Read all of it,


Name: John Carmack
Email: johnc@idsoftware.com
Description: Programmer
Project:
Last Updated: 02/11/2002 20:47:55 (Central Standard Time)
-------------------------------------------------------------------------------
February 11, 2002
-----------------

Last month I wrote the Radeon 8500 support for Doom. The bottom line is that
it will be a fine card for the game, but the details are sort of interesting.

I had a pre-production board before Siggraph last year, and we were discussing
the possibility of letting ATI show a Doom demo behind closed doors on it. We
were all very busy at the time, but I took a shot at bringing up support over
a weekend. I hadn't coded any of the support for the custom ATI extensions
yet, but I ran the game using only standard OpenGL calls (this is not a
supported path, because without bump mapping everything looks horrible) to see
how it would do. It didn't even draw the console correctly, because they had
driver bugs with texGen. I thought the odds were very long against having all
the new, untested extensions working properly, so I pushed off working on it
until they had revved the drivers a few more times.

My judgment was colored by the experience of bringing up Doom on the original
Radeon card a year earlier, which involved chasing a lot of driver bugs. Note
that ATI was very responsive, working closely with me on it, and we were able
to get everything resolved, but I still had no expectation that things would
work correctly the first time.

Nvidia's OpenGL drivers are my "gold standard", and it has been quite a while
since I have had to report a problem to them, and even their brand new
extensions work as documented the first time I try them. When I have a
problem on an Nvidia, I assume that it is my fault. With anyone else's
drivers, I assume it is their fault. This has turned out correct almost all
the time. I have heard more anecdotal reports of instability on some systems
with Nivida drivers recently, but I track stability separately from
correctness, because it can be influenced by so many outside factors.

ATI had been patiently pestering me about support for a few months, so last
month I finally took another stab at it. The standard OpenGL path worked
flawlessly, so I set about taking advantage of all the 8500 specific features.
As expected, I did run into more driver bugs, but ATI got me fixes rapidly,
and we soon had everything working properly. It is interesting to contrast
the Nvidia and ATI functionality:

The vertex program extensions provide almost the same functionality. The ATI
hardware is a little bit more capable, but not in any way that I care about.
The ATI extension interface is massively more painful to use than the text
parsing interface from nvidia. On the plus side, the ATI vertex programs are
invariant with the normal OpenGL vertex processing, which allowed me to reuse
a bunch of code. The Nvidia vertex programs can't be used in multipass
algorithms with standard OpenGL passes, because they generate tiny differences
in depth values, forcing you to implement EVERYTHING with vertex programs.
Nvidia is planning on making this optional in the future, at a slight speed
cost.

I have mixed feelings about the vertex object / vertex array range extensions.
ATI's extension seems more "right" in that it automatically handles
synchronization by default, and could be implemented as a wire protocol, but
there are advantages to the VAR extension being simply a hint. It is easy to
have a VAR program just fall back to normal virtual memory by not setting the
hint and using malloc, but ATI's extension requires different function calls
for using vertex objects and normal vertex arrays.

The fragment level processing is clearly way better on the 8500 than on the
Nvidia products, including the latest GF4. You have six individual textures,
but you can access the textures twice, giving up to eleven possible texture
accesses in a single pass, and the dependent texture operation is much more
sensible. This wound up being a perfect fit for Doom, because the standard
path could be implemented with six unique textures, but required one texture
(a normalization cube map) to be accessed twice. The vast majority of Doom
light / surface interaction rendering will be a single pass on the 8500, in
contrast to two or three passes, depending on the number of color components
in a light, for GF3/GF4 (*note GF4 bitching later on).

Initial performance testing was interesting. I set up three extreme cases to
exercise different characteristics:

A test of the non-textured stencil shadow speed showed a GF3 about 20% faster
than the 8500. I believe that Nvidia has a slightly higher performance memory
architecture.

A test of light interaction speed initially had the 8500 significantly slower
than the GF3, which was shocking due to the difference in pass count. ATI
identified some driver issues, and the speed came around so that the 8500 was
faster in all combinations of texture attributes, in some cases 30+% more.
This was about what I expected, given the large savings in memory traffic by
doing everything in a single pass.

A high polygon count scene that was more representative of real game graphics
under heavy load gave a surprising result. I was expecting ATI to clobber
Nvidia here due to the much lower triangle count and MUCH lower state change
functional overhead from the single pass interaction rendering, but they came
out slower. ATI has identified an issue that is likely causing the unexpected
performance, but it may not be something that can be worked around on current
hardware.

I can set up scenes and parameters where either card can win, but I think that
current Nvidia cards are still a somewhat safer bet for consistent performance
and quality.

On the topic of current Nvidia cards:

Do not buy a GeForce4-MX for Doom.

Nvidia has really made a mess of the naming conventions here. I always
thought it was bad enough that GF2 was just a speed bumped GF1, while GF3 had
significant architectural improvements over GF2. I expected GF4 to be the
speed bumped GF3, but calling the NV17 GF4-MX really sucks.

GF4-MX will still run Doom properly, but it will be using the NV10 codepath
with only two texture units and no vertex shaders. A GF3 or 8500 will be
much better performers. The GF4-MX may still be the card of choice for many
people depending on pricing, especially considering that many games won't use
four textures and vertex programs, but damn, I wish they had named it
something else.

As usual, there will be better cards available from both Nvidia and ATI by the
time we ship the game.


P.S. the high poly-bug is not a hardware bug and there is a fix the croteam (a.k.a. serious sam developers) have identified the driver specific problem with ATI and a fix will be released.

Very interesting email. It's clear he keeps an open mind, praising Ati at the proper times, even though he's had bad experiences in the past.

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>

yup, I read that e-mail a million times. ATI is sucking up to Carmack! GO ATI GO!

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor

Ya know, as long as any GF3 and Radeon 8500 can play Doom 3 at full settings without going under 30 FPS in heavy situations, then I can and we all agree to say that these cards are heavy-duty gladiators. This shows that they will last for over a year at full settings, because D3 will set a huge standard in graphical complexity and details and it definitly will squeeze out the max of any top-of-the-line card now. Even more than Aquanox ever did. I wouldn't mind playing on 640*480, as long as my Ti200 proves it can stand the future games, and Doom 3 anytime, it has done its job for easily 1 more year.

God I hope it works...
And I sure would like some Doom 3 pics of recent stuff, to see how it really is a demon card-eater, or videos if possible.

--
The other day I heard an explosion from the other side of town.... It was a 486 booting up...

I can almost garuntee you that the GF3 and 8500 won't be able to run Doom 3 at full setting at anything over 800x600 (even that would be pushing it). John Carmack has expressed in other interviews that he is counting on another generation of video cards coming out before Doom 3 can be run in all its glory. The biography-style interview that circulated last week had him saying something to the effect that a GF4 would be nice to have for the game, but still couldn't push the game to its limits. Kinda gives you an idea of just how wicked the engine is.

The holder of <i>this</i> pass is <b>exempt</b> from all forms of <font color=red>flaming</font color=red>.

First, I already posted about this.See "J.Carmakc speaks". But oh well. Second I am really looking forward to seeing the Doom 3 engine. I imagine it will be heavily used to make other games. Much like the Quake 3 engine was. As a mapper, Im not looking forward to the learning curve. But Im sure it will be worth it.
My only complaint about Carmack is he takes his sweet ass time puting a game out. Yeah, I know, his games arent buggy like some others. Also is he EVER going to make something other than Quake or Doom ?!? Clearly he is a genious, why cant he make something revolutionary instead of something evolutionary?

Benchmarks are like sex, everybody loves doing it, everybody thinks they are good at it.

Are you suggesting that it would be advisable to have a card better than the gf4 to play doom 3, that doesn't seem right to me.

P.S. your sig looks like a anti-flamethrower sig ;-)

I've got great news! People at SharkeyExtreme's SharkeyForum are saying that Carmack did show Doom3 behind closed doors on a GeForce3 at 1024*768*32 at 75FPS.

Also, Carmack was hinting that the R8500 will be abount 30% faster than that.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor <P ID="edit"><FONT SIZE=-1><EM>Edited by AMD_Man on 02/13/02 08:24 AM.</EM></FONT></P>

This is what I mean!
The thing is, you should not forget that there are code optimizations to be done! He says it won't run well, but think about it, it's still early and many games are that way until they are optimized and then run much better. We are also forgetting it's a DX8 game not DX9, so why the hell would DX8 cards not be able to play with it! That's like saying this car is so revolutionary, but only aliens in 20 years will be able to drive it! Even though we could drive it!
Otherwise like I said, even 640*480 isn't mindful to me, as it still will allow me to play something that is almost 2 years in advance from current game graphics.
AMD_MAN I do hope this article was true though!

And also, where on earth would Carmack find the right card if none works with the game smoothly? Think about the logic, you program something that won't work well, might as well trash it dontcha think? And it's also bad advertising to make new customers to GF3s, find out their card won't run the awaited Doom 3! And also since he commented in this letter on the cards, he did not give negative thoughts about them, and since R200 is the Ti500 competitor mainly, it means a Ti500 and GF3 are able to be used for D3.

In answer to the guy who asked why Carmack don't do other games, simple: They stick to where they rock at.
Kinda like Blizzard! They've been on 3 series ever since they rose in daylight. Always Warcraft-Diablo-Starcraft and repeat. Now Starcraft 2 should be something to look at once they start working on it!
AMD_MAN do you play many games? You sometimes hint as being more exploring but less of a gamer, so correct me if I'm wrong, and do tell what games you play!



--
The other day I heard an explosion from the other side of town.... It was a 486 booting up...<P ID="edit"><FONT SIZE=-1><EM>Edited by Eden on 02/13/02 04:05 PM.</EM></FONT></P>

Yes, I'm not a hardcore gamer, but I love to investigate new technologies. I bought the R8500 simply to review the card not that I actually need all that performance for games. I do play a good amount of games but I like to learn the technology behind the games I play.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor

Yes...and it looks like you will be getting a 30% increase in the poly-count once the 8500 patch is finalized!

I want to die like my Grandfather...in my sleep...not screaming in terror like his passengers.

What kind and what are the current ones you do play?
To me I play games not only for enjoyment but it's a hobby and in the future I plan to make a TV show using RPG elements to it, even battles using menus from a computer controlled! It will be totally a new experience but something worth investing and learning in it. And I love graphic technology so new games and new cards make me drool...

BTW Sacrifice still OWNZ in textures, and it's 2 years old!!

--
The other day I heard an explosion from the other side of town.... It was a 486 booting up...

Games I like to play include Quake III, UT, Messiah, AOE and AOEII, Star Trek: Armada, Star Trek: New Worlds, The Sims + Hot Date.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor

It's gonna rejuvenate the R8500! With the new drivers supposedly coming tomorrow (Feb. 14), and the R8500XT and RV250 coming in March, ATI will have more than enough competing products for the GF4Ti series plus upcoming "Ultra" cards.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor

Can't wait to see benchs then! It probably will have some competition to GF4 Ti levels for sure, and it does indeed unlock the R200 power. Remember not only it has more specs, but it runs at higher clock speeds so it should be offering more! ATI is indeed showing some responsibility nowadays!
I am wondering if their AthlonXP compatibility conspiracy is true or false now...
If you play Messiah and Star Trek NW, then you must have the Asus V8200s huh? I got these 3 games (including Sacrifice) but I did not like STNW, it lacked a lot of gameplay unlike other RTS'. Messiah's fun though however it stutters again too!
And as for Sacrifice, I dare you to finish it without difficulty...but it's a very original and nice game with humor and replayability.

--
The other day I heard an explosion from the other side of town.... It was a 486 booting up...

A 30% increase is believable (unlike 100%), but don’t get too excited yet. What’s that saying? Don’t count your framerates before the patch?

"Ignorance is bliss, but I tend to get screwed over."

I didn't state that Doom 3 wouldn't play on a GF3 or 8500, but I was saying that Carmack has stated that it will take a bleeding edge card to run the game at its max settings. He has publically said that Doom 3 will run on GeForce2s and Radeons. And you can't honestly tell me that you don't think id uses professional graphics cards to do the rendering in.

Also as far as the question over why they don't do anything other than Quake and Doom: Carmack also said that Doom 3 has, get this, a story line. He isn't talking about a "you're trapped in an alien world and need to escape" storyline, but a real modern type storyline. But he has also stated in personal interviews that he dislikes teamplay multiplayer stuff and prefers the deathmatch style of play. So I guess when you're the boss, you make games that you like.

The holder of <i>this</i> pass is <b>exempt</b> from all forms of <font color=red>flaming</font color=red>.

D3 had better continue the serie's unbeatable suspense and fright it had in Doom 2!! I want some dark areas and some background sounds out of nowhere! And I want the Cyber-Demon back! Or the Cacodemon, now that guy was scary!

--
The other day I heard an explosion from the other side of town.... It was a 486 booting up...