| [Files] Warp3D software render driver | Link |
| Posted on 19-Feb-2002 14:42 GMT by Teemu I. Yliselä | 158 comments (132k)
View flat (1, 2, 3, 4)
View list
Add comment | | ""This is a software driver for Warp3D, which enables:
using Warp3D applications, demos and games with PPC Amigas not equipped with a 3D GFx board, or not equiped with a GFx board at all (but requires P96 or CGX) [and] using Warp3D applications, demos and games not compatible with a 3D Gfx board (example : WipeOut and CGx4/Voodoo3)."" More details in the readme. |
|
Warp3D software render driver : Comment 4 of 158 | Posted by Detritus (194.222.62.98) on 19-Feb-2002 18:09:23 | In Reply to Comment 1 (David Scheibler):
Note also that the 68k version of Quake2 will not be released because physical 68k processors are too slow and it won't run on Amithlon due to an 'emulator bug'. Oh, and we do not 'officially support emulators'. Oh and we will auto-respond with 'See what happened to Sun' everytime we are asked about an x86 version of the OS.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 11 of 158 | Posted by SlimJim (130.235.59.132) on 19-Feb-2002 18:43:02 | In Reply to Comment 4 (Detritus):
Well it isn't up to Hyperion to make a x86 port of the OS, but Amiga Inc.
Hyperion only have a license for a PPC version. That they are opposed to the
idea of an x86 Amiga OS is irrelevant right now, really - I'm sure they would
consider it if Amiga Inc argumented/payed enough for it. That Amiga games sell
BETTER than Linux games (due to people buying cheap PC version, and pirating
linux exe) is a rather strong anti-x86 incentive as far as I am concerned.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 15 of 158 | Posted by redrumloa (209.86.184.182) on 19-Feb-2002 18:50:59 | In Reply to Comment 10 (Detritus):
Oh come on now! You going to force them to support MorphOS next? Why stop there? QNX was once the official next gen AmigaOS so why not force that port? Linux was also once the official path to an extent, so force that port(Oh yeah it already exists officially right?)! I've also heard that BeOS came within inches of being announced the official path at one point years ago, so god dammit it needs it to!
Reply to this comment | Top |
|
Warp3D software render driver : Comment 16 of 158 | Posted by Anonymous (80.196.135.192) on 19-Feb-2002 18:52:44 | In Reply to Comment 10 (Detritus):
Freespace is a real Hyperion game as opposed to Quake2, which just is
by the same ppl, and Freespace runs fine on Amithlon. And I'm pretty
sure (haven't tried) that Freespace 68k sw render isn't really usable
on 68k Amigas. So they are not *preventing* their games from running
under emulation. But software running fine on real amigas but not
under emulation must be up to the emulator to fix - you won't claim
that amihtlon is 100% amiga compatible, will you?
Reply to this comment | Top |
|
Warp3D software render driver : Comment 20 of 158 | Posted by Hans-Joerg Frieden (62.224.208.16) on 19-Feb-2002 20:18:49 | In Reply to Comment 9 (darklite):
sigh...
We never did anything to prevent our games from running on Amithlon, or MorphOS for the matter. Check your facts. See Freespace.
About Warp3D: Every supported system manufacturer paid for it. Amithlon didn't want to (in spite of advertising with Warp3D on the show in Cologne).
Do your math. Or don't. See if I care :-)
Reply to this comment | Top |
|
Warp3D software render driver : Comment 26 of 158 | Posted by Bill Hoggett (80.225.18.46) on 19-Feb-2002 22:57:24 | In Reply to Comment 25 (Anonymous):
Aaaargh!
Look people, give it up!! Amithlon does not have Warp3D
support because no one has bought a license. The reason no
one has bought a license has nothing to do with Hyperion.
Amithlon was demonstrated running Warp3D at Cologne, but it
was not advertised as having Warp3D support. There IS a
difference. Since Warp3D CAN be made to run on Amithlon,
we can assume it is not just PPC specific.
There is no debate that a license should be payable. No one
on either side disputes this.
Finally, AFAIK there has been no reported emulation bug that
prevents Quake 2 from running. It is common courtesy to
inform the author of any such bugs found, and if it exists
I can guarantee that Bernie will do whatever he can to fix
it.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 27 of 158 | Posted by Harald Frank (62.227.183.171) on 20-Feb-2002 00:02:05 | In Reply to Comment 26 (Bill Hoggett):
Hello Bill,
please just stop discussion about that. These guys
do *not want* that we get on realistic conditions
such a license.
There is already some discussion about that on the
Amiga-News.de comments ongoing. He know for sure
as you and me that i only demonstrated it on the
show to show ppl that the technology is there and
able to handle that. We never advertized that and
we never told anyone that we have any plans to make
such available without license. AND THAT do these
guys know for sure, because i seld told them, as
same phone call where i told them that NONE has
got these special drivers from me and that is was
NOT some kind of prometheus emulation as they think
at that time.
But anyway.. its useless.
btw, did you got from bernd the new drivers i gave
him for you and lasse ? did not get any response from
you side...
regards
VMC Harald Frank
Reply to this comment | Top |
|
Warp3D software render driver : Comment 28 of 158 | Posted by Steffen Haeuser (145.254.252.251) on 20-Feb-2002 00:15:49 | In Reply to Comment 4 (Detritus):
I *never* said
there would not be a 68k version of Quake 2. Actually I have Quake 2 for 68k
running and it works nicely on AmigaXL. It does not work on *Amithlon*, that
is true, and this I said, but it works fine on Amiga XL, the *second*
x86 Amiga Emulation available.
I reported this to Harald Frank already weeks ago, BTW, but he seems to
ignore me. BTW: My tester who is testing it on Amithlon *is* using the
latest version of Amithlon.
It all has to do with a specific Compiler-Setting, which I have to use for the
68k version (if I do not use this settings the game68k.dll has bugs, but if I
*do* use it it crashes on Amithlon - but still works fine on AmigaXL and on
physical 68k, though - on physical 68k it is too slow, of course). And as
we mentioned before - we do not officially support Amithlon, so if there
are bugs in the Emulation this is not *our* fault. And if the author of
Amithlon chooses to ignore this, I cannot do anything about it. If I find
the time I might even still try gcc 2.95.3 for 68k, if maybe this fixes the
problem, but I do not know if I find this time. And after all the flames
I get from Emulator-using people I do not know if I even WANT to invest that
extra work. And of course I do not know if using the newer compiler version
will fix anything, or if it will still crash on Amithlon.
And well - it DOES run on AmigaXL Emulation!
Well, as all Amithlon owners also have AmigaXL I do not see so much of a
problem :) They can run the game on AmigaXL.
I did not test on WinUAE, BTW, so I do not know if it runs there. Personally
I do not care about that x86 stuff anyways, but just to show you Emulation
guys how a nice guy I am I also did a 68k version :) - so please do not flame
me for actually doing a 68k version for a game which runs too slow on a
physical 68k version - next time I might decide in a different way else.
Steffen Haeuser
Reply to this comment | Top |
|
Warp3D software render driver : Comment 29 of 158 | Posted by Harald Frank (62.227.183.124) on 20-Feb-2002 01:23:12 | In Reply to Comment 28 (Steffen Haeuser):
Hello Steffen,
the problem you reported was already given to my partner Bernd Meyer
who is doing the cpu/JIT parts of Amithlon. He also got from me your
binarys to reproduce these problem on his system.
As of today i have not heard from him about that problem or if he
was able to fix that. So stop your "i dont know" asumes about we
ignore you or such untrue stuff. I told you by phone that we have
to ivestigate these problem to understand what is going wrong and
to see how we can fix that, first than we can tell you something
new and not before.
Best regards
VMC Harald Frank
Reply to this comment | Top |
|
Warp3D software render driver : Comment 31 of 158 | Posted by Bill Hoggett (80.225.27.56) on 20-Feb-2002 01:57:14 | In Reply to Comment 28 (Steffen Haeuser):
> I reported this to Harald Frank already weeks ago,
> BTW, but he seems to ignore me. BTW: My tester who
> is testing it on Amithlon *is* using the latest
> version of Amithlon.
Then your tester should know that actual emulation issues should be addressed to Bernd Meyer at bernie@amithlon.net, complete with details of the problem. It is Bernie who deals with the actual emulation engine, and if it crashes on Amithlon but not on XL or real Amigas there can only be two explanations:
(a) There is a bug in the JIT optimisations done on Amithlon which are missing from XL. This can be checked by switching off the JIT in Amithlon - it will run too slow to be really useful this way, but it is a good way to isolate JIT issues.
(b) Quake2 makes some call to the custom Amiga chipset which is emulated under XL but missing in Amithlon, which has no chipset emulation to speak of. As you mention that the problem results when using a particular compiler setting, this probably eliminates this possibility.
> Well, as all Amithlon owners also have AmigaXL I do
> not see so much of a problem :) They can run the game
> on AmigaXL.
Sigh. Ask your tester, and he should be able to explain. If I am forced to exit Amithlon and reboot to play Quake 2, why would I choose to boot into QNX and AmigaXL to run a software rendered Q2, when I could boot into Windows or Linux and run a fully 3D accelerated port of the game. Besides, not all Amithlon users bother installing QNX and AmigaXL as well.
> I did not test on WinUAE, BTW, so I do not know if it runs there.
It may be interesting to check. The WinUAE JIT engine is somewhere between the AmigaXL one - which is slower but more stable - and the Amithlon one, which is faster but also has lots of extra optimisations where new bugs may hide.
> Personally I do not care about that x86 stuff anyways,
> but just to show you Emulation guys how a nice guy I am
> I also did a 68k version :) - so please do not flame
> me for actually doing a 68k version for a game which runs
> too slow on a physical 68k version - next time I might
> decide in a different way else.
Nobody is flaming you for doing a 68k version. However, if I found a bug in your software you'd expect me to report it to you, with details of how it happened. If I reported it to Ben instead, you might not get to hear about it for some time. See?
Reply to this comment | Top |
|
Warp3D software render driver : Comment 34 of 158 | Posted by Steffen Haeuser (145.254.251.112) on 20-Feb-2002 08:19:20 | Harald: You could have at least replied to me. Last mail I got from you about the issue is "as far as I know he is using an old version" despite the fact that I earlier told to you that he is using the latest version now. I tried yesterday on irc to /msg to you that he is using the latest version now, but either you did not see the msg or you did not want to see. Anyways, I had the impression that nothing would happen on that bug. I also asked you several times if you could email me or Bolton the debug-tool you told me of earlier,
which might help identifying where in the source it actually crashes, but I never got a reply on that either. So after not getting any replies I assumed
you'd put the blame on me if the final Quake 2 version would not work on Amithlon. Prove me wrong!!! I still will include that 68k version with the
port :)
Bill: It is not a JIT issue, the problem also happens with JIT disabled. Neither is it a custom chip issue, as it only happens with a certain compiler setting (which I have to take, though, as if I don't do this setting, the game68k.dll has bugs). Strangely BTW the self-running demo does not crash the game68k.dll even with the setting being present. Only if I really start a new game the crash happens.
>Sigh. Ask your tester, and he should be able to explain. If I am forced to >exit Amithlon and reboot to play Quake 2, why would I choose to boot into QNX >and AmigaXL to run a software rendered Q2, when I could boot into Windows or >Linux and run a fully 3D accelerated port of the game. Besides, not all >Amithlon users bother installing QNX and AmigaXL as well.
But then why would you choose to boot into Amithlon to run a software rendered Q2 when they can run the Windows/Linux version as well ? That was our doubt on x86 Amiga Emulation from the start, see ? There's no Warp3D-licence for Amithlon, though the Amithlon-makers should contact the Friedens about a licence if they are interested in Warp3D.
>Nobody is flaming you for doing a 68k version. However, if I found a bug in >your software you'd expect me to report it to you, with details of how it >happened. If I reported it to Ben instead, you might not get to hear about it >for some time. See?
But then it is not a bug in *my* software. On a physical 68k (and on AmigaXL)
it runs fine. Emulation on x86 is not officially supported by Hyperion Entertainment.
Steffen Haeuser
Reply to this comment | Top |
|
Warp3D software render driver : Comment 37 of 158 | Posted by Bernd Meyer (130.194.226.138) on 20-Feb-2002 11:15:18 | In Reply to Comment 36 (nOw2):
Isn't it obvious? You don't have *any* hardware that anyone paid a Warp3D licence for....
Sorry, it's been a long and crappy day, and it's going to be a much longer and crappier night. And this whole stupidity really gets up my nose!
I don't have a problem with Hyperion saying "If you don't have any licensed hardware, you can't use Warp3D". That's cool, it's their decision, and while I personally doubt it's a wise one, that really doesn't matter.
What shits me is that Hyperion is now saying "You are welcome to use Warp3D if you have bought licenced hardware, *or* if you happen to own a PPC card, but not, repeat not, if you dare to run your Amiga on an x86". Someone please explain to me the difference between people with PPC accellerators and people with x86 accellerators, as far as Warp3D is concerned! That makes the whole "We don't oppose x86" stuff seem really really hollow....
Reply to this comment | Top |
|
Warp3D software render driver : Comment 39 of 158 | Posted by CD32Freak (202.160.12.120) on 20-Feb-2002 11:44:05 | "My name is Maximus Fatterus Agnusius, Commander of the Armies of the North and General of the Amiga Legions. Loyal servant to the true emperor, Jay Minerelius...mother to murdered daughters, wife to a murdered husband and I will have my vengence, in this life or the next."
Amiga Gladiator
:D
Reply to this comment | Top |
|
Warp3D software render driver : Comment 40 of 158 | Posted by Kay Are Ulvestad (129.241.102.40) on 20-Feb-2002 12:03:10 | In Reply to Comment 37 (Bernd Meyer):
>Isn't it obvious? You don't have *any* hardware that anyone paid a Warp3D >licence for....
It IS obvious. The only Amiga hardware Amithlon users have is emulated. So if someone is to pay a license, it should be the people who are selling the emulator.
>Sorry, it's been a long and crappy day, and it's going to be a much longer and >crappier night. And this whole stupidity really gets up my nose!
If it is any comfort, rest assured that Hyperion has to put up with a lot more stupidity than you do. Every time they are mentioned here on ANN, some mud-for-brains with no regard to facts start flaming them.
>I don't have a problem with Hyperion saying "If you don't have any licensed >hardware, you can't use Warp3D". That's cool, it's their decision, and while I >personally doubt it's a wise one, that really doesn't matter.
>What shits me is that Hyperion is now saying "You are welcome to use Warp3D if >you have bought licenced hardware, *or* if you happen to own a PPC card, but >not, repeat not, if you dare to run your Amiga on an x86". Someone please >explain to me the difference between people with PPC accellerators and people >with x86 accellerators, as far as Warp3D is concerned! That makes the >whole "We don't oppose x86" stuff seem really really hollow....
It is their call. If they don't want to make a software renderer for a dead processor range that's their choice.
Kay
Reply to this comment | Top |
|
Warp3D software render driver : Comment 41 of 158 | Posted by Bernd Meyer (130.194.226.138) on 20-Feb-2002 12:18:28 | In Reply to Comment 40 (Kay Are Ulvestad):
The guy had a CSPPC and a PIV, and couldn't get it to work....
Anyway --- It's not that the W3D team "doesn't want to make a software renderer for a dead processor". The software renderer for 68k *exists*, and was done by the same person who made this PPC one. Heck, you should hear the guy talking about how things fly on his Amithlon system.
What happens is that the W3D people for some reason forbid the release of the existing driver for 68k, and *at the same time* allow the release of the same software renderer, by the same author, for PPC. And as far as I can tell, their reason for not allowing the 68k software renderer out is "because then people could use Warp3D without having any licensed hardware".... which is all fine and good, *except* that the PPC software renderer, which they apparently don't have a problem with, allows that very same thing. It's just that those people without licensed hardware apparently are "good" people, because they have a PPC, as opposed to the "evil" people, who have an x86
All I am asking is for someone to explain this discrepancy. If that takes the form of "Well, we really don't like x86, and we can, so there!", fine. That would be honest. If it takes the form of "Oh, well, we can't really allow that, because W3D in software render on an Athlon 1700 runs faster than on any Amiga PCI card, and that would majorly upset our licensees", that's fine, too (I have no idea about performance figures one way or the other!). I just find it rather distasteful and insulting to the users' intelligence to cite "licensing issues" when those very same issues are ignored for the PPC software renderer.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 42 of 158 | Posted by Christoph Gutjahr (62.104.208.74) on 20-Feb-2002 13:33:19 | In Reply to Comment 41 (Bernd Meyer):
Bernd:
I see your point. But what do you think how many CSPPC/BPPC cards were sold since
Hyperion made W3D a commercial product ? There's simply nobody who will pay
licensing fees, and (practically) no PPC cards are sold anymore. That's why
it is fine to give the PPC guys a software-renderer.
But Amithlon sells in big numbers - see the difference ?
Reply to this comment | Top |
|
Warp3D software render driver : Comment 43 of 158 | Posted by Bernd Meyer (130.194.226.114) on 20-Feb-2002 13:44:15 | In Reply to Comment 42 (Christoph Gutjahr):
No, sorry, I *don't* see the difference. Either it is a commercial product, or it isn't. Or at least if you make it a commercial product for people you don't like the look of, and a non-commercial one for people you like the look of, then at least have the decency to come out and say so.
Please note that I have never ever uttered a word against the W3D team for treating W3D as a commercial product, and insisting on a license. I have been writing and selling software for, yikes, 18 years now, so I really have the highest respect for anyone else who does produce a good product and tries to make a few bucks of it. No problem *at all*.
Supposedly, the New Amiga Future (be it Pegasos, or AmigaONE, be it OS4.0 or MorphOS, or maybe even 3.9 in emulation) is just around the corner. And it's supposed to be built on fast PPC processors, and to use the very same gfx cards Amithlon uses. That's not a handful of old and decrepit Phase5 cards which nobody dare send to DCE for repairs, for fear of losing them completely. That's a whole bunch of new machines, or so it is said, which I get told again and again will, real soon now, outsell Amithlon. And gee whizz, isn't it funny --- they would have a W3D software renderer available. I am just *sure* that bplan and Eyetech will simply rush to their respective banks to transfer the licence fees to the W3D guys.....
Reply to this comment | Top |
|
Warp3D software render driver : Comment 44 of 158 | Posted by Ben Hermans/Hyperion (134.58.222.8) on 20-Feb-2002 14:00:48 | In Reply to Comment 43 (Bernd Meyer):
It should be pointed out that the software renderer was not done by Hyperion but by Stéphane Guillard.
As such, it doesn't fall within the scope of the license granted to Elbox, Matay and DCE.
Nor can it as we don't own it.
We did give Stéphane access to the source-code but that doesn't mean we own the software renderer.
Hence this entire discussion here starts from an incorrect basis.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 45 of 158 | Posted by Bernd Meyer (130.194.226.114) on 20-Feb-2002 14:21:16 | In Reply to Comment 44 (Ben Hermans/Hyperion):
Ben,
Could you please elaborate? From everything I have read publicly and privately from Stephane, it is you guys who basically told him not to release the 68k (or the x86, for that matter) software renderer due to "licensing problems".
Now you seem to say that none of this has anything to do with you, and that Stephane is free to do what he wants --- which, it was my impression, is very much to release those versions of his work.
Which is it? Either you guys *are* telling him what he can and can't release (at which point I'd really really like to have the apparent discrepancy explained), or you aren't. But if you are *not* stopping Stephane from releasing his work, then please say so unambiguously and publicly.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 46 of 158 | Posted by Ben Hermans/Hyperion (134.58.222.8) on 20-Feb-2002 15:35:26 | In Reply to Comment 45 (Bernd Meyer):
As far as I understand, Stéphane has agreed not to release the 68K software renderer because it is too slow for a physical 68K and would therefore only benefit Amithlon users.
Seeing as there is no license agreement for Warp3D for Amithlon, this decision is perfectly legitimate.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 47 of 158 | Posted by Georg Steger (195.254.227.69) on 20-Feb-2002 15:40:50 | In Reply to Comment 34 (Steffen Haeuser):
> Bill: It is not a JIT issue, the problem also happens with JIT disabled. Neither is it a custom chip issue, as it only happens with
> a certain compiler setting (which I have to take, though, as if I don't do this setting, the game68k.dll has bugs). Strangely BTW
> the self-running demo does not crash the game68k.dll even with the setting being present. Only if I really start a new game the
> crash happens.
This reminds me of something: years ago when Quake was ported to AROS it
used to crash. But only with egcs, not with gcc 2.7.2. Egcs was known to
be buggy so we were sure that it had to be a compiler bug. After all it
worked fine with gcc.
Guess what: ages later it turned out that there were some real bug somewhere
in AROS which caused this crash. And only because of luck this bug did not
cause any crash with gcc.
So actually it turned out not to be egcs's fault, althoug it initally looked
so.
--> the fact that you have to use a certain compiler setting (which one, BTW)
in order to get a non-buggy game68k.dll might indicate that there's real
bug in the code somewhere, ie. not necessarily a compiler issue.
Reply to this comment | Top |
|
Warp3D software render driver : Comment 48 of 158 | Posted by Bladerunner (80.133.21.64) on 20-Feb-2002 15:42:31 | In Reply to Comment 43 (Bernd Meyer):
But there *IS* a difference.. Matay, DCE, and even Elbox are paying the license fee for W3D.
Could you give a reason why W3D for Amithlon should be free? There is still a really bad discussion on Amiga-news about
that, and i still wonder if the ppc software renderer is legal, because i have to use a W3D V4.2 which is only legal if i use
a matay/dce/elbox board...
It would be nice to get a clear Statement from Hyperion about the W3D V4.2 Version..
But..it is one thing to have a free softwarerenderer for old Classic without 3D Hardware which will never be so
powerfull as a full supported W3D Archive for Amithlon.. Why don`t you still pay the license fee for Amithlon?
Reply to this comment | Top |
|
Warp3D software render driver : Comment 49 of 158 | Posted by David Scheibler (217.235.119.236) on 20-Feb-2002 15:59:20 | In Reply to Comment 48 (Bladerunner):
>Could you give a reason why W3D for Amithlon should be free?
For the same reason PPC card users get it for free? They also never paid any licencing fees and DCE license only covers Grex. At least that's my guess.
What if the Pegasos is released and the software driver works in some MorphOS emulation framework? Can't Stephane then distribute the PPC software driver anymore because bplan doesn't pay license fees?
Reply to this comment | Top |
|
Warp3D software render driver : Comment 50 of 158 | Posted by Bernd Meyer (130.194.226.136) on 20-Feb-2002 16:22:02 | In Reply to Comment 49 (David Scheibler):
Well put, David.
Let me stress again that I don't want anything for free, and have no problem at all with the W3D team or Hyperion not giving away their work for free!
The only thing I *do* have a problem with is that the software renderer for PPC somehow is OK (and I can hardly see the users of officially licensed and hardware-supported 3D cards being the target audience....), while a software renderer for Amithlon is somehow not OK.
Ben, a few more questions:
a) Who do you see as the target audience for the PPC software renderer?
b) Stephane "agreed" not to release the 68k (and x86) renderer --- agreed with whom? Agreed to whom's request?
c) Seeing as Warp3D 4.0 is freely distributable, would it, from your point of view, be OK to release a renderer for *that* version?
Reply to this comment | Top |
|
[Top] [Next 50] Add comments | - Links
- Support ANN
- ANN is hosted by Dreamhost. Sign up through this link and I will get a 10% commission on any account you purchase.
|
