things,
> CPU to PCI bus, sucked, limiting transfers to around 25MB/s (PCI bus
> is limited to 132MB/s at 33MHz and 32-bits wide).
Actually most VLB/PCI SVGA boards are in a 10-20MB/s range
on a 320x200x256 mode, when using 640x480x24bit perhaps
they slow down very much..
Dave Haynie
>> Now consider a PCI-based SVGA card. To do the same thing in 16-bit
>> pixels (also fairly good for video), you would need 14MB/s. For 24-bit
>> pixels, 21MB/s. When I was looking into available PCI graphics chips,
>> I found quite a few that could sustain 50MB/s into memory, no
>> problem.
>Actually most VLB/PCI SVGA boards are in a 10-20MB/s range
>on a 320x200x256 mode, when using 640x480x24bit perhaps
>they slow down very much..
Such as? Most PCI-based SVGA chips are far more advanced than
that. No VRAM-based chip slows down at all based on the video display
mode; they fetch pixels of the serial bus (like AAA could). The upper
half of the SVGA market is VRAM based.
640x480x24 bit requires about 26MB/s for full motion video, 30fps.
Modern SVGA chips have 10 times this bandwidth, or more. For example,
the DRAM-based 864 chip from S3 (in my PClone), which is fairly
low-end these days, does 264MB/s on its 64-bit local bus, though
clearly not all that through PCI (which does have a hard limit of
132MB/s, and again, not every PC's PCI bus interface chip can even
come close to sustaining that kind of rate, no matter what the video
chips can do).
The VRAM version of this chip has about 764MB/s of bandwidth available
on its local bus; PCI is once again the bottleneck, but you're not
going to find enough native work to even come close to saturating a
bus like that, even with PCI and blitters going full steam.
Next generation SVGA chips (really the point of my original article)
are going to cheaper RAM that supports VRAM level performance. So next
year your $200 video card (most of which have controller chips that go
for under $50 in quantity, sometimes way under) will have a native bus
bandwidth of around 400MB/s, and likely either 3D and/or motion video
(decompression, probably MPEG-2) support in hardware.
Jyrki Saarinen
> >Actually most VLB/PCI SVGA boards are in a 10-20MB/s range
> >on a 320x200x256 mode, when using 640x480x24bit perhaps
> >they slow down very much..
>
> Such as? Most PCI-based SVGA chips are far more advanced than
> that. No VRAM-based chip slows down at all based on the video display
> mode; they fetch pixels of the serial bus (like AAA could). The upper
> half of the SVGA market is VRAM based.
Are we talking about the same thing? I am talking about how
much one can move data to the video boards memory? And
this is usually about 10-20MB/s on VLB/PCI machines and
320x200x256 mode. Some cards like ET4000/W32p are faster,
like 30MB/s.
Etienne Vogt
Sorry to jump into this thread a bit late, but I don't have enough time
to keep up with the traffic in this group.
Dave Haynie (dave....@scala.com) wrote some time ago :
>[...]
>>(1) The AAA chipset offers some astounding features for multimedia
>>and games,
>
>It's questionable just how much you'd be paying for that, though. In
>most things, a modern 64-bit PCI chipset will be faster and
>cheaper. Next year's models will probably blow AAA away
>completely. While you need four chips and VRAM for a 32-bit AAA system
>(six chips for a 32/64-bit system), most SVGA solutions are single
>chips. The next generation is going to more advanced DRAM
>technologies, like RDRAM, SDRAM, etc. that offers VRAM performance at
>DRAM prices. The next generation is incorporating support for 3D and
>MPEG2 on-chip, things not even considered for AAA. Some are even
>supporting video features, AAA's last remaining advantage.
>
>The reasons for these things are simple -- AAA is old. It was started
>back in 1988 or so. Keep in mind what PCs did back then. Though never
>finished, AAA is probably the first 64-bit graphics chip set started,
>if never finished. Certainly for any personal computer. These days, a
>video chip generation is about one year. So although AAA started way
>ahead, it didn't magically stay there. Even Commodore had more
>advanced graphics projects, like Hombre, at the end.
It's true that today high end PC cards are faster than AAA. But AAA has
some other advantages.
- Compressed modes such as HAM or PACK-LUT
- Compatibility with ECS, so old software may run on it.
- And the most important one, unique to Amiga chipsets : _THE_COPPER_ !
Of course, you could emulate part of it in software, but this would be very
costly in CPU time. An alternative is to put it in a separate chip or to
integrate it in an existing SVGA chip.
Also AAA is not only a graphics subsystem. There's also a sound and I/O chip
in it. The sound part has 8 16 bit channels, the I/O part has a buffered
serial port and a pretty good floppy controller. This one is totally
programmable and can read every possible floppy format, even the infamous
low density MAC disks. How many PC controllers are able to read Amiga disks ?
>[...]
>>AAA in all Amigas from a low end A1200 style model to the high end
>>060 based machines would be a major plus point.
>
>AAA, at least as it exists today (or could exist with finished
>versions of today's chips) is too expensive for an A1200, and in fact
>wouldn't physically fit in one. You're taking about four large chips
>for the basic system, plus lots of support logic that has yet to be
>designed (a four-channel clock synthesizer, general Gary-style glue,
>smart buffering, etc).
AA is already 3 chips. One more should be possible in a low end machine.
And AAA chips use less power than current AA chips so they should even be
usable in a laptop.
I really think AAA + 060 should be the way to go for '96 machines. The
chipset was said to be 3 months to completion when ex C= decided to suspend
its development. If Amiga Technologies resume it now with new engineers,
then should be able to have it done for early 96. AFAIK only the Andrea
chip (DMA controller) needs major modifications.
I dont know how long it would take to integrate an SVGA chip into the
Amiga hardware, but if you need to design custom chips for Copper emulation
(Is an Amiga without Copper still an Amiga ?) or for the floppy controller
(Who wants to use the PC disk format ?) I guess it could take quite some
time and money.
Software wise, both AAA and SVGA require RTG to be fully functional in the
Amiga. But AAA is ECS compatible, so low-end users will be more likely to
upgrade to an AAA machine if they can run some of their old games.
A complete RTG OS was said to require over 12 months development time.
Basically intuition and graphics libraries have to be entirely rewritten.
But it should be possible to first write a partial RTG driver for AAA running
under the current OS that would allow up to 256 color screens.
The first AAA machines sold should then use a softkick boot rom (like the
first A3000s) so they could be easily upgraded to the full RTG OS when it
becomes available.
Etienne Vogt (Etienn...@obspm.fr)
Graduate Student at Meudon Observatory, France
A500/000 3.1 at home
A3000 & A4000/040 at work
MS-DOS : A Microsoft based, interrupt driven, non-reentrant program loader,
sometimes humourously referred to as an 'Operating System'.
(Ralph Babel, The Amiga Guru Book)
Dave Haynie
>> Such as? Most PCI-based SVGA chips are far more advanced than
>> that. No VRAM-based chip slows down at all based on the video display
>> mode; they fetch pixels of the serial bus (like AAA could). The upper
>> half of the SVGA market is VRAM based.
>Are we talking about the same thing?
Yup.
>I am talking about how much one can move data to the video boards
>memory?
That's the one. I spent a good bit of time researching this in my last
days at Commodore, because I had an application that needed quite a
bit of data/sec from PCI to VGA memory.
>And this is usually about 10-20MB/s on VLB/PCI machines and
>320x200x256 mode.
Again, the display mode doesn't matter with a modern chipset, or any
VRAM-based chipset. With older DRAM-based chips, sure, because just
like the Amiga, they had to steal memory cycles from the bus interface
to feed the display. This can't happen with VRAM, and it's unlikely
with a modern DRAM system. For example, the S3 864 has a 64-bit bus
with a bandwith of 264MB/s. While that's not all available to PCI
(which could handle a maximum of 132MB/s, peak, anyway), you rarely
can use that much for display and blitting. And you can turn off
blitting if necessary, to get more CPU to framebuffer bandwidth.
>Some cards like ET4000/W32p are faster, like 30MB/s.
The TSENG representative claimed they could get close to 100MB/s
between the PCI bus and DRAM. He wasn't certain it could actually hit
100MB/s (my target maximum rate), but it could get close, as long as
you used the interleaved memory. The W32p is a 32-bit chip, but uses
interleaved banks to get about 70% the bandwidth of DRAM-based 64-bit
SVGA chip. It also has an efficient set of FIFOs, to manage the four
channels into DRAM (normal graphics fetch, CPU access, blitter access,
and a secondary input for video-in-a-window. You clearly can't
saturate every channel at once and expect the thing to keep flying,
but it's a good design for a $25 chip.
If you have been having problems with bandwith to these things,
consider the rest of the system. A good portion of the PCI "host"
interfaces out there suck. The original Intel PCI chipset could manage
maybe 25MB/s, peak, on a good day. CPUs aren't the best thing to drive
long linear PCI transfers, anyway, since they're very poor at
collecting data into long burst transfers, with is the only way to get
high performance on PCI (or any other I/O bus, Zorro III included). Of
course, I wasn't using PCs or CPUs to drive PCI in my design. Too bad
the plug got pulled before there was a prototype, it would have been
everyone's favorite Zorro III graphics card :-)
Dave Haynie | ex-Commodore Engineering | See my first film
Sr. Systems Engineer | Class of '94 | "The Deathbed Vigil"
Scala Inc., US R&D | C= Failure n. See: Greed | in...@iam.com
"It's the little things that kill" -Bush
Richard R Spagnola
>course, I wasn't using PCs or CPUs to drive PCI in my design. Too bad
>the plug got pulled before there was a prototype, it would have been
>everyone's favorite Zorro III graphics card :-)
>
>Dave Haynie | ex-Commodore Engineering | See my first film
>Sr. Systems Engineer | Class of '94 | "The Deathbed Vigil"
>Scala Inc., US R&D | C= Failure n. See: Greed | in...@iam.com
>
> "It's the little things that kill" -Bush
Ok, give us the story on this project.
>course, I wasn't using PCs or CPUs to drive PCI in my design.
What were you using??
Thanks
\ /
- * -
/ \
Richard Spagnola ============================
Office Of Radiation Safety ======== RADIATION IS OUR FRIEND ========
Ohio State University ============================
rsp...@magnus.acs.ohio-state.edu \ /
- * -
/ \
Dave Haynie
> Sorry to jump into this thread a bit late, but I don't have enough time
>to keep up with the traffic in this group.
>>The reasons for these [deficiencies and cost] are simple -- AAA is old.
>It's true that today high end PC cards are faster than AAA. But AAA has
>some other advantages.
> - Compressed modes such as HAM or PACK-LUT
Made obsolete by MPEG in hardware. MPEG compressed motion video takes
far less memory than any HAM, PACKLUT or PACKHY mode ever dreamed of,
and looks like "video" rather than "computer graphics".
> - Compatibility with ECS, so old software may run on it.
A level of compatibility, targeted at OCS, not ECS. But even there,
it's impossible to be completely compatible. AAA supported a set of
OCS compatible registers, not an OCS (or ECS or AA) compatibility
mode. So it's only the very basic bit-level stuff that remains at all
compatible. Anything dependent on blitter timing or bus loading is
broken. Copperlists aren't even compatible when you go to the 64-bit
version of the system.
And just how much useful software does OCS or ECS compatibility buy
you? We're primarily takin' games here, folks. Most everything else
runs under RTG, or couldn't wait for anything better in the graphics
department. For the most part, no one play old games. They get so much
better every year, except for maybe one favorite everyone keeps
around, they get replaced.
> - And the most important one, unique to Amiga chipsets :
> _THE_COPPER_ !
Sure, you can't get a copper in any standard graphics chipset. But
you'll find, in a modern system, it's far less useful. Any PCI based
graphics chip is fast enough to given you screen banking out of Fast
RAM within a frame, so to the eye it's as fast as the copper would do
it, but not dependent on Chip RAM. With direct mapped graphics,
there's no need to overload color registers or make clever splits
between HAM and text-friendly display modes, etc. With all the extra
video buffer bandwidth available today, syncing to the exact
horizontal line isn't necessary for smooth graphics, and you can at
least double buffer with most off the shelf chipsets, cleanly.
OK, you might find a few uses for the copper, still. The question is,
what percentage of all Amiga software makes good and important use of
this. How much of that really sets the Amiga apart from other systems,
in the buyers minds? And is the $1 million bucks and year's effort to
deliever an expensive chipset that's otherwise underperforming today's
off-the-shelf graphics chips going to pay off? Are people going to buy
AAA cards for their Amiga 5000s, or rather go for the latest S3 or
Matrox with 3D, MPEG, etc. (sure, I'm assuming ESCOM does the
intelligent thing and get out a new system with PCI bus and RTG, at
the least). Would those resources benefit the Amiga better elsewhere?
You have to realize that there's not much room for screwups, if any,
if you want to see the Amiga take off again. It's way behind schedule
already, and any misdirection of effort is probably going to kill it
again, probably for good this time. Like Jack Tramiel used to say,
"Business is War", and in the PC business, you generally find the
bullets are constantly flying, there's no safe haven, and those that
do not adapt become causalties.
> Of course, you could emulate part of it in software,
What's to emulate? I have been running an RTG system since '91 or
so. It works good, and this was something designed outside the Amiga
company, with no access to graphics, Intuition, etc. When it's adapt
or die, I say it's time to evolve already.
>Also AAA is not only a graphics subsystem. There's also a sound and I/O chip
>in it. The sound part has 8 16 bit channels,
Sure does, but they're way behind the times on that. These are simply
DMA driven channels, like the OCS chipset, only fatter. I have a sound
card in my PC with 34 16-bit channels, including volume, pan, and
evelope settings for each one. Two play from main memory via (cheesy
ISA-bus) DMA channels, while 32 play from on-chip sample RAM, complete
with patch settings and all. In other words, folks, it's a
synthesizer, it'll do real MIDI or whatever else you want to throw at
it, eliminating the burden on the CPU. That's what the Amiga has
always been about, eh?
>the I/O part has a buffered serial port
A dime a dozen, this is a standard cell in anyone's IC library these
days, virually every PC out these days has two 16550s (or the
equivalent), which have 16 bytes of buffer, rather than the four byte
buffers on each of the AAA serial channels. The nice thing about the
Amiga serial port, it's device independent.
>and a pretty good floppy controller. This one is totally programmable
>and can read every possible floppy format, even the infamous low
>density MAC disks. How many PC controllers are able to read Amiga
>disks ?
The thing is, who's going to care all that much about conventional
floppies in another year. The future is somewhere around 100MB, as
witnessed by the hottest selling add-on drive this year, the IOMega
ZIP drive. Compaq's already looking into building something similar
and perhaps even cheaper. When you're talking about CD-quality audio
at 5MB/minute, compressed video at several times that, software
packages of 100MB or more, etc. the old floppy doesn't cut it, even if
you do go to 4MB density (which the AAA supports, but at a drive price
that could very soon look foolish next to one of these MOs). These bad
boys can usually support a variety of traditional floppy formats,
too. And they can live on SCSI, just like your HDs and your CD-ROM.
>>AAA, at least as it exists today (or could exist with finished
>>versions of today's chips) is too expensive for an A1200, and in fact
>>wouldn't physically fit in one. You're taking about four large chips
>>for the basic system, plus lots of support logic that has yet to be
>>designed (a four-channel clock synthesizer, general Gary-style glue,
>>smart buffering, etc).
> AA is already 3 chips. One more should be possible in a low end
> machine.
You're comparing two 84 and one 48 pin chip (AA) with two 144 and two
160 pin chips (AAA's anticipated final form). Not to mention all the
extra 32-bit buses running around.
>And AAA chips use less power than current AA chips
Where in the world did you get that idea? Just because they're CMOS?
Anyone care to guess whether or not the NMOS 68000 uses more power, or
less, than the CMOS 68040? The comparison is a valid one, for exactly
the same reasons.
>The chipset was said to be 3 months to completion when ex C= decided
> to suspend its development.
The chipset was approximately 3 months from the next prototype
turnaround, assuming those last tapeouts were real. That means, chips
come back, maybe they work, maybe there are still errors. If there
are, you (as ESCOM) have a problem, since you don't have any of the
designers around to fix them.
>If Amiga Technologies resume it now with new engineers,
>then should be able to have it done for early 96.
You might be underestimating the amount of work necessary to bring new
people up on a million+ transistor design. Sure, most of it's in a
high level design language, but most engineers know VHDL or Verilog,
few know M (a shame, too, it's superior to either of the "standards").
>AFAIK only the Andrea chip (DMA controller) needs major
>modifications.
Just where did you get your information? Keep in mind, I'm the only
person to have designed and built a AAA system. While I don't know
what's going on in the chips per se, I know a bit more than the
average guy on the virtual street. Like, for instance, that due to a
fairly major bug in Andrea, large portions of the last chipset were
untestable.
> A complete RTG OS was said to require over 12 months development
> time.
Don't believe what marketroids or others with alterior motives have to
say over the engineers who actually work on the stuff. I don't have
all the answers, but I can base what I am saying on 10.5 years of
computer design experience and a world class knowledge of the Amiga
architecture, both what's out there and what it could be.
>Basically intuition and graphics libraries have to be entirely
>rewritten.
Several companies already have proprietary RTG systems running on SVGA
chips. No matter what you claim, an existence proof tends to sway the
case toward an SVGA driver being easier to do than a reasonable AAA
driver, simply because the former already exists, and could be adapted
to any particular chipset in a matter of a day or two, if ESCOM chose
to license code from a third party.
Dave Haynie
>>course, I wasn't using PCs or CPUs to drive PCI in my design. Too bad
>>the plug got pulled before there was a prototype, it would have been
>>everyone's favorite Zorro III graphics card :-)
>Ok, give us the story on this project.
After AAA was cancelled (permanently tabled, whatever), we were given
the go-ahead to do some smaller projects. They were going to resurrect
the '040/'060 interface device from 1992, and design a Zorro III
graphics card using off the shelf graphics chips and either gate
arrays or FPLDs.
I got the graphics card project. Since graphics chips change all the
time, I decided the best course of action would be to design a general
Zorro III to PCI bus converter, then I could pick the SVGA chip of
choice, based on performance desires and the evaluation of various
chips in the software group. In fact, I had planned to put a normal
PCI card socket on the first prototype, to assist in evaluation, but
we never made it that far.
Along with the Zorro III to PCI converter, I had one other custom
circuit driving the PCI bus. This guy sucked up data from the video
slot, formatted it, FIFOed it, and eventually spit it out as long PCI
burst cycles. The idea here was that, without going to any real crazy
frame store architecture, this bit could provide scan conversion for
Amiga display modes. With an intelligent address generator (nothing
real fancy, just a counter with modulos), scan converted data could be
directed to a window within an SVGA display. So overscan wasn't a
problem, weird Amiga resoultions weren't a problem, as you could
always pick an SVGA mode that could contain them.
The scan converter could handle Amiga pixels down to 35ns, and it
could handle scanning of AA's 24-bit or ECS's 12-bit color data,
with optional conversion to 16 or 8 bit pixels, respectively. I needed
about 100MB/s throughput over PCI to graphics store to get a full
24-bit display converted at 35ns pixels. This is the only way to
convert HAM, for instance. SVGA chips fast enough were just coming on
the market around that time, though just barely.
Where it gets interesting, though, it's simply in the scan
conversion. Since the conversion takes place on the fly, with a number
of programmable options like line length, etc., this was also suitable
as a replacement for the fancy "Motivator" scan converter for AA that
Joe Augenbraun and others had worked on at Commodore. The idea here is
that you send over a high resolution display using slow scanning, like
an 800x600 display at a real 35Hz, rather than interlaced at
72Hz. Using this technique, plus AA's CLUT-bypass mode, you could get
displays like 800x600, 1024x768, even 1280x1024 in 8 bits/pixel,
without saturating the chip bus. Drivers for something like this are
easier than, for instance, the A2024/Moniterm, and would have allowed
a bridge between AA conversion and full RTG on this card.
And of course, once a full RTG was available, the card could be wholly
driven over Zorro III, using its native blitter, etc., for a dramatic
speedup over any scan-conversion related mode. And the scan
converter's still there to support legacy modes.
Adrian Demarais
: In <3rp5ra$5...@upsn2.u-psud.fr>, vo...@mesiom.obspm.fr (Etienne Vogt) writes:
: > Sorry to jump into this thread a bit late, but I don't have enough time
: >to keep up with the traffic in this group.
8<--- Mega-snip!! --- (long, interesting discussion of AAA)
As a temporary stopgap until new chipsets of whatever sort can
be created, how hard would an 8M AA Agnus be? That 8 (16?)M chip
jumper on my A4000 motherboard is really annoying.
Richard R Spagnola
Thanks for the info on the graphics card, hey how much money do ya want to
finish that design, your right that would have been everyone's favorite card.
You must be very sad that your old job melted-down. :(
>After AAA was cancelled (permanently tabled, whatever), we were given
>the go-ahead to do some smaller projects. They were going to resurrect
>the '040/'060 interface device from 1992, and design a Zorro III
>graphics card using off the shelf graphics chips and either gate
>arrays or FPLDs.
Ok whats the '040/'060 interface device from 1992
Dennis Grant
(Dave Haynie) wrote:
[much wisdom from Dave deleted]
Time for a reality check here folks.
The day of the custom hardware Amiga is *gone* Back when the common
denominator was a 16 bit processor on an 8 bit bus, it made sense to spend
millions on developing special hardware. But these days, 128 bit graphics
subsytems, 16 bit DSP's, and 64 bit data busses are off the shelf items.
It makes no sense to try and re-invent the wheel.
What stands the Amiga apart from the rest of the crowd these days is the
_operating system_ AmigaDOS does now what the "other guys" are evolving
toward, with the notable exception of RTG. It multitasks, it's device
independant, it's small, and it's fast.
An Amiga, based on a PCI bus and a PowerPC processor, built around the new
PowerPC spec developed by IBM and Apple, actually stands a chance. Take
AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
and add RTG, and you've got a winner.
The only special hardware I could see on a new Amiga would be a Zorro 3
slot or 2, alongside the PCI slots, so that older cards could be used.
That's it.
Hardware wars are over. Now it's the OS wars.
---------------------------------------------------------------------
Dennis Grant Cycor Tech (Amiga&Mac) Support / HTML & Graphic Design
dgr...@cycor.ca Amiga 4000/040/14/970/17"IDEK/2XCD-ROM/14.4 AmiTCP
http://www.cycor.ca/TCave/ Visit Trog's Cave!
Michael van Elst
>millions on developing special hardware. But these days, 128 bit graphics
>subsytems, 16 bit DSP's, and 64 bit data busses are off the shelf items.
>It makes no sense to try and re-invent the wheel.
This just says that everyone is now using "custom hardware".
>AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
>and add RTG, and you've got a winner.
Does it run Word ? Where are the Norton utilities ? Where can I copy DOOM ?
>Hardware wars are over. Now it's the OS wars.
Tell that the c00l c0d3rz :)
Regards,
--
Michael van Elst
Internet: mle...@serpens.rhein.de
"A potential Snark may lurk in every tree."
Jonas Elfström
>In <3rqfbq$d...@charm.magnus.acs.ohio-state.edu>, rspa...@magnus.acs.ohio-state.edu (Richard R Spagnola) writes:
>
-----8<---terrific text about Dave's videocard project----8<-----
>
>And of course, once a full RTG was available, the card could be wholly
>driven over Zorro III, using its native blitter, etc., for a dramatic
>speedup over any scan-conversion related mode. And the scan
>converter's still there to support legacy modes.
>
>Dave Haynie | ex-Commodore Engineering | See my first film
>Sr. Systems Engineer | Class of '94 | "The Deathbed Vigil"
>Scala Inc., US R&D | C= Failure n. See: Greed | in...@iam.com
>
> "It's the little things that kill" -Bush
>
How long time do you think you should need to finish this project
and how money would it cost? :-)
This card as an add-on for A3000 and A4000 would be a smash hit!
(and if you let the Cybergraphics team into it, you're having
a big seller, if the prize isn't astromical)
This could be a good investement for ESCOM...
--
Jonas Elfstroem !The more you suffer !ECS...@klecs1.ericsson.se
ERICSSON MOBILE !The more it shows you! ->voice: +46 19 584455<-
COMMUNICATIONS !really care Right?! /*amigafreak, R.E.M. */
This is my point of view not Ericsson's /*Offspring, SOM, DM */
Christian Brandt
In article <3822...@kone.fipnet.fi>, "Jyrki Saarinen" <jsaa...@kone.fipnet.fi> writes:
|> Are we talking about the same thing? I am talking about how
|> much one can move data to the video boards memory? And
|> this is usually about 10-20MB/s on VLB/PCI machines and
|> 320x200x256 mode. Some cards like ET4000/W32p are faster,
|> like 30MB/s.
Nice way to misinterpred this :-)
The "faster ET4000W32p" is rather a very lowcost-solution and
veryvery slow - 20-30MB/s.
The better S3 864 64Bit-Systems do around 40MB/s on PCI and
sometimes up to 60MB/s. There are still some 128Bit-Systems out
and they increase Speed by another factor of four (I must state,
internal memorytransfer, not external, we have to wait for
burst-mode 64Bit-PCI but then we get up to 260MB/s).
--
Christian Brandt
- The Stand - von und mit Thomas Boerkel - Das letzte Gefecht -
demnaechst in de.comp.sys.amiga.*
Windows ist ein Betriebssystem fuer Bloede und deswegen sehr erfolgreich.
Christian Brandt
|> > - Compressed modes such as HAM or PACK-LUT
|>
|> Made obsolete by MPEG in hardware. MPEG compressed motion video takes
|> far less memory than any HAM, PACKLUT or PACKHY mode ever dreamed of,
|> and looks like "video" rather than "computer graphics".
Thanks on your comments on ham, copper and ocs/aaa-compability.
Theres a rather awful discussion in a local echo conserning traditional
Amigacustomchips vs. Standardsolutions and your post fits perfect
to explain many things. And, maybe its a way to stop the local gibberish
like "copper makes an A500 supperior over every pc" :-)
Christian Brandt
Etienne Vogt
>> - And the most important one, unique to Amiga chipsets :
>> _THE_COPPER_ !
>
>Sure, you can't get a copper in any standard graphics chipset. But
>you'll find, in a modern system, it's far less useful. Any PCI based
>graphics chip is fast enough to given you screen banking out of Fast
>RAM within a frame, so to the eye it's as fast as the copper would do
>it, but not dependent on Chip RAM. With direct mapped graphics,
>there's no need to overload color registers or make clever splits
>between HAM and text-friendly display modes, etc. With all the extra
>video buffer bandwidth available today, syncing to the exact
>horizontal line isn't necessary for smooth graphics, and you can at
>least double buffer with most off the shelf chipsets, cleanly.
Can you drag screens nicely with these SVGA chips ? On the Picasso II, it
looks horrible as the palette can't be changed on the fly. And using true
color modes for everything would be a waste of resources.
AAA still supports planar screens so you don't have the overhead of a 256+
color screen when you only need fewer colors for a particular application.
>The thing is, who's going to care all that much about conventional
>floppies in another year. The future is somewhere around 100MB, as
>witnessed by the hottest selling add-on drive this year, the IOMega
>ZIP drive. Compaq's already looking into building something similar
>and perhaps even cheaper. When you're talking about CD-quality audio
>at 5MB/minute, compressed video at several times that, software
>packages of 100MB or more, etc. the old floppy doesn't cut it, even if
>you do go to 4MB density (which the AAA supports, but at a drive price
>that could very soon look foolish next to one of these MOs). These bad
>boys can usually support a variety of traditional floppy formats,
>too. And they can live on SCSI, just like your HDs and your CD-ROM.
You will still need floppies for some time (At least for '96 machines.)
Lots of software is still distributed on them. And when you need to exchange
"small amounts" of data with friends, a floppy disk is more convenient than
expensive MO disks. "floptical" drives are great, but they don't read the
Amiga floppy format, so they can't replace them completely as in a PC clone.
A new amiga will need a custom floppy controller or at least an "off the shelves"
one that supports raw reading from the drive.
>>Basically intuition and graphics libraries have to be entirely
>>rewritten.
>
>Several companies already have proprietary RTG systems running on SVGA
>chips. No matter what you claim, an existence proof tends to sway the
>case toward an SVGA driver being easier to do than a reasonable AAA
>driver, simply because the former already exists, and could be adapted
>to any particular chipset in a matter of a day or two, if ESCOM chose
>to license code from a third party.
These third parties "RTG systems" are hacks just like the one I proposed
for early AAA machines. For full RTG, you still need to rewrite intuition
and gfx, not just SetFunction()ing the existing libraries.
Sure, the Amiga will only survive if it evolves to keep up in hardware with
PCs and PowerMacs. But it also needs to be different, and software will not
do it if AmigaOS is to be ported to other platforms. A new Amiga will need
some hardware compatibility with current models and support for Zorro boards
and Amiga floppy format is the minimum people will expect.
Etienne Vogt (Etienn...@obspm.fr)
Graduate Student at Meudon Observatory, France
A500/000 3.1 & C= 128 at home
Maxwell Daymon
: "small amounts" of data with friends, a floppy disk is more convenient than
: expensive MO disks. "floptical" drives are great, but they don't read the
: Amiga floppy format, so they can't replace them completely as in a PC clone.
They were going to until Commodore f'ed it up. Computer Shopper (US clone
magazine) reported the Amiga as the first machine to ship with a
floptical as standard equipment - then Commodore dropped the whole thing.
I believe IOmega was all geared up to write a ROM version that works with
the Amiga diskchange, disk format, etc.
: A new amiga will need a custom floppy controller or at least an "off the shelves"
: one that supports raw reading from the drive.
The Floptical is SCSI. You just don't get much more flexible than that.
: These third parties "RTG systems" are hacks just like the one I proposed
: for early AAA machines. For full RTG, you still need to rewrite intuition
: and gfx, not just SetFunction()ing the existing libraries.
Well OBVIOUSLY they wouldn't be SetFunctioning if they had the source
code to recompile the stuff! In fact, with access to the OS they could do
a BETTER and more seamless job of RTG. The "hacks" are essentially full
RTG systems just waiting to be implemented.
--
+- Maxwell Daymon -+- mda...@rmii.com -+
Jon M. Taylor
Dennis Grant <dgr...@cycor.ca> wrote:
>In article <1995Jun19.1...@scala.scala.com>, dave....@scala.com
>(Dave Haynie) wrote:
>
>[much wisdom from Dave deleted]
>
>Time for a reality check here folks.
>
>The day of the custom hardware Amiga is *gone*
Then the AMiga itself is *gone*, IMHO.
>Back when the common
>denominator was a 16 bit processor on an 8 bit bus, it made sense to spend
>millions on developing special hardware. But these days, 128 bit graphics
>subsytems, 16 bit DSP's, and 64 bit data busses are off the shelf items.
>It makes no sense to try and re-invent the wheel.
What the hells the point of even *HAVING* an Amiga without the
custom hardware?
- The CPU is better, but unless you code to the metal this won't be visible,
- The expansion bus has been topped by PCI,
- The OS, while good, lacks too many modern features,
Without the chipset, all you have is a nonstandard system running
a nonstandard OS. It has no saleable features.
>What stands the Amiga apart from the rest of the crowd these days is the
>_operating system_
The Amiga OS has no virtual memory, task protection, device
independance/RTG, and many other "modern" OS features. We can appreciate
it for its compactness and speed, but other people will just see that it
is no better than windows when it comes to one task being able to bring
down the whole system.
>AmigaDOS does now what the "other guys" are evolving
>toward, with the notable exception of RTG. It multitasks, it's device
>independant, it's small, and it's fast.
See above.
>An Amiga, based on a PCI bus and a PowerPC processor, built around the new
>PowerPC spec developed by IBM and Apple, actually stands a chance.
Great. All this would be is another PPC clone box that happens
to run a port of AmigaDOS. If I get a PPC box, I ain't running anything
except Linux on it.
>Take
>AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
>and add RTG, and you've got a winner.
I fail to see any selling point at all for this system. The OS by
itself will go over like a lead balloon. Why do people want to give up
the chipset for the OS? I just don't get this.
--
************ Jon Taylor *************************************************
* "For something that has spread with all the forethought of kudzu, the *
* Internet isn't half bad." - Newsweek, 2/27/95 *************************
************************************ tay...@gaia.ecs.csus.edu **********
Glenn Saunders
: I fail to see any selling point at all for this system. The OS by
: itself will go over like a lead balloon. Why do people want to give up
: the chipset for the OS? I just don't get this.
Well, do you concede that you have to give up the 68K series in favor of
RISC? If you you pretty much have to give up AGA because how well will
AGA like being driven by a foreign CPU?
The OS is the only thing that will survive the move to RISC. AAA would
have been another AGA i.e. too little too late, and Hombre is still a
ways off yet which means we're going to have to deal with whatever chips
other 3rd parties provide in the meantime.
You talk about how the Amiga IS the chipset (which I do agree with you
on) but you offer no solutions that would allow an analogous chipset to
survive the necessary changes. Are you saying the Amiga is doomed? I'd
rather see SOMETHING survive than nothing. Even if it flops, I'm not
going to discourage Escom from doing it in the first place.
I think multiple 060s might work as a CISC stalling tactic as a 3rd party
upgrade but that's all. That doesn't solve ZII/ZIII bus speed limitations
or chipram bottlenecks, and so on. There is only so much patching you can
do onto the old workhorse before you realize the need to start over from
scratch.
Jon M. Taylor
Glenn Saunders <kri...@primenet.com> wrote:
>The mortal Jon M. Taylor wrote:
>: I fail to see any selling point at all for this system. The OS by
>: itself will go over like a lead balloon. Why do people want to give up
>: the chipset for the OS? I just don't get this.
>
>Well, do you concede that you have to give up the 68K series in favor of
>RISC?
Not at all.
>If you you pretty much have to give up AGA because how well will
>AGA like being driven by a foreign CPU?
Not well at all, especially by a RISC CPU. I really don't see
very many programs running on Amigas that are CPU-bound enough to make a
68060 nonviable.
>The OS is the only thing that will survive the move to RISC. AAA would
>have been another AGA i.e. too little too late, and Hombre is still a
>ways off yet which means we're going to have to deal with whatever chips
>other 3rd parties provide in the meantime.
Which means that we will have to run the graphics through an OS
layer, which kills speed/smoothness. Again, I see no advantage to using
an Amiga if this approach is taken.
>You talk about how the Amiga IS the chipset (which I do agree with you
>on) but you offer no solutions that would allow an analogous chipset to
>survive the necessary changes. Are you saying the Amiga is doomed?
Not at all. I think that staying with the 680x0 CPU family for
at least a couple more years would be the wisest move. The chips are
cheap, people know them well, the OS wouldn't need to be ported, and the
custom chipset wouldn't break.
>I'd
>rather see SOMETHING survive than nothing. Even if it flops, I'm not
>going to discourage Escom from doing it in the first place.
Me either. Hell, *anything* would be better than what we
had/have....
>I think multiple 060s might work as a CISC stalling tactic as a 3rd party
>upgrade but that's all. That doesn't solve ZII/ZIII bus speed limitations
>or chipram bottlenecks, and so on. There is only so much patching you can
>do onto the old workhorse before you realize the need to start over from
>scratch.
Which isn't a bad idea at all, especially for a low-end machine.
Maxwell Daymon
: Which means that we will have to run the graphics through an OS
: an Amiga if this approach is taken.
Then we'll lose you. So what? We'll gain others who see and respect the
Amiga for what it is. The Amiga market will get the people who were on the
fence - they liked the potential that was never fulfilled because
Commodore catered to a bunch of people who were trying really hard to
choose between a Nintendo and an Amiga. People like you who simply see the
Amiga as a toy have not helped. Yes, the Amiga got some short term sales
from that, but certainly not the long term support it needed to survive.
Had Commodore done it right and avoided people like you in the first
place, the Amiga might not be dead now.
: Not at all. I think that staying with the 680x0 CPU family for
: at least a couple more years would be the wisest move. The chips are
: cheap, people know them well, the OS wouldn't need to be ported, and the
: custom chipset wouldn't break.
Keeping such a chipset would be moronic at best. The Amiga will survive
as what YOU want it to be: a toy for the playroom. There's too much
wasted power in the OS for me to be "okay" with that. I'm happy that
Escom has a much higher opinion of the system and sees potential you are
incapable of seeing.
Richard Rauch
>In article <dgrant-2106...@bigmac.cycor.ca>,
>Dennis Grant <dgr...@cycor.ca> wrote:
>>Back when the common
>>denominator was a 16 bit processor on an 8 bit bus, it made sense to spend
>>millions on developing special hardware. But these days, 128 bit graphics
>>subsytems, 16 bit DSP's, and 64 bit data busses are off the shelf items.
>>It makes no sense to try and re-invent the wheel.
> What the hells the point of even *HAVING* an Amiga without the
>custom hardware?
If custom hardware is all it's about, what's the point in having one, now?
>- The CPU is better, but unless you code to the metal this won't be visible,
Maybe we won't be locked into 1 CPU. (^& But, if we are, it still is nice to
have a 68k in there, even from a C programmer's point of view, IMHO. One
generally gets to know the assembly language at least well enough to
read it, if not to patch machine-code by hand. (^& Having a "nice"
assembly language is nice.
>- The expansion bus has been topped by PCI,
Something to hope the Amiga encompasses then, yes?
>- The OS, while good, lacks too many modern features,
Let's see...multi-user, and virtual memory. Both are available through
3rd-party hacks (though it WOULD be nice to have in the OS itself). What
else? RTG? The whole point of this discussion is RTG, rather than hacking
with custom hardware, right? And here you are, arguing AGAINST RTG?
Suppose the Amiga OS were ported to a clone, and you found yourself
forced by desperation to buy a clone. Wouldn't you get the Amiga OS
for it, if only for personal use? I would give it very serious thought,
at the very least. Further, when porting the OS (to WHATEVER CPU),
it might be feasible to look into changes that "proper" software should
be able to encompass without too much hassle (which could be lost in
the overall concern of porting any given application). Once a large
enough body of "properly written" software existed, then the "native"
platform could also adopt the features, relatively painlessly.
It seems to me to be at least a very good possiblity, and that it would
be a positive thing. And, too, that it could be a singnificant
"migration path" for those 2 or 3 "modern OS features" you are pining for.
> Without the chipset, all you have is a nonstandard system running
>a nonstandard OS. It has no saleable features.
And with the chipset, you just add "nonstandard display hardware" (and,
incidentally, aging hardware). What's so nice about the bulit-in
hardware, really? Aside from being really nice for 1985-era technology
and giving us a base-level display system & sound system that is at
least usable, how much good does it really do us at this point?
>>An Amiga, based on a PCI bus and a PowerPC processor, built around the new
>>PowerPC spec developed by IBM and Apple, actually stands a chance.
> Great. All this would be is another PPC clone box that happens
>to run a port of AmigaDOS. If I get a PPC box, I ain't running anything
>except Linux on it.
Your choice, I suppose. Myself, I'd look very seriously at running an
Amiga OS on any platform that I had and that it would run on. (Right now,
both factors restrict me to the 68k...(^&)
>>Take
>>AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
>>and add RTG, and you've got a winner.
> I fail to see any selling point at all for this system. The OS by
>itself will go over like a lead balloon. Why do people want to give up
>the chipset for the OS? I just don't get this.
Better display, faster, same nice OS to program with, same
"ambidextrous" user-interface, less cost to develop new Amigas if they
are using standard parts (== less cost to the end-user).
Seems like a great improvement. If you won't buy an improved Amiga, then
it seems reasonable to count you out as a potential buyer for the _old_
Amiga...so what are you worrying about? You won't buy another one, anyway.
For myself, the biggest hinging point is probably the OS, and the hardware
can go hang, as long as Intuition runs nicely one whatever's produced.
--
"I probably don't know what I'm talking about." --r...@interstate.net
Glenn Saunders
: Which means that we will have to run the graphics through an OS
: layer, which kills speed/smoothness. Again, I see no advantage to using
: an Amiga if this approach is taken.
But they will have the advantage of faster CPUs, a faster BUS, and faster
GFX chips. It will be a net gain and a layered AmigaOS probably won't be
anywhere near as bad as it is in some other environments I won't name.
: Not at all. I think that staying with the 680x0 CPU family for
: at least a couple more years would be the wisest move. The chips are
: cheap, people know them well, the OS wouldn't need to be ported, and the
: custom chipset wouldn't break.
As far as number crunching, that's all well and good. But supposedly the
Amiga is a multimedia machine but with 8-bit sound and slow 8-bit planar
gfx in 1997 it will be laughed at as things like 3D cards enter the
mainstream. Don't get me wrong. I think AGA trudges on like a valiant
warrior when it comes to trying to keep up with multimedia with things
like HAM8 and 14-bit sound, no 8-bit VGA Klone w/8-bit soundblaster would
be able to pull such tricks, but it's not an ideal situation.
: >I think multiple 060s might work as a CISC stalling tactic as a 3rd party
: >upgrade but that's all. That doesn't solve ZII/ZIII bus speed limitations
: >or chipram bottlenecks, and so on. There is only so much patching you can
: >do onto the old workhorse before you realize the need to start over from
: >scratch.
: Which isn't a bad idea at all, especially for a low-end machine.
I think you have a false idea of chip prices.
The 68K series were always too expensive compared to their Intel
counterparts and was one factor in CBM using shittier, older variants of
the 68K rather than the newest and the most competitive ones to keep the
prices lower.
Despite the early statements about 060s supposedly being sold cheaper
than 040s, don't expect them to be cheap enough to be put into "low end"
Amigas. Maybe not even in 2 years time. I'll believe it when I see it.
God knows I'd like it to be so but it doesn't seem likely.
I don't necessarily think it's an either-or situation.
nyle landas
: Jon M. Taylor (tay...@ecs.ecs.csus.edu) wrote:
: : Which means that we will have to run the graphics through an OS
: : layer, which kills speed/smoothness. Again, I see no advantage to using
: : an Amiga if this approach is taken.
: Then we'll lose you. So what? We'll gain others who see and respect the
: Amiga for what it is. The Amiga market will get the people who were on the
<STUFF CUT>
: Had Commodore done it right and avoided people like you in the first
: place, the Amiga might not be dead now.
The Amiga can not afford to avoid any people who are potential customers.
Amiga Technologies will need every Niche it can find even JTaylor.
: : Not at all. I think that staying with the 680x0 CPU family for
: : at least a couple more years would be the wisest move. The chips are
: : cheap, people know them well, the OS wouldn't need to be ported, and the
: : custom chipset wouldn't break.
: Keeping such a chipset would be moronic at best. The Amiga will survive
: as what YOU want it to be: a toy for the playroom. There's too much
: wasted power in the OS for me to be "okay" with that. I'm happy that
: Escom has a much higher opinion of the system and sees potential you are
: incapable of seeing.
I agree we need to move RISC but I still cannot believe that we need to
totally scrap the hardware that makes an Amiga an Amiga. If we simply go
SVGA then why not run OS/2, Windows NT, Windows 95, etc. It's going to be a
hard sale to people who don't understand why AmigaDos is better and only
listen to the industry. If we go SVGA we'll loose a lot of the stuff that
made the Amiga #1 for video work. We need to keep the niche markets while
moving onto other markets as well.
In my mind this requires -
1.) Port to RISC
2.) RTG
3.) A new custom chipset
4.) PCI
With the new custom chipset you still supply a "lowest common denominator"
that will be available for the low-end to the high-end. With RTG and PCI
you supply inexpensive SCGA cards for people who need that kind of power.
As well as many other inexpensive add on cards. If we want it to be an
Amiga we can't sacrifice totally on either side.
I've seen two camps forming on this...
One says scrap the custom chips AmigaDos is everything
the other says keep the custom chips the combination is everything.
The Amiga can't afford to ignore either camp a compromise has to be drawn
somewhere and I think it's in the rough outline above.
Just some ideas,
-Nyle
/*******************************************************************\
|Nyle F. Landas |
|E-MAIL - ny...@servtech.com |
|URL - http://www.servtech.com/public/ami4000/ |
|FTP - ftp://ftp.servtech.com/pub/users/ami4000/ |
| |
| Build a better mouse trap and Bill Gates will steal all the cheese.|
\********************************************************************/
Brian King
> I fail to see any selling point at all for this system. The OS by
>itself will go over like a lead balloon. Why do people want to give up
>the chipset for the OS? I just don't get this.
Because alot of us having been using the OS *without* the chipset for quite
awhile! Basically, anybody out there with a graphics card knows that the
chipset bites for anything but video-compatible output.
The chipset doesn't provide:
- 16-bit sound
- support for high-density floppies
- FIFO-based UART (buffered serial port)
- Enhanced parallel port
- chunky video modes, high resolution video modes
Hell, the only part of the chipset some people use specifically is Paula,
because they have no choice, the chip-ram, and the CIA timers.
You don't want to spend money on designing all the custom chips to support
the above items, and still be stuck with the same basic OS and 68040/060.
Why not spend the money on OS enhancement (add VM and RTG and RTS), and
spend the money on a port to PCI/PowerPC?
Besides, ESCOM could sell the machine as a Commodore PowerPC, with OS/2 or
Win-NT installed (or even MacOS), and Amiga Technologies could sell the same
machine with AmigaOS. They may sell fewer machines (say only 500,000
instead of 1-million other PowerPC machines), but the hardware costs are
reduced due to amount produced...
--
+-------------------------------------+-----------------------------------+
| Brian D. King - Computer Engineer | Fourth Dimension Software, author |
| PBX/ISDN Software Development Group | of LhA-GUI, Cert. Amiga Developer |
| Mitel Corporation, Kanata, Ontario | ***** Brian...@Mitel.com ***** |
Mr Gavin Moran
: I agree we need to move RISC but I still cannot believe that we need to
: totally scrap the hardware that makes an Amiga an Amiga. If we simply go
: SVGA then why not run OS/2, Windows NT, Windows 95, etc. It's going to be a
: hard sale to people who don't understand why AmigaDos is better and only
: listen to the industry. If we go SVGA we'll loose a lot of the stuff that
: made the Amiga #1 for video work. We need to keep the niche markets while
: moving onto other markets as well.
: In my mind this requires -
: 1.) Port to RISC
: 2.) RTG
: 3.) A new custom chipset
: 4.) PCI
: With the new custom chipset you still supply a "lowest common denominator"
: that will be available for the low-end to the high-end. With RTG and PCI
: you supply inexpensive SCGA cards for people who need that kind of power.
: As well as many other inexpensive add on cards. If we want it to be an
: Amiga we can't sacrifice totally on either side.
My thoughts exactly.. A new custom chipset that enhances AGA to something
that can keep pace with SVGA cards for tmap games, and that offers higher
rez modes and faster blitter for the user interface. This keeps the low-end
Amiga competitive with cheap PCs, as well as improving the base gfx of the
high end. Add RTG to the OS and a PCI slot and the high end is catered for
in the host of SVGA chipsets coming out on PCI card. A port to RISC will
be needed to keep the Amiga in the frame as a 'workstation' computer, but is
that going to be a profitable option? Lets face it, if all you want is a
fast processor with a high rez gfx card in a box then the PC is always gonna
be a better bet. I'd be more than happy to have a 50mHz 030 in my Amiga,
never mind an 80mHz 060 so for the majority of people the 060 as top end
CPU isn't going to be a limitation.
Whether or not the new gfx chipset can match SVGA cards is utterly irrelevent- its there for video, multimedia and games and the AGA set already does most
of these far better than even the best PC SVGA sets. As long as its compatible
with AGA and offers better performance that will be enough.
Lets assume you could get an 'Amiga' that had a PPC604 and an SVGA chipset
in it, running Windows NT. Just what makes this an 'Amiga' and why would
ANYONE want to buy it when they could almost certainly get a PC or a Power PC
clone a _lot_ cheaper that would do exactly the same job.
gavan
--
email: G.M...@ee.qub.ac.uk | 'There can be only one!'
or gmo...@nyx.cs.du.edu | - The Highlander
Maxwell Daymon
: Maxwell Daymon (mda...@rainbow.rmii.com) wrote:
: : Then we'll lose you. So what? We'll gain others who see and respect the
: : Amiga for what it is. The Amiga market will get the people who were on the
: <STUFF CUT>
: : Had Commodore done it right and avoided people like you in the first
: : place, the Amiga might not be dead now.
: The Amiga can not afford to avoid any people who are potential customers.
: Amiga Technologies will need every Niche it can find even JTaylor.
Not true. It is possible to lose 2-to-1 (or worse) if you cater to the
wrong group. You just CAN'T please everyone - so please as many people as
you can who will be there when you need them.
: : Keeping such a chipset would be moronic at best. The Amiga will survive
: I agree we need to move RISC but I still cannot believe that we need to
: totally scrap the hardware that makes an Amiga an Amiga. If we simply go
: SVGA then why not run OS/2, Windows NT, Windows 95, etc. It's going to be a
Because it's not AmigaOS. I can run OS/2 and Win NT *now*, but I run
AmigaOS. In fact, I bought an Amiga just to do it. If people are that
uninterested in the Amiga that they would jump on WinXX or OS/2, it'd be
hard to convince them to buy a custom platform that's guaranteed obsolete
and non-upgradable out the door.
: listen to the industry. If we go SVGA we'll loose a lot of the stuff that
: made the Amiga #1 for video work. We need to keep the niche markets while
: moving onto other markets as well.
No, if we go SVGA *only* we'll lose out. If we go RTG - we can have the
best of all worlds.
: 1.) Port to RISC
: 2.) RTG
: 3.) A new custom chipset
: 4.) PCI
I agree, in a different order.
1) RTG (can be out the door faster than any of the others)
2) RISC
3) PCI
Maxwell Daymon
: Lets assume you could get an 'Amiga' that had a PPC604 and an SVGA chipset
: in it, running Windows NT. Just what makes this an 'Amiga' and why would
: ANYONE want to buy it when they could almost certainly get a PC or a Power PC
: clone a _lot_ cheaper that would do exactly the same job.
Wrong answer. If the Amiga was based on the stuff the rest of the
industry was made of, IT would be a lot cheaper. It's the custom hardware
and designs that make the Amiga expensive. If it adopted industry
standards it could EASILY compete at industry price levels.
AmigaOS. There are quite a few of us on the gfx card mailing lists who
have all but abandoned the Amiga chips (except for the boot menu) and
we're nothing but happy. AmigaOS is a 32-bit, pre-emptive, multi-tasking
OS with strong native support for hardware blitters and a solid design.
REGARDLESS of the hardware - I can't get that elsewhere.
I have PowerMacs and Pentium PCs available for my use whenever I want
them. I don't use them. They just don't let me be productive. The Amiga OS
*does* because I have been able to _overcome_ the Amiga hardware with
other devices.
Dave Haynie
>>> - And the most important one, unique to Amiga chipsets :
>>> _THE_COPPER_ !
>>Sure, you can't get a copper in any standard graphics chipset. But
>>you'll find, in a modern system, it's far less useful.
>Can you drag screens nicely with these SVGA chips ?
Dragging screens around is largely useless one you get to true color
displays, except perhaps for aesthetics. I explained why it's much
less useful for solving real problems today than it was 10 years ago
in a previous article. I'm not saying it isn't cute to be able to drag
screens, but it's not something to spend a million dollars and an
extra year to get.
>On the Picasso II, it looks horrible as the palette can't be changed
>on the fly.
Not between screens. In terms of real time, though, you can change the
palette of any SVGA display faster than the copper can change the AA
palette.
>And using true color modes for everything would be a waste of
>resources.
Why so? If you have dedicate screen memory, you're not wasting
anything using direct mapped color. Plus, mapped color modes are
always available.
>AAA still supports planar screens so you don't have the overhead of a
>256+ color screen when you only need fewer colors for a particular
>application.
SVGA also supports 1-4 bitplanes. Sure, anyone who's a big fan of 5,
6, or 7 bitplane screens would be SOL. Again, that's not something
anyone needs to spend big resourses for, it's just of too little
general use.
>>The thing is, who's going to care all that much about conventional
>>floppies in another year. The future is somewhere around 100MB, as
>>witnessed by the hottest selling add-on drive this year, the IOMega
>>ZIP drive.
>You will still need floppies for some time (At least for '96 machines.)
>Lots of software is still distributed on them.
Floptical drives read and write standard floppy disks. Any format you
care to have programmed into the drive's BIOS; most can support
Amiga formats just fine, though they don't necessarily have these
built-in at present. Another reason to have them backed by ESCOM.
>"floptical" drives are great, but they don't read the Amiga floppy
>format,
Again, unlike plain old PCs, the reason they don't read the Amiga
format is simple: no one told them how. Most of these drives can
support it, if programmed to do so.
>These third parties "RTG systems" are hacks just like the one I proposed
>for early AAA machines.
You can consider them hacks only because the folks who wrote them
didn't have graphics library source. As of OS 3.x, the API was
complete for RTG support, at least for what they had planned at the
time.
>For full RTG, you still need to rewrite intuition and gfx, not just
>SetFunction()ing the existing libraries.
That's what SetFunction() is doing. The main problem with patching
this way is that, unlike RTG, usually only one guy can do it. RTG was
intended to support add-in cards from multiple vendors, and simpler
graphics device drivers. Some of the RTG systems out there already
a device driver level, the monitors system, etc. Adapting one of
these, given sharing of source code between graphics and the RTG
system, would achieve the primary goals of C='s RTG system, very
quickly. We were considering just this approach in the latter days of
Commodore.
>Sure, the Amiga will only survive if it evolves to keep up in hardware with
>PCs and PowerMacs. But it also needs to be different, and software will not
>do it if AmigaOS is to be ported to other platforms. A new Amiga will need
>some hardware compatibility with current models and support for Zorro boards
>and Amiga floppy format is the minimum people will expect.
I disagree. As someone said a few notes back, the hardware wars are
over. The winner is PCI, and everything related to it. Apple has just
dumped NuBus is all new systems, opting for PCI. Sun, similarly, gave
the axe to S-Bus, in favor of PCI. Unless you count Commodore, they
were the last holdouts (C= wasn't a holdout, I decided to move toward
PCI back in '92, but of course we never quite made it there under the
C= management).
When there clearly is a business reason to do custom designs, they
belong in the Amiga. I just don't think you realize what kind of work
ESCOM has in front of them to have any hopes of longterm
success. They don't have the luxery of doing custom chips for any
other reason, and perhaps some aspects of the old Amiga fall in the
process. Regardless, you get a much better Amiga in the end, and when
that's the case, you shouldn't care what's under the hood.
Maxwell Daymon
: SVGA also supports 1-4 bitplanes. Sure, anyone who's a big fan of 5,
: 6, or 7 bitplane screens would be SOL. Again, that's not something
: anyone needs to spend big resourses for, it's just of too little
: general use.
The Picasso II does 5, 6 and 7 bitplanes in planar or chunky format.
No loss there. (?) Am I missing something?
Dave Haynie
>Etienne Vogt (vo...@mesiom.obspm.fr) wrote:
>: "small amounts" of data with friends, a floppy disk is more convenient than
>: expensive MO disks. "floptical" drives are great, but they don't read the
>: Amiga floppy format, so they can't replace them completely as in a PC clone.
>They were going to until Commodore f'ed it up. Computer Shopper (US clone
>magazine) reported the Amiga as the first machine to ship with a
>floptical as standard equipment - then Commodore dropped the whole thing.
The first such deal was in the works some three or four years ago,
with a company called Insite. They made one of the first flopticals
that also read standard floppies. C= was very interested in offering
these as a standard option over the normal DF0:; clearly they were a
better idea when it came to moving multimedia data by sneakernet,
etc. Even those these early ones only did 21MB or so, that was a big
deal difference over the normal floppy. The deal was nixed by new
management at some point or another, but I gather Insite didn't do all
that well anyhow.
>I believe IOmega was all geared up to write a ROM version that works with
>the Amiga diskchange, disk format, etc.
I don't know the ZIP specifically, but it's fairly easy to put a
couple of file systems in these things and let the drive detect what's
installed and adjust accordingly.
Dave Haynie
> I fail to see any selling point at all for this system. The OS by
>itself will go over like a lead balloon.
If you don't want it, don't buy it. But spend some time around a PC or
two before you decide the Amiga OS is all that dated. Sure, it needs
some work, but then again, they all do. There AmigaOS is small enough
to do the same things on home systems as Windows, only for half the
memory. And it can do realtime multimedia, which you simply can't do
under Windows very effectively. This doesn't make it a general
solution, but it does make it viable.
>Why do people want to give up the chipset for the OS?
The Amiga chipset is just plain outdated. While I'm sure ESCOM will
get some sales out of AA machines if they move fast, these aren't
going to last very long. Users are expecting more, and expecting it
fast, especially on the high end.
>I just don't get this.
What I don't get is this virtual "chipset worship" that seems to be a
common thread in these little discussions. Especially by folks who
don't quite realize just what it is down there in the Amiga
chips. Sure, they were years ahead of their time when first
introduced. Unfortunately, not 10 years ahead of their
time. Everything that only the Amiga could do back in 1985 is done
much, much better by chips that anyone can buy. The primary reason
people still like Amigas as general purpose computers (well, religious
issues aside) is that other systems don't take advantage of these
great standard chips. These OSs are slow (for various reasons, most of
them bad), they're poorly designed, and they're memory hungry.
Jyrki Saarinen
> The "faster ET4000W32p" is rather a very lowcost-solution and
> veryvery slow - 20-30MB/s.
>
> The better S3 864 64Bit-Systems do around 40MB/s on PCI and
> sometimes up to 60MB/s. There are still some 128Bit-Systems out
> and they increase Speed by another factor of four (I must state,
> internal memorytransfer, not external, we have to wait for
> burst-mode 64Bit-PCI but then we get up to 260MB/s).
I was talking about how many megabytes per second one
can copy to the videoram, not internal speeds or theoretical
speeds.
Cirrus Logic cards are usually about 10MB/s, so is my friends
S3 805 (VLB), ET4000/W32p-cards are faster.
I guess the speed depends quite much on the motherboard
implementation too.
--
Warped Amiga programmer - Stellar
Maxwell Daymon
: >magazine) reported the Amiga as the first machine to ship with a
: The first such deal was in the works some three or four years ago,
: with a company called Insite. They made one of the first flopticals
^^^^^^
Yes. This was the company.
: that also read standard floppies. C= was very interested in offering
: >I believe IOmega was all geared up to write a ROM version that works with
: >the Amiga diskchange, disk format, etc.
It was *Insite* that was getting a ROM version specifically to support
the Amiga specifics (including disk formats). Now all you can do is 720K
and 1.4M. No 880 or 1.76. :-(
The drive (Insite) is fairly slow, and the 21MB disks cost a lot
compared to newer technology, but it would have been great for the time.
jke...@news.epix.net
: If we simply go
: SVGA then why not run OS/2, Windows NT, Windows 95, etc. It's going to be a
: hard sale to people who don't understand why AmigaDos is better and only
: listen to the industry. If we go SVGA we'll loose a lot of the stuff that
: made the Amiga #1 for video work. We need to keep the niche markets while
: moving onto other markets as well.
How might the advertisment look? Come buy an Amiga! It's better than
the rest. Doesn't run Window/Dos/OS2/Mac software. It's incompatible
with your school or work computers. But it is better!
They're not going to find out that it is better 'cause they're not
going to buy it. If it can use other OS's too (expecially if AmigaOS
is default and runs concurrently with the others) they'll see AmigaOS.
My graphic card is alot faster than my ECS chips (and I hear the AGA
too) and can display anything they can. If I can display a screen at a
good speed, what do I care what hardware is doing it?
Someone mentioned that the AGA chips are inexpensive. Why not include
both AGA and SVGA in new Amigas? The Amiga should use SVGA by default,
but still have the AGA for old stuff. The SVGA should be in it's own
slot and be replaceable by other cards.
To me the Amiga is what I see on the screen and how it operates.
(Shared libraries, device drivers, ect.)
[Jim Kelley AKA Dr. Edge / jke...@epix.net]
Dennis Grant
M. Taylor) wrote:
> In article <dgrant-2106...@bigmac.cycor.ca>,
> Dennis Grant <dgr...@cycor.ca> wrote:
> >Back when the common
> >denominator was a 16 bit processor on an 8 bit bus, it made sense to spend
> >millions on developing special hardware. But these days, 128 bit graphics
> >subsytems, 16 bit DSP's, and 64 bit data busses are off the shelf items.
> >It makes no sense to try and re-invent the wheel.
>
> What the hells the point of even *HAVING* an Amiga without the
> custom hardware?
Look at it this way, when you can get anything you want with off the shelf
parts, cheaper than developing it yourself, what's the point of custom
hardware?
> Without the chipset, all you have is a nonstandard system running
> a nonstandard OS. It has no saleable features.
No, you have a _standard_ machine running an alternative OS.
> The Amiga OS has no virtual memory, task protection, device
> independance/RTG, and many other "modern" OS features. We can appreciate
> it for its compactness and speed, but other people will just see that it
> is no better than windows when it comes to one task being able to bring
> down the whole system.
My 4000 has virtual memory (Thanks to VMM) and will have RTG when my
Cybervision shows up.
Yes, that's cheating. The point is, these are things that can (and should)
be added to AmigaDOS fairly easily. (Especially if it's going through a
port to a PowerPC/PCI architechture)
As for task protection, that debate has been raging since 1985, so I won't
get into that.
However, I can say this: I have observed, with my own eyes, an Amiga
4000/030 beat the pants off a DX2/66 Windows box in real world
performance. That shouldn't have hapened, there's more horsepower in the
Intel box than 3 4000/030's. Windows is a huge anchor dragging behind any
machine that runs it. NT is worse. (My boss has a DX4/100 with 16Mb of RAM
that he's removing NT from because it's too slow) Win95 is better (Yes,
I've seen it) but it still has a way to go. MacOS is sluggish too. Linux
is fast, but it has no support.
AmigaDOS (updated and improved) provides a valid alternative OS, based on
speed alone. Can you imagine a PowerPC AmigaDOS Machine?
> >An Amiga, based on a PCI bus and a PowerPC processor, built around the new
> >PowerPC spec developed by IBM and Apple, actually stands a chance.
>
> Great. All this would be is another PPC clone box that happens
> to run a port of AmigaDOS. If I get a PPC box, I ain't running anything
> except Linux on it.
Enjoy your directory listings, 'cause that's all you'll be able to do with it.
>
> >Take
> >AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
> >and add RTG, and you've got a winner.
>
> I fail to see any selling point at all for this system. The OS by
> itself will go over like a lead balloon. Why do people want to give up
> the chipset for the OS? I just don't get this.
An OS, with a bunch of apps ready to go, that makes my machine run 5 times
faster and actually multitasks?
Sounds like a good selling point to me.
ChrisBig-KahunaRampson
: In my mind this requires -
: 1.) Port to RISC
: 2.) RTG
: 3.) A new custom chipset
: 4.) PCI
: With the new custom chipset you still supply a "lowest common denominator"
: that will be available for the low-end to the high-end. With RTG and PCI
: you supply inexpensive SCGA cards for people who need that kind of power.
: As well as many other inexpensive add on cards. If we want it to be an
: Amiga we can't sacrifice totally on either side.
The lowest common denominator is to take what you have and improve on it
(Not throw it all away and start from scratch). A "port" to a RISC based
system is an expensive and long-term effort, hardly lowest common den. I
guess people that need (or greed?) tip-top graphics have given up on the
co-processors already. I think they would be happy with a PC-CLone that
ran a good multi-tasking OS, as long as the GRAPHICS were there (pant,
pant). I say it may be a long time before AmigaTech can compete in that
arena, and maybe they don't want to. Super graphics is not the ONLY thing
that the Amiga is famous for. The A500 was designed to be a general
purpose machine that could do a lot for the money. I think that this is
the "roots" that AmiTech needs to get back to.
Heres my suggestion:
1. Evaluate if AAA is doable.
2. Make new machines with 060s.
3. Rewrite AmigaOS to use multiple CPUs (should be a smaller effort than
porting!)
4. Make the new Amigas the first home "parallel processing" machine.
5. Get young programmers interested in parallel programming.
6. Make new chipset (with eye toward RISC).
7. Port to RISC, PCI, etc.
This is a good "future" technology that AmiTech could leapfrog into. Could
be hard to do, but no pain, no gain.
: I've seen two camps forming on this...
: One says scrap the custom chips AmigaDos is everything
: the other says keep the custom chips the combination is everything.
: The Amiga can't afford to ignore either camp a compromise has to be drawn
: somewhere and I think it's in the rough outline above.
You can't please everyone, so someone will be "ignored". Probably me, sigh.
No COMPROMISES! Find the BEST solution, and just DO IT!
: Just some ideas,
: -Nyle
--
_/_/_/ _/ _/ Chris "Big Kahuna" Rampson
_/ _/_/_/ _/ _/_/ _/_/_/ ram...@ave050.ve.ford.com
_/ _/ _/ _/_/ _/ _/_/ NO cutdowns, flames, meaness
_/ _/ _/ _/ _/ _/_/ ONLY info. as I interpret it
_/_/_/ _/ _/ _/ _/ _/_/_/ "NOT an OFFICIAL FORD spokesperson"
Mr Gavin Moran
: > What the hells the point of even *HAVING* an Amiga without the
: > custom hardware?
: Look at it this way, when you can get anything you want with off the shelf
: parts, cheaper than developing it yourself, what's the point of custom
: hardware?
Whats the point of an Amiga then?
: > Without the chipset, all you have is a nonstandard system running
: > a nonstandard OS. It has no saleable features.
: No, you have a _standard_ machine running an alternative OS.
Why bother to make Amigas any more - just make the OS for PPC clones????
: However, I can say this: I have observed, with my own eyes, an Amiga
: 4000/030 beat the pants off a DX2/66 Windows box in real world
: performance. That shouldn't have hapened, there's more horsepower in the
: Intel box than 3 4000/030's. Windows is a huge anchor dragging behind any
: machine that runs it. NT is worse. (My boss has a DX4/100 with 16Mb of RAM
: that he's removing NT from because it's too slow) Win95 is better (Yes,
: I've seen it) but it still has a way to go. MacOS is sluggish too. Linux
: is fast, but it has no support.
: AmigaDOS (updated and improved) provides a valid alternative OS, based on
: speed alone. Can you imagine a PowerPC AmigaDOS Machine?
Compared to NT or MacOS it would have little software support and most
old Amiga apps wouldn;t work.
: > >An Amiga, based on a PCI bus and a PowerPC processor, built around the new
: > >PowerPC spec developed by IBM and Apple, actually stands a chance.
: >
: > Great. All this would be is another PPC clone box that happens
: > to run a port of AmigaDOS. If I get a PPC box, I ain't running anything
: > except Linux on it.
: Enjoy your directory listings, 'cause that's all you'll be able to do with it.
LINUX is incredibly well supported and you can easily port any unix stuff to
it. Likewise BSD.
: > I fail to see any selling point at all for this system. The OS by
: > itself will go over like a lead balloon. Why do people want to give up
: > the chipset for the OS? I just don't get this.
: An OS, with a bunch of apps ready to go, that makes my machine run 5 times
: faster and actually multitasks?
What apps would continue to work on a generic PPC machine under a PPC port
of Amiga OS? How would they compare to MacOS or Windows NT apps which are much
more plentiful on those platforms?
: Sounds like a good selling point to me.
Sounds like a recipe for disaster. What would make the Amiga PPC machine
any different to PowerMacs or 3rd party CHRP clones??? Why would people want
to buy it? What level of software support would AmigaOS have on such a
machine?
NT may be slow, but its _very_ powerful. On a fast PPC machine that speed
is not too much of a problem.
I say work on upping the gfx chipset to make the lowend machines more
attractive to home users, and add PCI and rtg to make the high-end machines
more useful to professionals. An 80mHz 060 should be more than enough for
most people as a top-end processor in the near future, and work could be done
in the meantime on a smooth transition to a PPC machine with a specialised
Amiga hardware card for real top-end work. By that point of course you
are competing with x other PPC clone makers so maybe the Amiga isn't really
cut out for the high-end workstation market?
Mr Gavin Moran
: Mr Gavin Moran (gavinm@vsprsun_14) wrote:
: : Lets assume you could get an 'Amiga' that had a PPC604 and an SVGA chipset
: : in it, running Windows NT. Just what makes this an 'Amiga' and why would
: : ANYONE want to buy it when they could almost certainly get a PC or a Power PC
: : clone a _lot_ cheaper that would do exactly the same job.
: ^^^^^^^^^^^^^
: Wrong answer. If the Amiga was based on the stuff the rest of the
: industry was made of, IT would be a lot cheaper. It's the custom hardware
: and designs that make the Amiga expensive. If it adopted industry
: standards it could EASILY compete at industry price levels.
If the Amiga was based on the stuff the rest of the industry was made of,
it wouldn't be an Amiga and you wouldn't be able to get a nice home machine
like the A1200 at such a good price.
: AmigaOS. There are quite a few of us on the gfx card mailing lists who
: have all but abandoned the Amiga chips (except for the boot menu) and
: we're nothing but happy. AmigaOS is a 32-bit, pre-emptive, multi-tasking
: OS with strong native support for hardware blitters and a solid design.
: REGARDLESS of the hardware - I can't get that elsewhere.
Amiga OS is nice for the home and multimedia user, for the general modern
business computer market it can't compete with NT which is multi-platform,
SMP, fully network ready and can run a range of apps which outnumber the Amiga
by about 30 to 1. As a home user, although I'd like a state of the art 60MIPS
machine with 24 bit gfx in 1280x1024 with ATM network I'm not prepared to pay
all that money and go to all that hassle.
The Amiga is cheap, user friendly, plays good games, well supported with PD
and FUN. A souped up workstation-class machine meets very few of those
criteria.
: I have PowerMacs and Pentium PCs available for my use whenever I want
: them. I don't use them. They just don't let me be productive. The Amiga OS
: *does* because I have been able to _overcome_ the Amiga hardware with
: other devices.
You're crazy to spend all that money on souping up an Amiga just for
home productivity. Very few people would be willing to spend that much money
and spend that much time building a top end system like that.
Nice though AmigaOS is it would DIE if it had to compete as another PPC
OS. It just doesn't have the features or software support that the PPC using
market is looking for.
By all means take the Amiga into the workstation market, but unless theres
a good cheap, fun, low-end HOME machine the Amiga is finished. As a workstation
type machine the Amiga makes no sense at all and in order to make it
competitive as such, you lose EVERYTHING that made the Amiga desirable to the
mass consumer market in the first place.
Mr Gavin Moran
: Depends. Even AGA has 8-bit color and can do 320x200, which is
: all most PC games use nowadays. I'd say that the Amiga *still* beats the
: average PC hands down in the game department, especially considering that
: you can get a 1200 for less than $400. As for the high-end, I stand by
: my assertion that the high end is a dead end for the Amiga.
I totally agree. The thing that made the Amiga great was that it was
*different* to the PC.
A lot of people seem to be saying, lets just clone the PC and run AmigaOS.
What a total waste of time .. the Amiga can never be _cheaper_ than the
broad PC market and if they just settle for trying to do the same things
its never gonna be better either.
If many people here get there way the Amiga will just be another version
of the Power Mac - powerful, but too expensive (and unsuited) for home users
and too different to the mainstream to compete effectively in the business
arena.
The Amiga needs to get back to its roots a PERSONAL computer for the home.
Something thats cheap, friendly, can do amazing games and graphics tricks
and has the potential to be expanded into a really powerful system IF THE
USER SHOULD WANT TO.
We need a redesigned chipset (compatible with AGA) for the low end which
gives the amiga extra gfx oomph for games and multimedia, and a high end
with the fastest 060s and PCI with RTG (which should take care of all the
people whining that amiga chipsets aren't up to SVGA standards of colour
depth and blitter speed).
Making the Amiga some sort of PowerMac clone is like giving it a death
sentence.
Richard Rauch
>that going to be a profitable option? Lets face it, if all you want is a
>fast processor with a high rez gfx card in a box then the PC is always gonna
>be a better bet. I'd be more than happy to have a 50mHz 030 in my Amiga,
But that's just the point, isn't it? Keeping up on hardware is not feasible.
I'm not intimately familiar with the details of MS-WINDOWS or OS/2, but their
memory & CPU requirements (aside from claims from users & supposed users)
make it clear that the Amiga has a "leaness" advantage, at least...relatively
speaking, it's a CPU performance multiplier & RAM multiplier, then.
Now, if the Amiga OS can be made to run on that "cheaper, faster, better"
hardware that you point to, why not? An MS-DOS or PowerMac, or whatever,
compatiblity-box might be run internally to support "well-behaved"
applications from other systems on the same hardware, and the OS could
be sold at low cost to "the masses". Developers who develop for the
Amiga would then no longer be making a decision that would limit them
to a niche piece of hardware. Suddenly, develping "native" Amiga applications
would be more common. (This argument is at least as valid as its alter-ego
that says Amiga users will leave in droves if the hardware were standard,
IMHO).
>never mind an 80mHz 060 so for the majority of people the 060 as top end
>CPU isn't going to be a limitation.
And, "640k ought to be enough for anybody," huh? (^&
> Whether or not the new gfx chipset can match SVGA cards is utterly
>irrelevent- its there for video, multimedia and games and the AGA set
>already does most
>of these far better than even the best PC SVGA sets. As long as its compatible
>with AGA and offers better performance that will be enough.
> Lets assume you could get an 'Amiga' that had a PPC604 and an SVGA chipset
>in it, running Windows NT. Just what makes this an 'Amiga' and why would
Who suggested lining MS' pockets? Run the Amiga OS on it. Being
able to run NT would be a positive thing from marketing's perspective,
and shouldn't be avoided, but if you throw out the OS, you don't have an
Amiga.
>ANYONE want to buy it when they could almost certainly get a PC or a Power PC
>clone a _lot_ cheaper that would do exactly the same job.
What you outline above is a generic PPC system, right? So why is it going
to be a "_lot_" more expensive than any other PPC system?
To the contrary, if the Amiga OS is kept, and if it stays lean, the Amiga
system should be cheaper than a comparable MS-WINDOWS-NT, or the like,
due to reduced RAM required for a usable system, yes?
Gary Bates
: In article <1995Jun19.1...@scala.scala.com>, dave....@scala.com
: (Dave Haynie) wrote:
: [much wisdom from Dave deleted]
: Time for a reality check here folks.
: The day of the custom hardware Amiga is *gone* Back when the common
: denominator was a 16 bit processor on an 8 bit bus, it made sense to spend
: millions on developing special hardware. But these days, 128 bit graphics
: subsytems, 16 bit DSP's, and 64 bit data busses are off the shelf items.
: It makes no sense to try and re-invent the wheel.
: What stands the Amiga apart from the rest of the crowd these days is the
: _operating system_ AmigaDOS does now what the "other guys" are evolving
: toward, with the notable exception of RTG. It multitasks, it's device
: independant, it's small, and it's fast.
: An Amiga, based on a PCI bus and a PowerPC processor, built around the new
: PowerPC spec developed by IBM and Apple, actually stands a chance. Take
: AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
: and add RTG, and you've got a winner.
: The only special hardware I could see on a new Amiga would be a Zorro 3
: slot or 2, alongside the PCI slots, so that older cards could be used.
: That's it.
: Hardware wars are over. Now it's the OS wars.
One of the wisest followups I've seen in months and months. If we are to
have an OS war I'm sure we will come out very favourably. AmigaDOS is a
fantastic OS, make no bones about it. Even Zsolt and Dave would have to
agree that AmigaDOS is the best bit about the Amiga.
--
.-------------------------------------------------------------------------.
! E-Mail: ga...@batesg.demon.co.uk __ !
! Private Node running on an Amiga 1200 __/// B-Spline meshes are my !
! 33Mhz'030/50Mhz'82/10Megs/1.7GbSCSI \XX/ life. !
`-------------------------------------------------------------------------'
Mr Gavin Moran
: Now, if the Amiga OS can be made to run on that "cheaper, faster, better"
: hardware that you point to, why not? An MS-DOS or PowerMac, or whatever,
Because NO-ONE in the wider computer market would want such an OS. If you
have spent on a PPC machine then you WILL have the resources to run NT or
whatever well and the fact the AmigaOS might need less memory is irrelevent,
especially when you compare the amount of business software that NT will let
you run vs the amount of commercial business software the AmigaOS would let
you run.....
: >never mind an 80mHz 060 so for the majority of people the 060 as top end
: >CPU isn't going to be a limitation.
: And, "640k ought to be enough for anybody," huh? (^&
Obviously for some who want outright raw power it won't be, but for the
next few years no typical user should _need_ anything more than an 060. the
Amiga market isn't generally speaking a 'power' market like the PC or PPC
clone markets.
: Who suggested lining MS' pockets? Run the Amiga OS on it. Being
: able to run NT would be a positive thing from marketing's perspective,
: and shouldn't be avoided, but if you throw out the OS, you don't have an
: Amiga.
Why bother making the Amiga at all then? Just run AmigaOS on a powerMac or
similar clone. Apple seem to be going that way as an OS company, but the
Mac hardware was never much to shout about.
: >ANYONE want to buy it when they could almost certainly get a PC or a Power PC
: >clone a _lot_ cheaper that would do exactly the same job.
: What you outline above is a generic PPC system, right? So why is it going
: to be a "_lot_" more expensive than any other PPC system?
OK, at best it costs the same as a Mac or PPC clone but I bet there will
always be the discounters who can do clones just that bit cheaper. Why buy
'Amiga' hardware?
: To the contrary, if the Amiga OS is kept, and if it stays lean, the Amiga
: system should be cheaper than a comparable MS-WINDOWS-NT, or the like,
: due to reduced RAM required for a usable system, yes?
Hows it going to stay lean? Maybe the OS might need less but you will be
able to run FAR LESS software on such a machine. Power users aren't going to
go for that and home/fun users probably won't be willing to spend on a PPC
spec machine.
Mats Pettersson
>In <3sekte$5...@upsn2.u-psud.fr>, vo...@mesiom.obspm.fr (Etienne Vogt) writes:
>>In <1995Jun19.1...@scala.scala.com>, dave....@scala.com (Dave Haynie) wrote:
>>AAA still supports planar screens so you don't have the overhead of a
>>256+ color screen when you only need fewer colors for a particular
>>application.
>
>SVGA also supports 1-4 bitplanes. Sure, anyone who's a big fan of 5,
>6, or 7 bitplane screens would be SOL. Again, that's not something
>anyone needs to spend big resourses for, it's just of too little
>general use.
These SVGA graphics cards for PCI, do they still have 64kB segments
or is that only a problem under MS-DOS/ISA?
By the way... Are you designing any hardware rigth now?
Mats Pettersson (who has an A1000 with fancy autographs inside :-)
Chris Gray
(Dave Haynie) writes:
> >And using true color modes for everything would be a waste of
> >resources.
>
> Why so? If you have dedicate screen memory, you're not wasting
> anything using direct mapped color. Plus, mapped color modes are
> always available.
Programming within the windowing system becomes a lot harder if you have
only modes using lots of colors. Not because of the colours, but because
of the memory required for backup buffers. Opening a full-screen window
suddenly requires a couple of Meg of RAM for a backup buffer. If you
don't want to do that, then you have to be prepared for arbitrary
expose-events throughout your application. (In Intuition-speak, the
difference is between SIMPLE_REFRESH and SMART_REFRESH). The latter
is a *lot* harder for the programmer, and could greatly affect the
number of people who are willing/able to program for the system.
--
Chris Gray c...@ami-cg.GraySage.Edmonton.AB.CA
Michael van Elst
> Yes, but I should hope that you of all people would admit that it
>is the chipset and not the OS that is primarily responsible for the
>Amiga's multimedia potential.
Not at all. Multimedia doesn't play on TV sets only and in all other
respects PC hardware is more advanced or found better solutions.
> Depends. Even AGA has 8-bit color and can do 320x200, which is
>all most PC games use nowadays.
We were talking about multimedia potential, not games.
Games on the other hand do not need high resolution. They need
a large bandwidth to video memory (every SVGA chipset is faster than
AGA) and they need lots of CPU power for real-time effects.
>you can get a 1200 for less than $400.
A PC tailored for playing Amiga-style games could be as cheap.
But PC users probably have more demands and more money. That's
why $1200 computers are popular.
>sitting on top of a huge OS layer. I can't really explain it, it's kind of
>like musical taste. Either you like it or you don't.
These people were left where electronic hobbyists were left 10 years ago.
>(though there ARE some), but there are folks who live and breathe Amiga
>hardware refs. I myself am only mediocre in this regard.
These people are either exaggerating or really do not care about
"hardware refs" when they can get something better..
> I am not defending either the OCS or AGA in particular so much as
>the general idea of a computer that has that sort of a chipset standard.
Restricting yourself to a chipset standard lets you die immediately.
Restricting yourself to a software standard isn't much better.
You have to evolve. Software lets you evolve faster without changing
the hardware all the time and it also helps you to evolve while the
hardware is changing.
> Really? Tell that to the PC game-coders, who are still struggling
>to get their games to run in 640x480 on anything less than a P120 because
>there is no hardware video standard over VGA.
Didn't you just say that games use 320x200 ?
>still defend even the ancient OCS - it may be old, but the fact that so
>many Amigas all came with the absolute same version of it made (and still
>makes) up for a hell of a lot.
It kept the cycle counters busy..
--
Michael van Elst
Internet: mle...@serpens.rhein.de
"A potential Snark may lurk in every tree."
Dave Haynie
>In article <3skjk4$7...@nnrp3.primenet.com>,
>>: itself will go over like a lead balloon. Why do people want to give up
>>: the chipset for the OS? I just don't get this.
>>Well, do you concede that you have to give up the 68K series in favor of
>>RISC?
> Not at all.
You should. How can you argue that no one will want a ported AmigaOS,
and yet claim that folks will continue to buy much slower
hardware. The plain fact of the matter is that the 68K series is out
of steam, at least for the high end. Sure, the 68060 can breathe a bit
of life into the A4000, but it's not going to last beyond that. The
A4000T/060 (what ESCOM could deliver next fall) is going to be
competing with 80x86 machine with 2x-3x the CPU power, at the same
price. Not to mention the RISC systems significantly beyond that (ok,
the main competitor, PowerPC, is currently around twice the Pentium
performance, matching high-end for high-end).
>>If you you pretty much have to give up AGA because how well will
>>AGA like being driven by a foreign CPU?
> Not well at all, especially by a RISC CPU.
Not again. Sometimes y'all give me a headache, talking about hardware
without a clear understanding of it. The AA chipset is very CPU
independent. All the CPU-specific stuff is driven by Gary and some
PALs -- chip select, bus termination, buffer controls, etc. Keep in
mind that the original Agnus was designed to hook directly to the
68000 bus. There's not much more in common between the 68000 and the
68030 buses, or the '030 and '040 for that matter, than there is
between an arbitrary 680x0 and an arbitrary RISC bus. In fact, if
anything, the 68040/68060 bus looks more like a typical RISC CPU bus.
So don't start making pronouncements about Amiga architecture that
just aren't true.
In practical terms, one might ask who'd bother to build a RISC-based
system with AA on the motherboard. Sure, that's senseless, though it's
equally senseless to put any graphics chips on the motherboard these
days, at the high end anyway. Put AA on a PCI module if you're going
to have it, give the user a choice. If there were a RISC port of the
AmigaOS, there would certainly be a CPU module or two to drop into
the A3000 architecture machines. Probably not much more difficult than
the 68040/68060 modules.
>I really don't see very many programs running on Amigas that are
>CPU-bound enough to make a 68060 nonviable.
Sure, and no one needs more that 64K of memory. The point isn't
necessarily what you need in terms of CPU performance. It's both new
applications, especially realtime multimedia, and it's competition. If
the other guy sells a 100MIPS machine, your 25MIPS machine is going to
have a hard time at the same price. Sure, MIPS is a marketing number,
but after all, marketing is much of what sells machines, at least when
a real company is selling them. No one needs an '060 for
wordprocessing. But it's not enough juice for full motion software
MPEG. And 10 '060s aren't necessarily enough power for LightWave
users.
>>The OS is the only thing that will survive the move to RISC. AAA would
>>have been another AGA i.e. too little too late, and Hombre is still a
>>ways off yet which means we're going to have to deal with whatever chips
>>other 3rd parties provide in the meantime.
>
> Which means that we will have to run the graphics through an OS
>layer, which kills speed/smoothness.
News flash -- we do that now, except for games and killer demos. There
are plenty of 3rd party graphics boards providing their own RTG layer,
and every one of them still on the market is considerably faster than
AA at everything. That kind of blows away your complaints.
> Not at all. I think that staying with the 680x0 CPU family for
>at least a couple more years would be the wisest move. The chips are
>cheap,
Not compared to many RISC chips. You can get 68040-class performance
for about $15 in several RISC families. Sure, the 680x0 line is
falling in price, but it's not falling as fast on the low end, and
it's not improving anywhere near as fast on the high end. It can't;
it's just too expensive to design a 680x0 part that can compete. The
only reason Intel's still keeping pace (eg, the gap's there, but it's
not substantially widening) is that they plow millions into CPU
development, which they get back thanks to a captive
market. Motorola's thrust, now, has nothing to do with user-level
computer systems, it's for embedded controller applications.
Jon M. Taylor
Dave Haynie <dave....@scala.com> wrote:
>In <3sjco5$h...@csusac.ecs.csus.edu>, tay...@ecs.ecs.csus.edu (Jon M. Taylor) writes:
>
>> I fail to see any selling point at all for this system. The OS by
>>itself will go over like a lead balloon.
>
>two before you decide the Amiga OS is all that dated.
I have been using both PCs and Amigas since the Amiga 1000 came
out.
>Sure, it needs
>some work, but then again, they all do.
I perhaps have come off a bit to harshly on the Amiga OS. I don't
think it's all that bad, and in it's day it was second to none. However,
the OS has never been what attracted me to the Amiga, and it just isn't an
issue for me. I have transferred a great deal of my fanaticism over to
Linux as well |->.
>There AmigaOS is small enough
>to do the same things on home systems as Windows, only for half the
>memory.
I remember the first day I got my shiny new A1000 with the
then-brand-new V1.1 OS version. I hadn't as yet upgraded from 256K RAM to
512K, and even with that I remember repeatedly opening up multiple tiny
instances of the little random line-drawing workbench demo just to see how
many I could get before something screwed up. I recall getting over fifty
running at once....
>And it can do realtime multimedia, which you simply can't do
>under Windows very effectively. This doesn't make it a general
>solution, but it does make it viable.
Yes, but I should hope that you of all people would admit that it
is the chipset and not the OS that is primarily responsible for the
Amiga's multimedia potential.
>>Why do people want to give up the chipset for the OS?
>
>The Amiga chipset is just plain outdated. While I'm sure ESCOM will
>get some sales out of AA machines if they move fast, these aren't
>going to last very long. Users are expecting more, and expecting it
>fast, especially on the high end.
Depends. Even AGA has 8-bit color and can do 320x200, which is
all most PC games use nowadays. I'd say that the Amiga *still* beats the
average PC hands down in the game department, especially considering that
you can get a 1200 for less than $400. As for the high-end, I stand by
my assertion that the high end is a dead end for the Amiga.
>>I just don't get this.
>
>What I don't get is this virtual "chipset worship" that seems to be a
>common thread in these little discussions.
It does seem like there is a quite distinct subset of us that
strongly prefer assembly language to C, counting vertical refresh cycles
to calculating big-O values, and coding to the bare metal instead of
sitting on top of a huge OS layer. I can't really explain it, it's kind of
like musical taste. Either you like it or you don't.
>Especially by folks who
>don't quite realize just what it is down there in the Amiga
>chips.
I doubt that many people know the chipset as well as you do
(though there ARE some), but there are folks who live and breathe Amiga
hardware refs. I myself am only mediocre in this regard.
>Sure, they were years ahead of their time when first
>introduced. Unfortunately, not 10 years ahead of their
>time.
I am not defending either the OCS or AGA in particular so much as
the general idea of a computer that has that sort of a chipset standard.
>Everything that only the Amiga could do back in 1985 is done
>much, much better by chips that anyone can buy.
Really? Tell that to the PC game-coders, who are still struggling
to get their games to run in 640x480 on anything less than a P120 because
there is no hardware video standard over VGA.
>The primary reason
>people still like Amigas as general purpose computers (well, religious
>issues aside) is that other systems don't take advantage of these
>great standard chips.
They *aren't* standard! The only across-the-board standard
interface common to all clone video cards that I know of is the VESA
interface, which only gives the programmer access to higher-res
screenmodes in a dumb frame buffer manner. No standardized blitting,
clipping, line drawing, display scrolling, sprites, or most of the other
standard Amiga chipset features. Each card implements these functions
differently and "standardizes" them via the Windows GDI. This is why I
still defend even the ancient OCS - it may be old, but the fact that so
many Amigas all came with the absolute same version of it made (and still
makes) up for a hell of a lot.
Most PC games, even the newest ones, *still* aren't as smooth as
7-year-old Amiga scrollers! As long as that is the case (a while yet), as
long as I can get a $50 used A500 at a garage sale that kicks the ass of a
$2000 P90 (except in games that *need* lots of CPU, which is all the PC
seems to have anymore), I think even AGA will suit me just fine. Having a
standard chipset with the coprocessed features like AGA has means a hell
of a lot more to me than being able to do 1280x1024x16M.
>These OSs are slow (for various reasons, most of
>them bad), they're poorly designed, and they're memory hungry.
I won't argue here, other than to say that I am willing to throw a
few more megs of RAM at my OS to get some of the features that AmigaDOS
lacks. I really do consider Amigas and PCs as completely different
beasts.
--
************ Jon Taylor *************************************************
* "For something that has spread with all the forethought of kudzu, the *
* Internet isn't half bad." - Newsweek, 2/27/95 *************************
************************************ tay...@gaia.ecs.csus.edu **********
Jorrit Tyberghein
>4000/030 beat the pants off a DX2/66 Windows box in real world
>performance. That shouldn't have hapened, there's more horsepower in the
>Intel box than 3 4000/030's. Windows is a huge anchor dragging behind any
>machine that runs it. NT is worse. (My boss has a DX4/100 with 16Mb of RAM
>that he's removing NT from because it's too slow) Win95 is better (Yes,
>I've seen it) but it still has a way to go. MacOS is sluggish too. Linux
>is fast, but it has no support.
What do you mean? No support? Linux is supported by the whole Linux
community which is getting rather big at this moment. Oh, but there is no
commercial support! I'd rather have it this way.
>AmigaDOS (updated and improved) provides a valid alternative OS, based on
>speed alone. Can you imagine a PowerPC AmigaDOS Machine?
Yes, it would be VERY nice.
>> >An Amiga, based on a PCI bus and a PowerPC processor, built around the new
>> >PowerPC spec developed by IBM and Apple, actually stands a chance.
>>
>> Great. All this would be is another PPC clone box that happens
>> to run a port of AmigaDOS. If I get a PPC box, I ain't running anything
>> except Linux on it.
>
>Enjoy your directory listings, 'cause that's all you'll be able to do with it.
That's not entirely true. On Linux the source for (almost) everything is
available and free. Someone only has to port the compiler for Linux to PPC
(maybe this has already been done) and you're ready to start compiling. I
compiled most of my software on my Intel Linux box myself.
>> >Take
>> >AmigaDOS, port it to PowerPC, keep a 68000 compatibility mode (a la Apple)
>> >and add RTG, and you've got a winner.
>>
>> I fail to see any selling point at all for this system. The OS by
>> itself will go over like a lead balloon. Why do people want to give up
>> the chipset for the OS? I just don't get this.
When I bought my first Amiga 500 some years ago, I bought it for the OS.
I couldn't care less about the chipset and sound. I use my Workbench in 4
colors and that's all I need. If there is a more modern version of AmigaOS
ported to any processor then that's probably the OS I'm going to buy.
>An OS, with a bunch of apps ready to go, that makes my machine run 5 times
>faster and actually multitasks?
>
>Sounds like a good selling point to me.
This is a good selling point and can be achieved with a port to a faster
series of processors (like the PPC or other RISC processors).
Greetings,
Jorrit.T...@uz.kuleuven.ac.be
AmigaOS and Linux user.
--
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Launchpad is an experimental internet BBS. The views of its users do not
necessarily represent those of UNC-Chapel Hill, OIT, or the SysOps.
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jon M. Taylor
>: layer, which kills speed/smoothness. Again, I see no advantage to using
>: an Amiga if this approach is taken.
>
>Then we'll lose you. So what? We'll gain others who see and respect the
>Amiga for what it is.
This is so classic. The Amiga-using net populace is *highly*
atypical of the Amiga market in general, you know. I always get a chuckle
out of seeing people say things like "I think Escom should make the next
Amigas with 100Mhz PPC 615s, ZIP drives, AAAAAA graphics, SCSI VIII on the
motherboard, zorro 10 bus slots, and a built-in video toaster" and then
wonder why Escom isn't on the net soliciting their opinions....
>The Amiga market will get the people who were on the
>fence - they liked the potential that was never fulfilled because
>Commodore catered to a bunch of people who were trying really hard to
>choose between a Nintendo and an Amiga. People like you who simply see the
>Amiga as a toy have not helped.
I represent the majority of Amiga owners, people who aren't as
likely to be on the net and get dismissed because of it.
>Yes, the Amiga got some short term sales
>from that, but certainly not the long term support it needed to survive.
Funny how C= only started dying after the A500 was
discontinued....
>Had Commodore done it right and avoided people like you in the first
>place, the Amiga might not be dead now.
It isn't dead now. Fortunately, it appears as if Escom is going
to concentrate their efforts on Europe for a little while, where people
seem to mostly see the Amiga in the same light as I do. I feel confident
that Escom is going to flog the 1200 a lot harder than the 4000, for some
strange reason.
>: Not at all. I think that staying with the 680x0 CPU family for
>: at least a couple more years would be the wisest move. The chips are
>: cheap, people know them well, the OS wouldn't need to be ported, and the
>: custom chipset wouldn't break.
>
>Keeping such a chipset would be moronic at best. The Amiga will survive
>as what YOU want it to be: a toy for the playroom.
Well, I suppose that you are entitled to hope that the Amiga goes
in your direction, but I wouldn't hold my breath too long if I were you.
Escom will do what any business (well except for CBM) does, and focus
their marketing efforts on the segment of the market they are selling the
most machines to, and that is the low end. They will do this if for no
other reason than they have *much* less competition and a far larger
established customer base there - it's *easier*! Smart companies will go
for the easy money every time.
>There's too much
>wasted power in the OS for me to be "okay" with that.
I'm not trying to tell you that you have to like what I like about
the Amiga, I am merely stating my opinions and making a prediction or two.
>I'm happy that
>Escom has a much higher opinion of the system and sees potential you are
>incapable of seeing.
It certainly doesn't look to me like they are going too far afield
from where I sit. You can be they have done a lot of market research WRT
the Amiga in Europe and the US, and they could hardly have failed to
notice the breakdown of the installed base. What I am worried about is
that they will crank out a shitload of 1200s to saturate the market, not
ever improve the line, and then drop the Amiga when the 1200 sales get too
low. They might even be able to make decent money off such a strategy,
given how little they paid for the corpse of CBM. I don't think that is
likely, but given the number of cruel turns we Amiga users have had to go
through in the last several years I am not going to relax quite yet.
ChrisBig-KahunaRampson
: I totally agree. The thing that made the Amiga great was that it was
: *different* to the PC.
: A lot of people seem to be saying, lets just clone the PC and run AmigaOS.
: What a total waste of time .. the Amiga can never be _cheaper_ than the
: broad PC market and if they just settle for trying to do the same things
: its never gonna be better either.
: The Amiga needs to get back to its roots a PERSONAL computer for the home.
: Something thats cheap, friendly, can do amazing games and graphics tricks
: and has the potential to be expanded into a really powerful system IF THE
: USER SHOULD WANT TO.
: We need a redesigned chipset (compatible with AGA) for the low end which
: gives the amiga extra gfx oomph for games and multimedia, and a high end
: with the fastest 060s and PCI with RTG (which should take care of all the
: people whining that amiga chipsets aren't up to SVGA standards of colour
: depth and blitter speed).
: Making the Amiga some sort of PowerMac clone is like giving it a death
: sentence.
Hear Hear! Evolution, not Revolution. Evolve the Amiga into a multi-
processor platform! Parallel computing anyone?
: gavan
: --
: email: G.M...@ee.qub.ac.uk | 'There can be only one!'
: or gmo...@nyx.cs.du.edu | - The Highlander
--
Michael van Elst
> I totally agree. The thing that made the Amiga great was that it was
>*different* to the PC.
Other computers were *different* too.
> A lot of people seem to be saying, lets just clone the PC and run AmigaOS.
Noone really says this. But making an Amiga that can use PC hardware
is a different thing.
>What a total waste of time ..
Well, better used than counting cycles..
> The Amiga needs to get back to its roots a PERSONAL computer for the home.
Didn't other say it has to get back to its roots: a GAME CONSOLE ?
For every potential market you can find arguments why the Amiga is
successful there or will fail miserably.
Regards,
Marcel Offermans
>> Can you drag screens nicely with these SVGA chips ?
> Dragging screens around is largely useless one you get to true color
> displays, except perhaps for aesthetics. I explained why it's much less
> useful for solving real problems today than it was 10 years ago in a
> previous article. I'm not saying it isn't cute to be able to drag
> screens, but it's not something to spend a million dollars and an extra
> year to get.
I'm quite used to using a larger virtual workbench. Would something like
this still be possible with SVGA chip technology. I'm not talking about
dragging screens down to reveal other screens, but I do want to be able to
scroll around on a big screen.
>> And using true color modes for everything would be a waste of resources.
> Why so? If you have dedicate screen memory, you're not wasting anything
> using direct mapped color. Plus, mapped color modes are always
> available.
Well, memory costs money. I like the idea better that you can use memory
for graphics and `other things'. But I can see that this possibly is more
expensive than to have dedicated graphics and `normal' memory.
I agree with you that Amiga Technologies should concentrate on using
existing technologies as much as possible.
Greetings, Marcel
nyle landas
: : I agree we need to move RISC but I still cannot believe that we need to
: : totally scrap the hardware that makes an Amiga an Amiga. If we simply go
: : SVGA then why not run OS/2, Windows NT, Windows 95, etc. It's going to be a
: : hard sale to people who don't understand why AmigaDos is better and only
: : listen to the industry. If we go SVGA we'll loose a lot of the stuff that
: : made the Amiga #1 for video work. We need to keep the niche markets while
: : moving onto other markets as well.
: : In my mind this requires -
: : 1.) Port to RISC
: : 2.) RTG
: : 3.) A new custom chipset
: : 4.) PCI
: : With the new custom chipset you still supply a "lowest common denominator"
: : that will be available for the low-end to the high-end. With RTG and PCI
: : you supply inexpensive SCGA cards for people who need that kind of power.
: : As well as many other inexpensive add on cards. If we want it to be an
: : Amiga we can't sacrifice totally on either side.
: My thoughts exactly.. A new custom chipset that enhances AGA to something
<CUT>
: high end. Add RTG to the OS and a PCI slot and the high end is catered for
: in the host of SVGA chipsets coming out on PCI card. A port to RISC will
: be needed to keep the Amiga in the frame as a 'workstation' computer, but is
: that going to be a profitable option? Lets face it, if all you want is a
If we plan on moving into new markets and ever having more power than the
060 then yes, we need this move. IF they go PPC which I herby declare is
what I think Amiga Technologies will do. 8 ) Yeah, like I'm an expert or
soemthing. They can make their own reference platform and sell PowerMac
clones and Amigas based on the basic design. This will cut costs as others
have mentioned.
: fast processor with a high rez gfx card in a box then the PC is always gonna
: be a better bet. I'd be more than happy to have a 50mHz 030 in my Amiga,
: never mind an 80mHz 060 so for the majority of people the 060 as top end
: CPU isn't going to be a limitation.
For low end users the Motorola line is viable but for people looking to do
high-end multimedia and video work the more CPU power the better. This
market is where the Amiga has shinned we can't ignore it and the port to
RISC is critical if we want to keep it.
: Whether or not the new gfx chipset can match SVGA cards is utterly irrelevent- its there for video, multimedia and games and the AGA set already does most
: of these far better than even the best PC SVGA sets. As long as its compatible
: with AGA and offers better performance that will be enough.
: Lets assume you could get an 'Amiga' that had a PPC604 and an SVGA chipset
: in it, running Windows NT. Just what makes this an 'Amiga' and why would
: ANYONE want to buy it when they could almost certainly get a PC or a Power PC
: clone a _lot_ cheaper that would do exactly the same job.
Well if Amiga Technoligies/ESCOM can make one reference platform with the
PPC and sell
1.) Mac clones
2.) Amigas
3.) Windows NT servers
Then they'll lower the overall cost of production for each which is good all
around. Hopefully the Amiga will sell best. :)
-Nyle
--
/*******************************************************************\
|Nyle F. Landas |
|E-MAIL - ny...@servtech.com |
|URL - http://www.servtech.com/public/ami4000/ |
|FTP - ftp://ftp.servtech.com/pub/users/ami4000/ |
| |
| Build a better mouse trap and Bill Gates will steal all the cheese.|
\********************************************************************/
nyle landas
: : : Then we'll lose you. So what? We'll gain others who see and respect the
: : : Amiga for what it is. The Amiga market will get the people who were on the
: : <STUFF CUT>
: : : Had Commodore done it right and avoided people like you in the first
: : : place, the Amiga might not be dead now.
: : The Amiga can not afford to avoid any people who are potential customers.
: : Amiga Technologies will need every Niche it can find even JTaylor.
: Not true. It is possible to lose 2-to-1 (or worse) if you cater to the
: wrong group. You just CAN'T please everyone - so please as many people as
: you can who will be there when you need them.
That only holds in this case if keeping a custom chipset will loose
customers. I and others feel that not keeping it will loose customers
especially in the low end.
: : : Keeping such a chipset would be moronic at best. The Amiga will survive
: : I agree we need to move RISC but I still cannot believe that we need to
: : totally scrap the hardware that makes an Amiga an Amiga. If we simply go
: : SVGA then why not run OS/2, Windows NT, Windows 95, etc. It's going to be a
: Because it's not AmigaOS. I can run OS/2 and Win NT *now*, but I run
: AmigaOS. In fact, I bought an Amiga just to do it. If people are that
: uninterested in the Amiga that they would jump on WinXX or OS/2, it'd be
: hard to convince them to buy a custom platform that's guaranteed obsolete
: and non-upgradable out the door.
I love AmigaDos too, but it's not the only reason I bought or own an Amiga
the custom hardware played a much larger part in my purchase and others
purchase to. As I said in the remainder of my post the Amiga comunity is
currently split among AmigaDos survivalist and AmigaDos/Custom Chipset combo
survivalistst. It seems to be a failry even ratio and we can't afford to
loose 1/2 of the potential customers. A custom chipset does not gurantee
obsolete and non-upgradeability if we add RTG/PCI as suggested. This would
allow a common denominator for all machines and power-users who need SVGA
can add a PCI card and retarget everything. 8 )
: : listen to the industry. If we go SVGA we'll loose a lot of the stuff that
: : made the Amiga #1 for video work. We need to keep the niche markets while
: : moving onto other markets as well.
: No, if we go SVGA *only* we'll lose out. If we go RTG - we can have the
: best of all worlds.
Well if you dump the custom chipset you'll pretty much be reserved to SVGA
only. Since the custom chips are what are used for most video work. Yes,
if we go RTG we can have the best of both worlds but only if we upgrade the
custom chipset. :)
: I agree, in a different order.
: 1) RTG (can be out the door faster than any of the others)
: 2) RISC
: 3) PCI
You seem to suggest not dumping the custom chips and then you leave them out
of future enchancements how about -
1.) RTG 2.) RISC 3.) PCI [Done at same time virtually]
4.) Enhanced custom chipset
nyle landas
: AmigaOS. There are quite a few of us on the gfx card mailing lists who
: have all but abandoned the Amiga chips (except for the boot menu) and
: we're nothing but happy. AmigaOS is a 32-bit, pre-emptive, multi-tasking
: OS with strong native support for hardware blitters and a solid design.
: REGARDLESS of the hardware - I can't get that elsewhere.
How many of those people are usinf those same Video cards to do video work?
Most of the people using video cards are doing so for DTP, CAD, Internet,
etc. not video work. We can't ignore this market.
: I have PowerMacs and Pentium PCs available for my use whenever I want
: them. I don't use them. They just don't let me be productive. The Amiga OS
: *does* because I have been able to _overcome_ the Amiga hardware with
: other devices.
As I said before - 8 ) Just because you only want AmigaDos doesn't mean
there aren't many others out there who don't need the custom chips...
Especially the low-end users for who the custom chips are vital. We need
that lowest common-denominator and yes, because Commodore let development
fall so far behind and didn't release advancements even when available, many
"hacks" came about to make the Amiga more powerful. This doesn't
automatically mean the custom chipset could not have been kept competitive
or still be made competetive now.
SVGA
and Multi-I/O cards were adopted by so many Amiga users simply because C=
didn't update the custom chips with faster blitter and faster ports... These
can be added to the custom chips now. 8 ) And could have been done in 1992.
If the Amiga goes SVGA only I may not follow and many others feel the same
way this is something ESCOM will have to be very careful about. I don't
want another non-elegant horsepower above all designed computer if I did I'd
own a PCC or a Mac not the Amiga with it's great Hardware/OS combo.
Richard Rauch
>: Now, if the Amiga OS can be made to run on that "cheaper, faster, better"
>: hardware that you point to, why not? An MS-DOS or PowerMac, or whatever,
> Because NO-ONE in the wider computer market would want such an OS. If you
>have spent on a PPC machine then you WILL have the resources to run NT or
>whatever well and the fact the AmigaOS might need less memory is irrelevent,
And it won't matter to you, say, that for the same performance, you can shave
several megs from the system (how many megs does NT require just to boot?
how much does a meg of RAM go for, for such a hign-end system?)? Or,
alternatively, for the same cost, you can have your system handle larger
working sets without swapping to disk (i.e., for the same money, MUCH
better performance).
>especially when you compare the amount of business software that NT will let
>you run vs the amount of commercial business software the AmigaOS would let
>you run.....
There is no limit to the amount of commercial business software that the
Amiga OS will let you run. It's not a limitation of the Amiga OS, despite
how you craft your sentence. It's that developing for a completely niche
system is unappealing for developers. With a port of the OS, they _could_
develop for the Amiga OS with a little less trouble, and with a lot better
chance of gaining back their investment (because "everyone" could then
run the Amiga OS, and so "everyone" is a potential customer of their system).
Aside from people who already have the Amiga OS on their system for personal
preference/use, suppose you look at it this way:
Buy MS-WINDOWS-NT for <however much 1>, plus foo-application for <however
much 2>.
OR:
Buy the Amiga OS port for <however much 3>, plus foo-application for
<however much 4>, AND get significant performance boosts (or reduced
system requirements, or both).
Assuming <however much 2> and <however much 4> are about the same
(the same product, after all), and that the Amiga OS port is no more
expensive than MS-WINDOWS-NT, and all features are otherwise the
same, why not?
>: Who suggested lining MS' pockets? Run the Amiga OS on it. Being
>: able to run NT would be a positive thing from marketing's perspective,
>: and shouldn't be avoided, but if you throw out the OS, you don't have an
>: Amiga.
> Why bother making the Amiga at all then? Just run AmigaOS on a powerMac or
I think I answered this elsewhere. "Making the Amiga" means making the
OS, as far as I'm concerned. Other than that, yeah, that might be where
I'd like to see things head, eventually. As a user.
>similar clone. Apple seem to be going that way as an OS company, but the
>Mac hardware was never much to shout about.
The little I've used Mac's, their OS isn't very good (or nice), either.
But that's just an opinion. My guess as to why they are so popular is:
They offered a GUI out-of-the-box before any other home/personal computer
did. On top of that, it had Apple's name-recognition to build on. And,
too, I'm of the cynical mind that if you take a mediocre product & price
it high enough, you'll get people who'll think it _must_ be good because
of the high price.
(Not that I have enough experience with Mac's to say if they are really
as mediocre as my low exposure led me to believe. (^&)
>: What you outline above is a generic PPC system, right? So why is it going
>: to be a "_lot_" more expensive than any other PPC system?
> OK, at best it costs the same as a Mac or PPC clone but I bet there will
>always be the discounters who can do clones just that bit cheaper. Why buy
>'Amiga' hardware?
Very possibly, I wouldn't. If I could get the OS to run on <any CPU>, I'd
probably be seriously looking for a new CPU to run my Amiga on. (^&
(I'd avoid an Intel system, and I've only marginally watched the development
of alternative systems, so I dunno what I'd buy, right now. If the Amiga
OS had been in the process of being ported to one or more other CPU's,
I would have kept a serious watch on other hardware.)
Of course, since porting the OS sounds like a major project, I probably
won't have the impetus to go looking for non-Amiga hardware. But,
cheaper, faster, better graphics displays would be nice.
>: To the contrary, if the Amiga OS is kept, and if it stays lean, the Amiga
>: system should be cheaper than a comparable MS-WINDOWS-NT, or the like,
>: due to reduced RAM required for a usable system, yes?
> Hows it going to stay lean? Maybe the OS might need less but you will be
I'm assuming it can manage to keep its form. I am not employed by anyone
to maintain the Amiga OS. If you are, perhaps you can you can spell out
why it can't remain lean?
>able to run FAR LESS software on such a machine. Power users aren't going to
>go for that and home/fun users probably won't be willing to spend on a PPC
Home/fun users should probably get a game console. Stripping that silly
keyboard, mouse, floppy/HD, and other pointless features (including the
OS) would greatly reduce the cost of their toy. (And the CD32 won't
be back in production until '96, I thought I read, so you won't be
likely to see any Amiga systems like that, except used ones. Even then,
they include the OS. (^&)
As for others: They may want the standard hardware to run any OS's (hope-
fully including the Amiga OS) that run on the hardware. If they buy
it under an Amiga Tech. label, a Commodore/Escom label, or an IBM (or
Apple, ...) label...it won't make much difference. (^&
>gavan
>--
>email: G.M...@ee.qub.ac.uk | 'There can be only one!'
> or gmo...@nyx.cs.du.edu | - The Highlander
--
nyle landas
: In article <1995Jun26....@scala.scala.com> dave....@scala.com
: (Dave Haynie) writes:
: > >And using true color modes for everything would be a waste of
: > >resources.
: >
: > Why so? If you have dedicate screen memory, you're not wasting
: > anything using direct mapped color. Plus, mapped color modes are
: > always available.
: Programming within the windowing system becomes a lot harder if you have
: only modes using lots of colors. Not because of the colours, but because
: of the memory required for backup buffers. Opening a full-screen window
: suddenly requires a couple of Meg of RAM for a backup buffer. If you
Anyone else catch the new standard being devloped on PC compatibles to merge
video RAM and system RAM like the Amiga and Mac??? Mmmm, I guess they
finally decided using some of that 2-4MB for programs when it's un-used
would be a good idea. If we now go to dedicated video ram we'll be
considered out-dated. :)
Keep the chipram but extend the barier to 16MB...
Christian Brandt
In article <3sps9v$l...@news-s01.ny.us.ibm.net>, mat...@ibm.net (Mats Pettersson) writes:
|> >6, or 7 bitplane screens would be SOL. Again, that's not something
|> >anyone needs to spend big resourses for, it's just of too little
|> >general use.
|>
|> These SVGA graphics cards for PCI, do they still have 64kB segments
|> or is that only a problem under MS-DOS/ISA?
Even many, maybe all, ISA, EISA and VLB VGA-Solutions can be memory-mapped.
The 64k-Segments are only needed if you operate in Real-Mode, which
is the native 8086-Mode. But Amigas generally do not run in 8086-Mode :-)...
|> By the way... Are you designing any hardware rigth now?
|> Mats Pettersson (who has an A1000 with fancy autographs inside :-)
--
Christian Brandt (who has the same autographs inside his A1000 :-)
Mr Gavin Moran
: >: Now, if the Amiga OS can be made to run on that "cheaper, faster, better"
: >: hardware that you point to, why not? An MS-DOS or PowerMac, or whatever,
: > Because NO-ONE in the wider computer market would want such an OS. If you
: >have spent on a PPC machine then you WILL have the resources to run NT or
: >whatever well and the fact the AmigaOS might need less memory is irrelevent,
: And it won't matter to you, say, that for the same performance, you can shave
: several megs from the system (how many megs does NT require just to boot?
: how much does a meg of RAM go for, for such a hign-end system?)? Or,
: alternatively, for the same cost, you can have your system handle larger
: working sets without swapping to disk (i.e., for the same money, MUCH
: better performance).
Actually, the memory requirements for AmigaOS on a PPC machine would be
higher than at present, especially if 68k emulation code was included. And
by the time you add in TCP/IP, proper network printing/filesharing, some sort
of display postscript etc you'd need yet more memory. Performance is
irrelevent if you can't get the software you want to run on the OS.
NT performs nicely on a 60mHz pentium, on an PPC chip at similar clockspeed
it should also be quite usable.
: >especially when you compare the amount of business software that NT will let
: >you run vs the amount of commercial business software the AmigaOS would let
: >you run.....
: There is no limit to the amount of commercial business software that the
: Amiga OS will let you run. It's not a limitation of the Amiga OS, despite
: how you craft your sentence. It's that developing for a completely niche
: system is unappealing for developers.
Exactly, and AmigaOS would be a niche system on PPC machines. Lets not
forget it doesn't even have true RTG or TCP/IP right now, by the time its
ported to PPC code with all the necessary embelishments, where will NT and
MacOS be in terms of installed base and specs?
: With a port of the OS, they _could_
: develop for the Amiga OS with a little less trouble, and with a lot better
: chance of gaining back their investment (because "everyone" could then
: run the Amiga OS, and so "everyone" is a potential customer of their system).
: Aside from people who already have the Amiga OS on their system for personal
: preference/use, suppose you look at it this way:
No, they;d say 'we can develop for NT or AmigaOS - which gives us the most
sales potential?' Doesn't take too long to figure what the answer would be.
: Buy MS-WINDOWS-NT for <however much 1>, plus foo-application for <however
: much 2>.
: OR:
: Buy the Amiga OS port for <however much 3>, plus foo-application for
: <however much 4>, AND get significant performance boosts (or reduced
: system requirements, or both).
Assuming the applications you want are available on AmigaOS and are current
with their NT or MacOS revisions.
: Assuming <however much 2> and <however much 4> are about the same
: (the same product, after all), and that the Amiga OS port is no more
: expensive than MS-WINDOWS-NT, and all features are otherwise the
: same, why not?
For the same reason OS/2 hasn't really set the world on fire - theres just
too much of an installed base of competitive OS products. And even though
OS/2 is clearly superior to Win 3.x, its not apparent that AmigaOS could
beat NT in anything but raw performance. Features wise NT is already a lot
better and would have improved even more by the time a PPC port of AmigaOS
was finished.
: > Why bother making the Amiga at all then? Just run AmigaOS on a powerMac or
: I think I answered this elsewhere. "Making the Amiga" means making the
: OS, as far as I'm concerned. Other than that, yeah, that might be where
: I'd like to see things head, eventually. As a user.
Thats your view, mine is that I bought the Amiga coz of the cool hardware
which could do things totally different to what PCs could do.
: >: system should be cheaper than a comparable MS-WINDOWS-NT, or the like,
: >: due to reduced RAM required for a usable system, yes?
: > Hows it going to stay lean? Maybe the OS might need less but you will be
: I'm assuming it can manage to keep its form. I am not employed by anyone
: to maintain the Amiga OS. If you are, perhaps you can you can spell out
: why it can't remain lean?
TCP/IP would need to be integral, Network File Sharing and printing would
need to be integral. Hardware memory protection would need to be integral,
the core would have to be portable, it would need SMP capability, RTG would
need to be available, you couldn't assume you had the nice tight custom
chipset to ease the strain on the CPU for I/O and gfx functions.....
: >able to run FAR LESS software on such a machine. Power users aren't going to
: >go for that and home/fun users probably won't be willing to spend on a PPC
: Home/fun users should probably get a game console. Stripping that silly
: keyboard, mouse, floppy/HD, and other pointless features (including the
: OS) would greatly reduce the cost of their toy. (And the CD32 won't
: be back in production until '96, I thought I read, so you won't be
: likely to see any Amiga systems like that, except used ones. Even then,
: they include the OS. (^&)
Home/Fun is where the Amiga is AT. Thats why it sold so well in Europe.
Thats why it still exists today. I certainly wouldn't have a computer at
home if I didn't enjoy using it - I have access to powerful PCs, Macs,
SUNS etc at work.
Dave Haynie
>Dave Haynie (dave....@scala.com) wrote:
>: SVGA also supports 1-4 bitplanes. Sure, anyone who's a big fan of 5,
>: 6, or 7 bitplane screens would be SOL. Again, that's not something
>: anyone needs to spend big resourses for, it's just of too little
>: general use.
>The Picasso II does 5, 6 and 7 bitplanes in planar or chunky format.
>No loss there. (?) Am I missing something?
IBM's original VGA specification (which all SVGA chipsets support)
requires planar modes to four bitplanes. While I suppose there's no
stopping any chip company from supporting more planes, certainly most
of them don't. While any RTG solution is likely to support planar
models at 5, 6, and 7 bits/pixel, you probably don't have that in
hardware. Or maybe they're using a weird chip...
Dave Haynie
>> The "faster ET4000W32p" is rather a very lowcost-solution and
>> veryvery slow - 20-30MB/s.
>> The better S3 864 64Bit-Systems do around 40MB/s on PCI and
>> sometimes up to 60MB/s.
>I was talking about how many megabytes per second one can copy to the
>videoram, not internal speeds or theoretical speeds.
So are we.
The S3 '864 runs a 264MB/s private bus. Some of this is devoted to
display, some to blitter, and some to the PCI channel. The '964 does
somewhere around 764MB/s on its bus, but that's the big advantage of
VRAM. Think of it as 264MB/s for PCI and blitter, since VRAM offloads
all display activity from the normal parallel, random access memory
interface. These are real speeds.
In 6 months or so, the first Rambus-based video cards will
come out with similar performance, at DRAM prices (a single RDRAM
channel can provide 500MB/s transfers sustained for blocks of 4K or
more, depending on the part, the number of parts on the bus, and how
well they designed the controller to use the Rambus).
>Cirrus Logic cards are usually about 10MB/s, so is my friends
>S3 805 (VLB),
The S3 '805, and Cirrus logic parts of similar vintage, are ancient,
by PC standards. Those are first generation DRAM-based 32-bit parts,
designed to much slower memory and chip speeds. Like the Amiga chips,
they spend a substantial portion of their bandwidth running display
refresh.
>ET4000/W32p-cards are faster.
Those are second generation 32-bit parts, which can deliver much
better performance than most of the 32-bit parts out there. Mainly
because the SVGA market is so competitive, you rarely get a second
pass at the same technology level. Three years ago, it was ISA bus
parts with blitters, DRAM and VRAM models. Two years ago, 32-bit VLB
parts, DRAM and VRAM models. Last year, 64-bit VLB/PCI parts, DRAM and
VRAM models. This year there's already one 128-bit PCI-based card. The
next generation cards are all going to various "exotic" DRAMs to
deliver VRAM-or-better performance for DRAM-card prices.
>I guess the speed depends quite much on the motherboard
>implementation too.
What you're running into is simply OLD TECHNOLOGY, nothing more. You
wouldn't expect to get 50MB/s though to Amiga Chip RAM, either. It
wasn't all that long ago that typical VGA cards were slow, by
comparison, to the Amiga's frame buffer. But they've been growing at
an amazing pace, with a new generation every year or two. You're
simply way behind the times. Remember, as I mentioned a few articles
ago, I looked into this stuff, in detail, a year ago. I'll wager
that's a bit more scientifically valid than checking out what my
buddies have in their systems.
Maxwell Daymon
And look where the Amiga is now.
: It isn't dead now. Fortunately, it appears as if Escom is going
All but dead.
: >Keeping such a chipset would be moronic at best. The Amiga will survive
: >as what YOU want it to be: a toy for the playroom.
: Well, I suppose that you are entitled to hope that the Amiga goes
: in your direction, but I wouldn't hold my breath too long if I were you.
I won't hold my breath at all, but any time I can use my Amiga over my PC
- I do. The Mac isn't so bad for some things, but I'd go crazy ONLY with
that OS after a while.
: Escom will do what any business (well except for CBM) does, and focus
: their marketing efforts on the segment of the market they are selling the
: most machines to, and that is the low end. They will do this if for no
They may or may not. Wise companies sell to who they will get profit and
longevity from.
: other reason than they have *much* less competition and a far larger
: established customer base there - it's *easier*! Smart companies will go
: for the easy money every time.
No, dumb companies will. Smart companies go for the solid money.
: It certainly doesn't look to me like they are going too far afield
: from where I sit. You can be they have done a lot of market research WRT
: the Amiga in Europe and the US, and they could hardly have failed to
: notice the breakdown of the installed base. What I am worried about is
The installed base wasn't solid. Amiga one day, Nintendo the next. How
many A500's are simply sitting in closets owned by people who won't buy
another one?
: that they will crank out a shitload of 1200s to saturate the market, not
: ever improve the line, and then drop the Amiga when the 1200 sales get too
: low. They might even be able to make decent money off such a strategy,
: given how little they paid for the corpse of CBM. I don't think that is
: likely, but given the number of cruel turns we Amiga users have had to go
: through in the last several years I am not going to relax quite yet.
I only hope that if the Amiga dies as a professional machine, there's an
alternative to the crap Microsoft calls software. Frankly, using
Microsoft programs is like a punishment, and if the Mac hasn't improved,
I think I'd be happier doing everything with paper and pencil until said
companies aquire intelligence and put out something usable.
Until then, I'll support the underdog because the underdog is doing it
better.
--
+- Maxwell Daymon -+- mda...@rmii.com -+
Maxwell Daymon
: is the chipset and not the OS that is primarily responsible for the
: Amiga's multimedia potential.
NOT according to Scala - the premeir multimedia company. They biggest
stumbling block for MM on the PC was the OS. That's why they created MMOS
and are now doing MM on the PC.
NOT according to one particular Amiga engineer who is RESPONSIBLE for the
integration of such chips.
: Depends. Even AGA has 8-bit color and can do 320x200, which is
Slowly. Very, very slowly.
: you can get a 1200 for less than $400. As for the high-end, I stand by
: my assertion that the high end is a dead end for the Amiga.
I stand by my assertion that it is not.
: >What I don't get is this virtual "chipset worship" that seems to be a
: >common thread in these little discussions.
: It does seem like there is a quite distinct subset of us that
: strongly prefer assembly language to C, counting vertical refresh cycles
: to calculating big-O values, and coding to the bare metal instead of
: sitting on top of a huge OS layer. I can't really explain it, it's kind of
: like musical taste. Either you like it or you don't.
But when it comes to working or not - that's a different story.
Processors today don't lend themselves to the architecture of yesterday.
: >Everything that only the Amiga could do back in 1985 is done
: >much, much better by chips that anyone can buy.
: Really? Tell that to the PC game-coders, who are still struggling
: to get their games to run in 640x480 on anything less than a P120 because
: there is no hardware video standard over VGA.
That's not a chipset problem. That's a Microsoft problem. That's what
happens when a monopoly control the industry.
: >The primary reason
: >people still like Amigas as general purpose computers (well, religious
: >issues aside) is that other systems don't take advantage of these
: >great standard chips.
: They *aren't* standard! The only across-the-board standard
Standardized WITHIN the Amiga.
: screenmodes in a dumb frame buffer manner. No standardized blitting,
: clipping, line drawing, display scrolling, sprites, or most of the other
: standard Amiga chipset features. Each card implements these functions
And this is from poor design and implementation. That's IT.
: I won't argue here, other than to say that I am willing to throw a
: few more megs of RAM at my OS to get some of the features that AmigaDOS
: lacks. I really do consider Amigas and PCs as completely different
: beasts.
So do I, that's why I want a high end Amiga. There's too many companies
concentrating on low end games. (Sega, Sony, Nintendo) for Ami to survive
there. Clones are priced too low for the Amiga to fit inbetween.
Dave Haynie
>In article <3sleii$b...@natasha.rmii.com>,
>Maxwell Daymon <mda...@rainbow.rmii.com> wrote:
>>Jon M. Taylor (tay...@ecs.ecs.csus.edu) wrote:
> Funny how C= only started dying after the A500 was
>discontinued....
Well, it was a bit before that. The cancellation of the A500 was a
clear symptom of the cancer at work within Commodore. After all, it
was still selling well. This was the first time in Commodore's history
that a successful product was cancelled, in favor another product (the
A600) that was largely disliked within the company (outside of
management, anyhow). This was nothing but politics, personal greed for
money and power, and bad management. There was quite a bit of that,
and that's what killed Commodore.
>What I am worried about is that they will crank out a shitload of
>1200s to saturate the market, not ever improve the line, and then
>drop the Amiga when the 1200 sales get too low. They might even be
>able to make decent money off such a strategy, given how little they
>paid for the corpse of CBM.
They certainly could do that. As they've alread announced, they'll
make A4000Ts first, and for both Europe and NA. This, I believe, is
primarily because they can -- you can use many, many standard parts to
make an A4000T. For the A1200, you need custom plastic, that's
especially annoying if they haven't salvaged the die for these from
the Phillipines (or if they've worn too much and need replacing).
The volume market is clearly with the A1200, and it makes sense. The
A4000T is a niche machine, it can't compete head to head on
performance or graphics with any up-to-date personal computer. The
A1200, on the other hand, is cheaper than the standard PC or Mac by
far, works on the TV, and doesn't need 8MB of RAM for a base setup
(which Windows95 systems will, no matter what Microsoft says right
now).
>I don't think that is likely, but given the number of cruel turns we
>Amiga users have had to go through in the last several years I am not
>going to relax quite yet.
Somehow, Commodore's management group in the early 90s emerged as a
world-class Bozo magnet. Like you couldn't get in there unless you had
proven record of failure or, at least, inertness. In the fall of '92
there was not one manager I can think of (except Jeff Porter, who had
been running the show, but was cast aside by the new guys and had
little power to correct the mistakes of the others), you'd want to let
near your computer, much less determine it's ultimate fate. You would
have thought the fun was over by the time the liquidator took over,
but as we all know, that was a mess in itself.
Hopefully ESCOM will do better. But it's easy to be jaded at this
point, even if you weren't in the thick of things.
Michael van Elst
> Most of whats 'happening' in multimedia is designed to play on domestic TV
>sets or needs BIG screens - that means a lot of expense if you are using
>dedicated computer monitors.
Probably depends on the definition of "Multimedia".
> SVGA chipsets may have a larger bandwidth but the vast majority of games
>DONT USE THEM.
No ? Why do you think so ?
>They simply use the 256 colour MCGA mode, thus theres little
>to no speed advantage over AGA in that sense.
Of course there is a speed advantage..
> uh uh, NO current PC could play Amiga-style games as well as the Amiga.
No ? I beg to differ.
ChrisBig-KahunaRampson
people actually talked about. If you want to argue the future of Amigas,
read the thread I started. I welcome all opposing viewpoints from
Amigans, I do not welcome Mac or PC users straying from the subject!
Dave Haynie
>These SVGA graphics cards for PCI, do they still have 64kB segments
>or is that only a problem under MS-DOS/ISA?
While I believe most have some option for segmented operation, they
all run linearly. Real OSs on the PC have been using them that way for
quite some time. Even ISA bus supported 24-bits worth of addressing,
it's MS-DOS that was the real problem
>By the way... Are you designing any hardware rigth now?
Yes, but nothing interesting. I'm also working on a new DiskSalv
release, some rock music, and shortly, I'll be up at the AmiJAM show
in Calgary...
Michael van Elst
>I'm quite used to using a larger virtual workbench. Would something like
>this still be possible with SVGA chip technology. I'm not talking about
>dragging screens down to reveal other screens, but I do want to be able to
>scroll around on a big screen.
That's possible with pretty _any_ SVGA chip. Of course your display
is limited by the amount of video memory on the card. A 4MB card
could display a 2048x2048 screen in 256 colors.
Neil Clark
Michael van Elst <mle...@serpens.rhein.de> wrote:
>
>Close to every system uses SIMPLE_REFRESH (i.e.: no backup buffers).
Doesn't mean it's the best way to attack the problem though. If I
remember correctly, Foley et al in "Computer Graphics: Principles and
Practice" lists the advantages and disadvantages of both. I personally
think the Amiga has the best system offering a choice of both simple and
smart refresh, where each can be used in the appropriate situation.
--
Neil Clark
Telepresence researcher
University of Strathclyde
"I can create a fire anywhere with a few hours and two old cliches"
Dave Haynie
>In article <1995Jun26....@scala.scala.com>,
>Dave Haynie <dave....@scala.com> wrote:
>>In <3sjco5$h...@csusac.ecs.csus.edu>, tay...@ecs.ecs.csus.edu (Jon M. Taylor) writes:
> I perhaps have come off a bit to harshly on the Amiga OS. I don't
>think it's all that bad, and in it's day it was second to none. However,
>the OS has never been what attracted me to the Amiga, and it just isn't an
>issue for me. I have transferred a great deal of my fanaticism over to
>Linux as well |->.
Linux is UNIX, without the trademarks. If you like UNIX, great, lots
of people do. I used UNIX for years before I started with AmigaOS, and
I like AmigaOS better. It just sits better with my philosophy about
how things should be done in a computer OS. To me, UNIX does lots of
things wrong. Sure, the list isn't as long as the list for Windows,
it's probably about as long as the OS/2 list... A UNIX fanatic can
certainly tear into AmigaOS for all he's worth, too.
>>And it can do realtime multimedia, which you simply can't do
>>under Windows very effectively. This doesn't make it a general
>>solution, but it does make it viable.
> Yes, but I should hope that you of all people would admit that it
>is the chipset and not the OS that is primarily responsible for the
>Amiga's multimedia potential.
You're dead wrong on that one. You haven't been listening. At best,
Amiga's CPU, chipset, etc. are considerably slower than anything
that's been popular in the PC market for several years. Yet, I can do
multimedia things on the Amiga you can't do on the PC. Period. That is
due only to one thing -- the Amiga's OS. Sure, the hardware is a good
design, but at least much of what you get on a PC these days is very
good, hardware-wise, too.
The Amiga's chipset helped make real multimedia possible when it
couldn't be done on "standard" hardware (which really wasn't all that
standard back then, "standard" meant "PClone", today it means PCI,
which includes PClone, Apple/IBM/Motorola PowerPC, Sun SPARC, DEC
Alpha, etc. The hardware and software were designed together, and they
play very well together, both supporting the nothing that everything
should be fast and nothing should have to wait.
Well, the chipset isn't fast anymore. I suspect if the Amiga chipset
goes back into production, it'll be the slowest graphics chipset for a
general purpose computer still in production.
>>What I don't get is this virtual "chipset worship" that seems to be a
>>common thread in these little discussions.
> It does seem like there is a quite distinct subset of us that
>strongly prefer assembly language to C, counting vertical refresh cycles
>to calculating big-O values, and coding to the bare metal instead of
>sitting on top of a huge OS layer. I can't really explain it, it's kind of
>like musical taste. Either you like it or you don't.
You do all that on Linux? That's called hacking, it's fun, sure. No
problems. But it's not going to keep ESCOM in the Amiga business.
> I doubt that many people know the chipset as well as you do
>(though there ARE some),
Of course there are. I didn't design the chips, I don't program them
directly. There are possibly two other guys, George Robbins and Greg
Berlin, who know them (well, ECS and AA, not AAA) from the system
point of view as well as I do. Though neither of these guys knows the
software side as well, though George contributed more to the actual
design of the chips. You need to understand both the hardware and the
software to get The Big Picture. That was a longstanding C= tradition,
and also one at Amiga, which is why the Amiga worked so well. At most
companies, there's a big wall between the two, if they do both at all.
>>Sure, they were years ahead of their time when first
>>introduced. Unfortunately , not 10 years ahead of their
>>time.
> I am not defending either the OCS or AGA in particular so much as
>the general idea of a computer that has that sort of a chipset standard.
Clearly, the popular system defines the standard. AA is well
supported, yet it's only available on a small percentage of
Amigas. Anything else that goes into the low end will have the same
level of support. On the high end, it's not necessary. If a particular
game won't run on my RTG board, I won't run that game.
>>Everything that only the Amiga could do back in 1985 is done
>>much, much better by chips that anyone can buy.
>
> Really? Tell that to the PC game-coders, who are still struggling
>to get their games to run in 640x480 on anything less than a P120 because
>there is no hardware video standard over VGA.
And yet, the standard for new games is the PC. Under MS-DOS, of
course, they don't want Windows in the way. And I've seen quite a few
games that blow away anything I've seen on the Amiga (not that I care,
I don't play games). Seen Myst or DOOM 2 for the Amiga yet? Game
coders are blowing the old systems away, very simply. If you don't
have a '486DX2-66, local bus SVGA, and 16-bit sound with General MIDI,
you're lucky if a modern game plays well at all. If you're not keeping
your hardware to date, there's little reason to assume you buy much
software.
>>The primary reason
>>people still like Amigas as general purpose computers (well, religious
>>issues aside) is that other systems don't take advantage of these
>>great standard chips.
> They *aren't* standard!
They're standard in hardware. That's the point. And they can be set up
for standard video modes.
>The only across-the-board standard interface common to all clone
>video cards that I know of is the VESA interface, which only gives
>the programmer access to higher-res screenmodes in a dumb frame buffer
>manner. No standardized blitting, clipping, line drawing, display
>scrolling, sprites, or most of the other standard Amiga chipset
>features.
And yet, any modern PC can do these things an order of magnitude
faster, with the CPU, than the Amiga can with it's custom hardware.
Most game companies have their own internal libraries to deal with
this stuff, identifying the popular display chips if they need to do
hardware blits, etc. They're moving to general standards, though, at
the game API level. You need this for 3D, even game boxes are using 3D
libraries. This is also why Microsoft let game developers design the
game APIs for Windows95.
>Each card implements these functions differently and "standardizes"
>them via the Windows GDI.
Don't judge how the Amiga would support these things based on how
Microsoft did it. Windows is a very bad OS just about everywhere. Ok,
not everything is bad about it, just nearly everything. No reason to
expect GDI to be different. But they're getting better, for games
anyway, for Windows95, simply to attract games to the OS. Because most
PC games shut down Windows and run along in '386 mode, calling MS-DOS
when necessary. That's the only way to do real multimedia type stuff
on a PC right now. That's an OS issue alone, nothing all that wrong
with the PC hardware.
And going to another Amiga chipset doesn't eliminate
the issues, either. You need to poke different registers to do 32-bit
blits on the AAA chipset, you need an abstraction layer to set up
copperlists, etc.
Dave Haynie
>A "port" to a RISC based system is an expensive and long-term effort,
Presumably, you balance these factors between "you want it when?" and
"it's gonna cost me what?". With enough good programmers, you're not
talking a lifetime's work here.
>3. Rewrite AmigaOS to use multiple CPUs (should be a smaller effort than
> porting!)
Ha! All kinds of stuff would break. Little things like Disable() and
Forbid(), which aren't necessarily library calls (eg, you can't
replace them in existing code) don't know the first thing about SMP.
Neither does Paula, the device that manages interrupts for the system.
That's lots of new code, and it doesn't really solve the
problem. There are real performance and cost issues with SMP (eg, do
you want it fast, or do you want it cheap) just from the hardware
point of view. And even then, there are plenty of RISC processors that
are faster than two '040s, for the same money. Hell, there are plenty
of RISC processors that are faster than two '040s, for less than half
the money.
And I like the idea parallel processing, don't get me wrong. I had a
skunkworks project called "Gemini", designed to let us play around
with this in a less intrusive implementation. The DSP subsystem in the
A3000+ supported multiple DS3210/3207s (though the motherboard had
space for one, the Zorro III card supported two).
Dave Haynie
>Jon M. Taylor (tay...@ecs.ecs.csus.edu) wrote:
> I totally agree. The thing that made the Amiga great was that it was
>*different* to the PC.
No, the thing that made it great was that it was better and cheaper
than a PC. The problem is, PCs have gotten better, the Amiga hasn't
kept up.
>What a total waste of time .. the Amiga can never be _cheaper_ than the
>broad PC market
Of course it can. You don't think a chip company's going to charge
more for a chip because it's used in the Amiga, do you? That's the
point -- the PC market has created a market of commodity parts. In
essence, you aren't going to pay more than a part is worth, and you
get economies of scale no single company can touch. About the only
thing in the PC market that isn't commodity are high-end 80x86
chips. Fancy that. You could use the same expensive parts, cheaper
glue chips (Amigas don't have quite the architectural overhead of PCs
in terms of what MUST be on the motherboard; and you're only talking
one gate array anyway), and a cheaper processor. It's not like there's
one company you're competing with, there are tons of them. Compaq
alone is larger than Apple. This is one of the main reasons Apple
switched over to the PCI bus and a RISC chip that at least has a
chance of hitting substantial volumes (hint: PowerPC systems have been
out for a bit over a year, and it's already the highest volume RISC in
any desktop computer).
>and if they just settle for trying to do the same things its never
>gonna be better either.
Using the same parts doesn't mean you're doing everything exactly the
same way. It does mean there's a chance to compete on a level ground,
if you can run other OSs.
> If many people here get there way the Amiga will just be another version
>of the Power Mac - powerful, but too expensive (and unsuited) for home users
>and too different to the mainstream to compete effectively in the business
>arena.
No, you're totally backwards here. The PowerPC is not an expensive
microprocessor. And there are going to be very cheap versions, not
necessarily capable of running MacOS. But they'd drive a RISC A1200
just fine. And the AmigaOS is small enough to bring this kind of
performance home for a reasonable price. They can make a mass market
home system with PowerPC much in the same way they've made the A1200,
and it price points that no other PPC machine would dream of hitting
(like, you can get an A1200 for about the price of the 16MB SIMM that
every PowerMac needs).
Berend Ozceri
Amiga from ESCO.. by Dave Hay...@scala.com
> >is the chipset and not the OS that is primarily responsible for the
> >Amiga's multimedia potential.
I disagree. I'll ask you a question, and I'd like to ask you to answer
it honestly:
Here's a simplified scenario: If somebody sneaked into your room one day
while you were gone, replaced your Amiga's internals with a motherboard
that had readily available off-the-shelf parts (you know, used a
VGA/SVGA chipset for graphics plus a VGA/NTSC video encoder for video
output, a typical 16-bit/44.1 kHZ audio chipset for audio, 16550s or
other UARTS for ports, and an improved Buster for Zorro suport, etc.)
and somehow, magically transformed the OS so that everything supported
this new hardware, would you notice a difference for the worse?
You got to realize that although the Amiga is a computer with soul, the
hardware is still a blackbox that can be implemented in different ways
(some are better, some are worse). It is much more the
kernel/OS/application layers that give us the "Amiga" experience as we
know it.
In the days that the Amiga came out, the custom chipset was
revolutionary and it was what made the kernel/OS/application combination
possible becuase there was no readily available off-the-shelf component
choice that gave the same performance. You can't say the same thing
today.
Berend Ozceri
Carnegie Mellon University
Byron Montgomerie
: Mr Gavin Moran (gavinm@vsprsun_14) wrote:
: : Lets assume you could get an 'Amiga' that had a PPC604 and an SVGA chipset
: : in it, running Windows NT. Just what makes this an 'Amiga' and why would
: : ANYONE want to buy it when they could almost certainly get a PC or a Power PC
: : clone a _lot_ cheaper that would do exactly the same job.
: Wrong answer. If the Amiga was based on the stuff the rest of the
: industry was made of, IT would be a lot cheaper. It's the custom hardware
: and designs that make the Amiga expensive. If it adopted industry
: standards it could EASILY compete at industry price levels.
Well, they have to save money somewhere, I am not sure that using all the
same chips as the competition would really do that. :) All that is left is
who can get the parts cheapest and assemble them the cheapest with the
greatest quality. :)
I think eventually there will be an upper limit on the number of colors and
screen resolutions/bandwidth that will be appropriate for a home computer.
If the amiga gets to that point eventually then the need for extra video cards
would be limited. I mean having a built-in graphics card with standard specs
that is also optimised in with the rest of the hardware would be an advantage
when the practical limits on graphics cards are reached. In the mean time,
adding the capability to add external cards and updating the os to reflect it
(RTG) can't hurt. Being able to sell amigas as PCs with built-in graphics
cards wouldn't hurt either if the specs were right and the right cpu was used.
I do tend to view PCs with their graphics cards in the same vein as a human
that has to use an artificial limb. No matter how good the artificial limb,
it just isn't the same, and you feel the corresponding sympathy. :)
Regards,
BM
Bruce.Warren
Presumably if you add too many bitplanes to the display, it'll become pretty
memory hungry.
I hope it's a quicker chipset not just more planes
Bruce.
Michael van Elst
>and somehow, magically transformed the OS so that everything supported
>this new hardware, would you notice a difference for the worse?
Yes of course.. the machine suddenly grew slots and the c00l c0d3r
demos won't run anymore.
Mr Gavin Moran
: gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
: > SVGA chipsets may have a larger bandwidth but the vast majority of games
: >DONT USE THEM.
: No ? Why do you think so ?
Because theres no standard above 640x480x16 colours on the PC and most
modern games need 256 colours. Plus typical arcade games run best at 320x200
or thereabouts. Even the mighty Sony PSX runs a lowres mode for all those
great games.
: >They simply use the 256 colour MCGA mode, thus theres little
: >to no speed advantage over AGA in that sense.
: Of course there is a speed advantage..
No there isn;t. Using MCGA mode means NO BLITTER. If you want to use
640x480 then you have to work in 64K page segments.
Rob
>gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
>
>> SVGA chipsets may have a larger bandwidth but the vast majority of games
>>DONT USE THEM.
>
>No ? Why do you think so ?
>
>>to no speed advantage over AGA in that sense.
>
>Of course there is a speed advantage..
>
Yes, I have to agree. Although possibly the speed difference is partially to do with the fact
that a 486 chip runs a hell of a lot faster than an EC020 8)
>> uh uh, NO current PC could play Amiga-style games as well as the Amiga.
>
>No ? I beg to differ.
>
I think he means things like Lionheart that use 8 planes of sixteen way parallax scrolling
with a copper list back drop for sky and big alien muvers that cover half the screen. Parallax
of that kind would give your average Pentium a bit of a headache, plus the fact that it`s
running in well over 256 colours cause of the copper.
On the other hand, he may be talking about things like DUNE II or Monkey Island 2 which all
multitask quite happily ontop of WB (Amiga + M to flip WB to the front etc or just drag the
screen down).
Rob.
--
::~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*::
\@ <*=# R.P.N...@UK03.wins.icl.co.uk #=*> @/
{@}<>_<>__<>__<>__<>__<>__<>__<>__<>__<>__<>__<>_<>(@)
Christian Brandt
In article <3srhs7$13...@info4.rus.uni-stuttgart.de>, gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
|> Actually, the memory requirements for AmigaOS on a PPC machine would be
|> higher than at present, especially if 68k emulation code was included. And
|> by the time you add in TCP/IP, proper network printing/filesharing, some sort
|> of display postscript etc you'd need yet more memory. Performance is
|> irrelevent if you can't get the software you want to run on the OS.
|> NT performs nicely on a 60mHz pentium, on an PPC chip at similar clockspeed
|> it should also be quite usable.
Already out: AmiTCP, many Tools, needing just a nice allcovering
Installer-Script. And a 68k-Emulation is not to hard or too powerhungry.
|> : Aside from people who already have the Amiga OS on their system for personal
|> : preference/use, suppose you look at it this way:
|>
|> No, they;d say 'we can develop for NT or AmigaOS - which gives us the most
|> sales potential?' Doesn't take too long to figure what the answer would be.
Wouldn't they say
"Noone has a $5000 NT-Machine, but many have $1000 Amigas!"?
|> : Buy the Amiga OS port for <however much 3>, plus foo-application for
|> : <however much 4>, AND get significant performance boosts (or reduced
|> : system requirements, or both).
|>
|> Assuming the applications you want are available on AmigaOS and are current
|> with their NT or MacOS revisions.
FINAL Buisnessamigasoftware...
|> : Assuming <however much 2> and <however much 4> are about the same
|> : (the same product, after all), and that the Amiga OS port is no more
|> : expensive than MS-WINDOWS-NT, and all features are otherwise the
|> : same, why not?
|>
|> For the same reason OS/2 hasn't really set the world on fire - theres just
|> too much of an installed base of competitive OS products.
OS/2 is nearly the same as Windows and depends too much on WinOS2.
Also, its much too powerhungry. Windows made the success, because
it was cheap, need little resources (2MB RAM, 10Mb HD) and was easy
to install.
|> beat NT in anything but raw performance. Features wise NT is already a lot
|> better and would have improved even more by the time a PPC port of AmigaOS
|> was finished.
The difference between AmigaOS and NT area 24MB RAM and 1GB Harddisk, not
to mention $800 for a WinNT 3.5-Server/Client-Licence.
|> : I think I answered this elsewhere. "Making the Amiga" means making the
|> : OS, as far as I'm concerned. Other than that, yeah, that might be where
|> : I'd like to see things head, eventually. As a user.
|>
|> Thats your view, mine is that I bought the Amiga coz of the cool hardware
|> which could do things totally different to what PCs could do.
What, you bought the Amiga because of its infant hardware and
the lame behaviour compared with PCs could do?
Giggle..
|> TCP/IP would need to be integral, Network File Sharing and printing would
|> need to be integral. Hardware memory protection would need to be integral,
|> the core would have to be portable, it would need SMP capability, RTG would
|> need to be available, you couldn't assume you had the nice tight custom
|> chipset to ease the strain on the CPU for I/O and gfx functions.....
all stated Software would just need 1-2MB extra RAM, thats nothing compared
with NT/OS2/Unix, which need 10-20MB RAM for this.
And classic Customchips are a struggle just slowing the system down.
Even a $10-Chip like the CL542x is 10 times faster than the "ultralame"
Amigacustomset, a Mid-Range S3 (864-968) is even running 30-40 times
faster. Guess why... The Companies making CL and S3 have several hundred
employes and Developer, THERE ARE MORE PEOPLE WORKING ON A SIMPLE
GRAPHICCHIP THAN WERE EVER WORKING ON AMIGACOMPUTERS.
Stop Amigaslowdown, bann Customchips!
And, yes, get a live.
|> Home/Fun is where the Amiga is AT. Thats why it sold so well in Europe.
|> Thats why it still exists today. I certainly wouldn't have a computer at
|> home if I didn't enjoy using it - I have access to powerful PCs, Macs,
|> SUNS etc at work.
Oh, good argument:
"I have powerfull computers at work, so I don't need one at home"
--
Christian Brandt
Christian Brandt
In article <5007...@wbmt41.wbmt.tudelft.nl>, "Marcel Offermans" <mar...@wbmt41.wbmt.tudelft.nl> writes:
|> I'm quite used to using a larger virtual workbench. Would something like
|> this still be possible with SVGA chip technology. I'm not talking about
|> dragging screens down to reveal other screens, but I do want to be able to
|> scroll around on a big screen.
No Problem, most VGAs can do scrolling supersmooth, I am using
fvwm with a resolution of 1024x768 and scrolling 2048x1536,
Most VGAs can zoom, scale or expand screenregions (realtime ofcoz),
Antialiasing is nothing new, fast blitters, many times faster than amiga,
mpeg-decoders, cluts-converters etcpp....
|> > Why so? If you have dedicate screen memory, you're not wasting anything
|> > using direct mapped color. Plus, mapped color modes are always
|> > available.
|>
|> Well, memory costs money. I like the idea better that you can use memory
|> for graphics and `other things'. But I can see that this possibly is more
|> expensive than to have dedicated graphics and `normal' memory.
sharing memory costs money, limits ram-access for processor and video
and limits resolution and colors.
|> I agree with you that Amiga Technologies should concentrate on using
|> existing technologies as much as possible.
--
Christian Brandt
- The Stand - von und mit Thomas Boerkel - Das letzte Gefecht -
demnaechst in de.comp.sys.amiga.*
Windows ist ein Betriebssystem fuer Bloede und deswegen sehr erfolgreich.
Mr Gavin Moran
: In article <3srhs7$13...@info4.rus.uni-stuttgart.de>, gavinm@vsprsun_14 ( Mr Gavin Moran) writes:
: |> Actually, the memory requirements for AmigaOS on a PPC machine would be
: |> higher than at present, especially if 68k emulation code was included. And
: |> by the time you add in TCP/IP, proper network printing/filesharing, some sort
: |> of display postscript etc you'd need yet more memory. Performance is
: |> irrelevent if you can't get the software you want to run on the OS.
: |> NT performs nicely on a 60mHz pentium, on an PPC chip at similar clockspeed
: |> it should also be quite usable.
: Already out: AmiTCP, many Tools, needing just a nice allcovering
: Installer-Script. And a 68k-Emulation is not to hard or too powerhungry.
So AMITCP and various 3rd party tools are all stable enough to be included
in an OS release? And ESCOM could support them? And what if the OS was ported
to PPC - these tools would need to be recoded. Who would do that?
:|>:Aside from people who already have the Amiga OS on their system for personal
:|>:preference/use, suppose you look at it this way:
:|>
:|> No, they;d say 'we can develop for NT or AmigaOS - which gives us the most
:|>sales potential?' Doesn't take too long to figure what the answer would be.
:Wouldn't they say
: "Noone has a $5000 NT-Machine, but many have $1000 Amigas!"?
Lets see, even assuming AmigaOS required _much_ less memory than NT the
only way an AMIGA PPC clone would be cheaper than a Mac or NT machine is by
virtue of having 4-8 megs less memory as a requirement. That means it'll
be maybe 15% cheaper. And with so many less machines running AmigaOS there'll
be less of a market.
: |> For the same reason OS/2 hasn't really set the world on fire - theres just
: |> too much of an installed base of competitive OS products.
: OS/2 is nearly the same as Windows and depends too much on WinOS2.
OS/2 is a full 32 bit multitasking OS - its vastly different and superior
to Windows 3.x. It depends on WinOS2 because THERES FAR MORE SOFTWARE FOR
WINDOWS - hence it must make an attempt to support that software. Compared
to OS/2, AmigaOS has even _less_ support in business software.
: Also, its much too powerhungry. Windows made the success, because
: it was cheap, need little resources (2MB RAM, 10Mb HD) and was easy
: to install.
Compared to Windows its not that much more power hungry. Why do you think
PCs have been getting more RAM and faster CPUs over the last few years? -
To run Windows........
: |> beat NT in anything but raw performance. Features wise NT is already a lot
: |> better and would have improved even more by the time a PPC port of AmigaOS
: |> was finished.
: The difference between AmigaOS and NT area 24MB RAM and 1GB Harddisk, not
: to mention $800 for a WinNT 3.5-Server/Client-Licence.
People don't think that large amounts of RAM are unreasonable anymore. Most
would consider 16mb to be the 'base' for a modern office computer and its
getting hard to locate SCSI disks <500megs. You don't need an NT server in
a basic office setup, and since when did such a thing as an AmigaOS server
edition exist?
: |> Thats your view, mine is that I bought the Amiga coz of the cool hardware
: |> which could do things totally different to what PCs could do.
: What, you bought the Amiga because of its infant hardware and
: the lame behaviour compared with PCs could do?
: Giggle..
I bought the Amiga because it was innovative, exciting and different. Your
PPC Amiga would simply be yet another clone.
: |> TCP/IP would need to be integral, Network File Sharing and printing would
: |> need to be integral. Hardware memory protection would need to be integral,
: |> the core would have to be portable, it would need SMP capability, RTG would
: |> need to be available, you couldn't assume you had the nice tight custom
: |> chipset to ease the strain on the CPU for I/O and gfx functions.....
: all stated Software would just need 1-2MB extra RAM, thats nothing compared
: with NT/OS2/Unix, which need 10-20MB RAM for this.
Its not as simple as that. And on a PPC machine the memory needed would
increase, esp. if you include a 68k emulator so you can at least run some
software on the system. I'm betting the minimum usable PPC AmigaOS system
with decent features would need around 8-10 meg of memory compared with
16 for NT. Thats not an awful lot of saving and given the few apps compared
to NT theres not much point in going for AmigaOS.
: And classic Customchips are a struggle just slowing the system down.
: Even a $10-Chip like the CL542x is 10 times faster than the "ultralame"
: Amigacustomset, a Mid-Range S3 (864-968) is even running 30-40 times
: faster. Guess why... The Companies making CL and S3 have several hundred
: employes and Developer, THERE ARE MORE PEOPLE WORKING ON A SIMPLE
: GRAPHICCHIP THAN WERE EVER WORKING ON AMIGACOMPUTERS.
: Stop Amigaslowdown, bann Customchips!
If you've bought an Amiga with slots you can simply put a modern SVGA
chipset in on a card. If RTG is supported in the OS it doesn't matter if
the basic chipset isn't as fast as SVGA, especially if PCI is integrated.
: |> Home/Fun is where the Amiga is AT. Thats why it sold so well in Europe.
: |> Thats why it still exists today. I certainly wouldn't have a computer at
: |> home if I didn't enjoy using it - I have access to powerful PCs, Macs,
: |> SUNS etc at work.
: Oh, good argument:
: "I have powerfull computers at work, so I don't need one at home"
If I wanted to buy a powerful home computer it would be a PC, Amigas don't
make sense as power machines.
ESCOM need to up the specs for the Amiga chipset, put PCI in the hardware
and RTG in the OS. Hook users with cheap A1200 type enhanced machines
and then sell A4000 class machines to them with 060s. With the money thet
make they can develope a powerful SG Indy type machine at reasonable
prices with a version of AmigaOS for it. Then sell that machine as a
gfx workstation and work at scaling down the architecture so that eventually
they can offer a cheap version to the home consumer market.
Trying to just make a PPC clone with AmigaOS is just like hanging yourself,
they need to up existing Amiga standards and encourage people to migrate to
the new platform gradually.
Jonas Elfström
>
>: > SVGA chipsets may have a larger bandwidth but the vast majority of games
>: >DONT USE THEM.
>
>: No ? Why do you think so ?
>
>modern games need 256 colours. Plus typical arcade games run best at 320x200
>or thereabouts. Even the mighty Sony PSX runs a lowres mode for all those
>great games.
>
I saw Ridge Racer a while ago and it didn't look like 320x200 to me,
more like 320x256 + overscan. That is a great diff. just look at the
gigantic pixels in PeeCee games and then look at the amigapixels in
the "same" res.
>
>
>: >They simply use the 256 colour MCGA mode, thus theres little
>: >to no speed advantage over AGA in that sense.
>
>: Of course there is a speed advantage..
>
> No there isn;t. Using MCGA mode means NO BLITTER. If you want to use
>640x480 then you have to work in 64K page segments.
>
Almost all (maybe all) PC gfx-boards has a linear mode for 640x480 the
trick is how to use them and having drivers for all kinds of hardware... It sucks!
There seems to be more standard in 320x200 64k because many new released games comes
with that.
>gavan
>--
>email: G.M...@ee.qub.ac.uk | 'There can be only one!'
> or gmo...@nyx.cs.du.edu | - The Highlander
--
Jonas Elfstroem !The more you suffer !ECS...@klecs1.ericsson.se
ERICSSON MOBILE !The more it shows you! ->voice: +46 19 584455<-
COMMUNICATIONS !really care Right?! /*amigafreak, R.E.M. */
This is my point of view not Ericsson's /*Offspring, SOM, DM */
nyle landas
: >> And using true color modes for everything would be a waste of resources.
: > Why so? If you have dedicate screen memory, you're not wasting anything
: > using direct mapped color. Plus, mapped color modes are always
: > available.
: Well, memory costs money. I like the idea better that you can use memory
: for graphics and `other things'. But I can see that this possibly is more
: expensive than to have dedicated graphics and `normal' memory.
And the PC compatible industry is now looking at making motherboards with a
standard that would allow grahpics memory to be used as system memory as
well. Like the Amiga and Macs if we go video ram only then system RAM we'll
be seen as moving backwards. I realize CHIP and FAST cause some problems
but for low end systems they save a lot of money. The lowenders can't
afford to pay for 1MB or 2MB of video RAM to sit around and just do video
when it could be used for system memory. And using 24-bit colors for an
8-bit terminal is overkill. Let's not go totally anti-amiga there are still
good principals in the Amiga hardware/ OS combo.
: I agree with you that Amiga Technologies should concentrate on using
: existing technologies as much as possible.
As long as it doesn't mean loosing features that currently exist I'm all for
this.
--
/*******************************************************************\
|Nyle F. Landas |
|E-MAIL - ny...@servtech.com |
|URL - http://www.servtech.com/public/ami4000/ |
|FTP - ftp://ftp.servtech.com/pub/users/ami4000/ |
| |
| Build a better mouse trap and Bill Gates will steal all the cheese.|
\********************************************************************/
nyle landas
: >How many of those people are usinf those same Video cards to do video work?
: >Most of the people using video cards are doing so for DTP, CAD, Internet,
: >etc. not video work. We can't ignore this market.
: No, we can't and we don't have to. It is fairly easy to build a video
: compatible display, it's not even expensive.
: But for video work you do not need the old chipset.
So, NewTek was wrong and the custom chipset was useless for use with the
toaster. They should have just built it for the clones?
: >As I said before - 8 ) Just because you only want AmigaDos doesn't mean
: >there aren't many others out there who don't need the custom chips...
: Very few people want the custom chips. Most people are more concerned
: with fast and powerful graphics.
That's where RTG and PCI come in add a cheap SVGA card if you want it.
: >didn't update the custom chips with faster blitter and faster ports... These
: >can be added to the custom chips now. 8 ) And could have been done in 1992.
: Of course it _can_ be done now. But who would pay for it ? The low-end
: users ?
Yes, if we sell 500,000 A1400s with AAA I imagine that'd pay for it's
development. There are problems with dumping the custom chipsets for the
low-end market. We shouldn't ignore one market for the other.
: >I don't
: >want another non-elegant horsepower above all designed computer if I did I'd
: >own a PCC or a Mac not the Amiga with it's great Hardware/OS combo.
: Maybe you should look deeper into standard hardware.
When OCS came out in 1985 it was elegant.
When ECS came out it was a nice enhancements..
AGA is a joke..
AAA would have been nice.
What I'm refering to is that the other major platforms rely on a central CPU
ideal and by doing so force horsepower over elegance. If AmigaDos is simply
ported to RISC and SVGA are we much better?
-Nyle
ChrisBig-KahunaRampson
: The Power architecture is capable of *REAL* parallel processing. Here comes
: the Amiga! Right on the heels of everyone else!
: (Maybe you should read up on the PowerPC options before denouncing them.
: You wouldn't appear so ill-informed)
I don't appreciate being called names. I have not said ANYTHING about the
PowerPC chip and what it is capable of. I was advocating parallel processing
on the Amiga, and I said NOTHING about the PowerPC in this regard! The Power
PC is a good chip. It is basically the same chip the IBM first used in their
UNIX workstations (PowerRISC?), but more evolved. I work with all of the
different UNIX workstations here at work, so I do know what I'm talking about!
Here's a question. If you had a PowerPC CHRP box that ran your Lightwave,
SCALA,animations and super (DSP) sound, and it multitasked well, would
you be happy? I think you would. What if that OS was OS/2? You would go
for it? OK here's the million dollar conclusion. THERE IS NO AMIGA HERE!
There is also NO ROOM for an amiga in this market. If you can do what
you want in OS/2, why buy AmigaOS (or pirate it)? This is why I say that
CHRP is not right for Amiga, but the next Amiga could have a PowerPC chip
in it.
You pissed me off dude! But read the sig, I won't do it to you.
: --
: +- Maxwell Daymon -+- mda...@rmii.com -+
--
Michael van Elst
>How many of those people are usinf those same Video cards to do video work?
>Most of the people using video cards are doing so for DTP, CAD, Internet,
>etc. not video work. We can't ignore this market.
No, we can't and we don't have to. It is fairly easy to build a video
compatible display, it's not even expensive.
But for video work you do not need the old chipset.
>As I said before - 8 ) Just because you only want AmigaDos doesn't mean
>there aren't many others out there who don't need the custom chips...
Very few people want the custom chips. Most people are more concerned
with fast and powerful graphics.
>didn't update the custom chips with faster blitter and faster ports... These
>can be added to the custom chips now. 8 ) And could have been done in 1992.
Of course it _can_ be done now. But who would pay for it ? The low-end
users ?
>I don't
>want another non-elegant horsepower above all designed computer if I did I'd
>own a PCC or a Mac not the Amiga with it's great Hardware/OS combo.
Maybe you should look deeper into standard hardware.
Regards,
Maxwell Daymon
: : Making the Amiga some sort of PowerMac clone is like giving it a death
: : sentence.
: Hear Hear! Evolution, not Revolution. Evolve the Amiga into a multi-
: processor platform! Parallel computing anyone?
The Power architecture is capable of *REAL* parallel processing. Here comes
the Amiga! Right on the heels of everyone else!
(Maybe you should read up on the PowerPC options before denouncing them.
You wouldn't appear so ill-informed)
--
Maxwell Daymon
: These SVGA graphics cards for PCI, do they still have 64kB segments
: or is that only a problem under MS-DOS/ISA?
Considering that current Amiga gfx cards don't even SUPPORT 64K chunks
(except the Picasso II) I'd say your guess is a good one.
PCI is far beyond those 64K chunks.