Doom 3 shader mod for ++ performance

[SuperRob]

Read my above two posts for #1 and #3.

For #2, if it was a worthwhile optimization, it could have been proposed to id software through ATI's channel with them and implemented officially in the next patch without all of this hoopla. It would have helped gamers just as much that way.

And I say again, since when did we ever care about things being done through official channels? In fact, I thought we prided ourselves on specifically NOT staying in official channels. Isn't that what being HARD is all about?

 

[trungracingdev]

Bcoz this is John Carmack's creation not ATi

and please do not say " shit " this word.

Again, This is John Carmack's creation not ATi , and we care as well as John Carmack

If u dun give a shit about Carmack then dun play this game and dun make any useless and unneccessary comments.

I have to care about JC to play DOOM 3? Since when was this a freakin requirement?

 

[tranCendenZ]

I never said I knew more, I just think its funny you always present the same side of the argument with the same pr lines.

PR highlights the best facets of a card, so it only makes sense that those facets would be repeated when discussing the best facets of a card in an argument.

I really don't understand your perspective on this, its obvious you are an nvida fan, with good reason, you spent your hard earned money on a 6800gt and you think it’s the best choice and you’re portraying that opinion.

I like Nvidia better than ATI currently, yes.

That’s fine and all, but you are taking it so far to the next level (shutting off anything positive about ati) it is just stupid.

??? How does anything I posted above shut off anything positive about ATI? According to you this has nothing to do with ATI, and whether the code was publically or privately sent to Carmack, if it was worthwhile it would have been made available to ATI users.

Like reverend said : the code could not be changed so easily if carmak actually didn’t want it changed.

Very true. It's just sketchy that an ATI employee is messing around with the graphics code on a public forum. If Humus was a random coder thats one thing, but he works for ATI. He could have easily contacted Carmack privately - in fact he probably would have gotten a faster response than doing all the public posting and experimenting. Again, Carmack might not give a crap, but out of respect if I was in Humus' position I would propose the changes to him privately instead of in a public forum since I would have the means at my fingertips. The tone of Humus' first posts also implied that Carmack's code needed to be "fixed" as if something was wrong with it, which also wasn't too cool. And his "fix" didn't even work right at first, and it still doesn't look identical.

So basically you are hoping for optimization wars 3000, which company can out-optimize the other with just minor IQ impact.

Who doesn't want that?

Might be fun, but in the end if it goes far the devs/publishers lose control and the end user could lose out as optimizations get more and more aggressive as the companies compete.

 

[KakimotoR]

I have to care about JC to play DOOM 3? Since when was this a freakin requirement?

I am telling u to show some respects to John Carmack

and u said u dun give a shit about wat John Carmack thinks,and u are playing his game and his engine and that's all done by his thinking and without his thinking , and imagination and talent, DOOM3 wouldn't even be here. If u dun give a shit about wat he thinks then dun play the game.

 

[^eMpTy^]

And I say again, since when did we ever care about things being done through official channels? In fact, I thought we prided ourselves on specifically NOT staying in official channels. Isn't that what being HARD is all about?

I've said this before but I think the problem here was that he threw up a big thread that promised a 40% performance boost before he had done any testing or even bothered to run the benchmark several times to make sure he could duplicate the performance increase...

I think ideally you'd want to first test it, make sure it's consistent...check for image quality issues on different cards...and THEN make a big deal out of it...

But whatever...none of it matters now...it seems to be turning out well enough...

 

[Three D. Marks]

Here's an idea. Shut your fanholes and go play Doom3.

 

[101998]

??? How does anything I posted above shut off anything positive about ATI? According to you this has nothing to do with ATI, and whether the code was publically or privately sent to Carmack, if it was worthwhile it would have been made available to ATI users.
.

It doesn't have anything to do with ati, you are the one who keeps bringing them into this, not me. You are the one who keeps saying how Humus is disrespecting blah blah blah blah blah. That is what I don't understand, its not that you are saying nvidia is better, you are saying ati is a company with no morals and employs people who are unscrupulous. What is the point?

The sad thing is that if the situation were reversed you would be singing the praises about how nvidia employee’s really care and love their jobs and how its so great they did something to help the community. You can’t even deny that. Funny thing is I would be right there with you agreeing, yeah nvidia has really turned around from the company that made drivers that didn’t render parts of a benchmark to really making a helpful addition to the community.

Bah, I'll never understand zealots like you.

 

[101998]

Here's an idea. Shut your fanholes and go play Doom3.

That rocks.

 

[SuperRob]

I've said this before but I think the problem here was that he threw up a big thread that promised a 40% performance boost before he had done any testing or even bothered to run the benchmark several times to make sure he could duplicate the performance increase...

I think ideally you'd want to first test it, make sure it's consistent...check for image quality issues on different cards...and THEN make a big deal out of it...

But whatever...none of it matters now...it seems to be turning out well enough...

If he were doing this through official channels, then I'm sure he would have done exactly that, but I don't think he ever "promised" anything. As a guy doing this in his free time, I don't see a problem with it. Sure, he should have hedged his language, maybe saying "up to a 40% increase". At the same time, I think we're all smart enough to know to take any estimations of performance increases with a grain of salt, since it's always dependent on hardware configurations. Some people will get more, some people will get less.

 

[trungracingdev]

I am telling u to show some respects to John Carmack

and u said u dun give a shit about wat John Carmack thinks,and u are playing his game and his engine and that's all done by his thinking and without his thinking , and imagination and talent, DOOM3 wouldn't even be here. If u dun give a shit about wat he thinks then dun play the game.

I rarely know the name of any developers, bc frankly....I dont need to and I dont care. If theres a hack that can improve performance, I'll use it. I dont see how thats disrespecting anyone.

 

[trungracingdev]

It doesn't have anything to do with ati, you are the one who keeps bringing them into this, not me. You are the one who keeps saying how Humus is disrespecting blah blah blah blah blah. That is what I don't understand, its not that you are saying nvidia is better, you are saying ati is a company with no morals and employs people who are unscrupulous. What is the point?

The sad thing is that if the situation were reversed you would be singing the praises about how nvidia employee’s really care and love their jobs and how its so great they did something to help the community. You can’t even deny that. Funny thing is I would be right there with you agreeing, yeah nvidia has really turned around from the company that made drivers that didn’t render parts of a benchmark to really making a helpful addition to the community.

Bah, I'll never understand zealots like you.

word...

although he'll either ignore it....or try to twist ur wording some way to make himself look better...

 

[CrimandEvil]

Some of you should be reminded of this:

This thread has gotten so far off track it isn't even funny.

The issue at hand is that:

a.) someone claimed to have found a performance tweak

b.) it causes artifacts

c.) people are getting different results, some people like it...some people don't. Right now it seems 50-50.

d.) the "hack" post could've been worded a little better, it did come off as bashing Carmack whcih upsets people.


PLEASE get off the "you have an ATi card, that makes you stupid" or the "GeForce owners lick balls" and all that other crap. Lay off the bashing.

Cool?

You can go back to your flaming now

 

[CrimandEvil]

If theres a hack that can improve performance, I'll use it. I dont see how thats disrespecting anyone.

How about this then?

let me see if I got this right, enter h4X0r get 40% increase in FPS that is actually a 5% increase but your image quality drops significantly and there are artifacts everywhere as well as complete areas of the map lit up bright white (when they weren't before)....

A warning to many of you considering this, there is a good deal of visual artifacts from this "hack"...I know everyone is singing it praise, but if Carmack himself had done this we would all be bitching to high heaven about "lame programming" that produces visual artifacts.


I have my very own hack that gives you 40% increas in FPS with no artifacts GUARANTEED:

Start DOOM 3 as you normally would

hit escape to skip opening animation and bring up the menu

at the menu select options

select "system" from the submenu

under the "video quality" setting move it from:

ULTRA QUALITY down to HIGH QUALITY

or

HIGH QUALITY down to MEDIUM QUALITY

(and so on)

next, select the "screen size" option setting and move it from:

1600 x 1200 down to 1280 x 1024

or

1280 x 1024 down to 1024 x 768

(and so on)





.....


I tried this, and I got a 40% increase in frame rate with no visual artifiacts

 

[KakimotoR]

I rarely know the name of any developers, bc frankly....I dont need to and I dont care. If theres a hack that can improve performance, I'll use it. I dont see how thats disrespecting anyone.

I dun care if u use hack or not, that's ur choice, but just not to disrespect anyone.

Yeah and go try set ur IQ to Low quality that gain u performance

 

[trungracingdev]

How about this then?

very lame...

 

[CrimandEvil]

How is it lame?
Your the one arguing over three extra frames.

 

[trungracingdev]

How is it lame?
Your the one arguing over three extra frames.

i'll let u figure this out on ur own...

 

[CrimandEvil]

I have, like I said:

Your the one arguing over three extra frames.

lol

 

[SuperRob]

Well, this had definitely gone far past civil. Time to leave.

 

[GabooN]

IQ with and without the HACK PK file.

1024x768 High Quality
9800 NoN Pro
CATS 4.9b


I am getting some nice gains with this hack and have been using it for at least 5 hours of DooM3 gameplay today. I didn’t notice any loss of IQ in game with this hack and I wanted to compare SS to see if there was a difference and I am afraid there is a IQ difference with this hack on my system. I do not notice this IQ difference while playing the game NOR do I play this game SS by SS so I prolly would never of noticed any IQ differences. It seems that with the hack ON the lighting isn’t as sharp as with the hack off, as far as any kind of tearing or snow I have not had any issues like that at all. I personal will continue to play with this hack as it is very hard to tell the difference while enjoying the atmosphere of the game. As for Humas I don’t care if he worked at Nvidia or ATi I am glad he is trying to tweak the game and in turn trying to help out the community. Below are the SS..

 

[’m‚³‚ñ]

IQ with and without the HACK PK file.

1024x768 High Quality
9800 NoN Pro
CATS 4.9b


I am getting some nice gains with this hack and have been using it for at least 5 hours of DooM3 gameplay today. I didn’t notice any loss of IQ in game with this hack and I wanted to compare SS to see if there was a difference and I am afraid there is a IQ difference with this hack on my system. I do not notice this IQ difference while playing the game NOR do I play this game SS by SS so I prolly would never of noticed any IQ differences. It seems that with the hack ON the lighting isn’t as sharp as with the hack off, as far as any kind of tearing or snow I have not had any issues like that at all. I personal will continue to play with this hack as it is very hard to tell the difference while enjoying the atmosphere of the game. As for Humas I don’t care if he worked at Nvidia or ATi I am glad he is trying to tweak the game and in turn trying to help out the community. Below are the SS..






[3MG]http://www.ocgamers.net/site/files/iq1.jpg[/IMG]



[3MG]http://www.ocgamers.net/site/files/iqa.jpg[/IMG]



[3MG]http://www.ocgamers.net/site/files/iqb.jpg[/IMG]

Same spot looked better on my friend's 5700 Ultra.

Seriously though, if the game is playable, why bother with the hack? If you're getting performance issues, to the point where you can't tell if you're using a 9800 Pro or a 5200, then bitch and moan about it. Otherwise, just play the fucking game.

 

[Xenozx]

IQ with and without the HACK PK file.

1024x768 High Quality
9800 NoN Pro
CATS 4.9b


I am getting some nice gains with this hack and have been using it for at least 5 hours of DooM3 gameplay today. I didn’t notice any loss of IQ in game with this hack and I wanted to compare SS to see if there was a difference and I am afraid there is a IQ difference with this hack on my system. I do not notice this IQ difference while playing the game NOR do I play this game SS by SS so I prolly would never of noticed any IQ differences. It seems that with the hack ON the lighting isn’t as sharp as with the hack off, as far as any kind of tearing or snow I have not had any issues like that at all. I personal will continue to play with this hack as it is very hard to tell the difference while enjoying the atmosphere of the game. As for Humas I don’t care if he worked at Nvidia or ATi I am glad he is trying to tweak the game and in turn trying to help out the community. Below are the SS..

I think its funny how you used 2 different screen shots to try to prove a point. your results are shit and i can prove it.

take the 2 no hack screens. Measure the top from where it says in yellow "No Hack IQ" compared to the top of the image. They dont measure out. If oyu where using the same image WHICH YOU SHOULD BE" they would be the same distance apart but for some reason they are not, I dont know why your playing games but this is BS and you know it! especially on a 9800 pro. stop playing with peoples heads. "Oh i get 28 fps increase this makes ATI better, well not really but ill make them think i do!" BS!

edit:

just to point it out clearly

edit2: crap sorry about the picture thingy. but you can see what im talking about there.

 

[The Batman]

I think its funny how you used 2 different screen shots to try to prove a point. your results are shit and i can prove it.

take the 2 no hack screens. Measure the top from where it says in yellow "No Hack IQ" compared to the top of the image. They dont measure out. If oyu where using the same image WHICH YOU SHOULD BE" they would be the same distance apart but for some reason they are not, I dont know why your playing games but this is BS and you know it! especially on a 9800 pro. stop playing with peoples heads. "Oh i get 28 fps increase this makes ATI better, well not really but ill make them think i do!" BS!

edit:

just to point it out clearly

edit2: crap sorry about the picture thingy. but you can see what im talking about there.

Heh, that reminds of another asswhad I ran into a few posts back. He was ranting on about how the 4.9s and improved his Pro's performance to the point where it could almost match my GT. He even posted screenshots as proof. Moron didn't bother to notice that the screenshot with the higher score was at 1024x768 while the screenshot with the lower score was at 1600x1200. Ah well, I got a good laugh out of that.

 

[Xtasy0]

I think its funny how you used 2 different screen shots to try to prove a point. your results are shit and i can prove it.

take the 2 no hack screens. Measure the top from where it says in yellow "No Hack IQ" compared to the top of the image. They dont measure out. If oyu where using the same image WHICH YOU SHOULD BE" they would be the same distance apart but for some reason they are not, I dont know why your playing games but this is BS and you know it! especially on a 9800 pro. stop playing with peoples heads. "Oh i get 28 fps increase this makes ATI better, well not really but ill make them think i do!" BS!

edit:

just to point it out clearly

http://home.comcast.net/~Xenozx/liar.jpg

edit2: crap sorry about the picture thingy. but you can see what im talking about there.

uhm, perhaps i'm missing your point, but maybe he took the screenshot, then did the side by side image and then added the text seperately in each image? whos to say he had to add the "no hack iq" text before making the side by side image?

personally i'd do it that way, i'd take my screenshot, save it as a psd file, then i'd take my other and save it as psd then i'd splice them together for the side by side shot and the very last thing i'd do would be add the text....so i dont see how the text spacing in two different images proves anything?

 

[w00t69]

I would lol.

And i doubt over half his customers are ATI users. nVidia has a slightly higher market share then ATI, although they are about tied right now i believe. DOOM 3 is also an nVidia branded and endorsed game so i can safely say more nVidia users are running it then ATI users. Especially after the DOOM 3 benchmarks.

PLUS, nVidia and ATI are just a small portion of the GPU market. They both have around like 23% or so market share. Intel owns the majority of the graphics market and VIA has a small chunk as well.

Will DOOM 3 even run on an Intel integrated graphics setup? lol

He's not exactly screwing anyone over by not adding it in ither.

Its just like the SM 3.0/2.0b path in Far Cry. They give extra performance but they aren't really needed.

And it looks like this HACK in D3 for the R300/R400/NV30 cores lowers the IQ so that will be a major reason not to implement it.

The answer is no, it will not run on Intels' integrated or their agp solutions. None of them even support hardware T & L. It has a hard time running Halo even. As a matter of fact i don't believe it can run Halo...

It does not lower the IQ, he has shown that there is zero pixel difference between both screenshots already. Also from personal experience, this thing does work, and the cases where people are getting 3-5 fps boosts are with shadows on. I have tested both with and without the tweak with shadows off both times (helps tremendously in multiplayer which is why they are off).

Settings are 1024 x 768 High detail (no shadows)

Without tweak first run: 38.7 fps
Without tweak second run: 44.1 fps

With past tweak first run: 45.4 fps
With past tweak second run: 51.9 fps

With latest tweak first run: 46.3 fps
With latest tweak second run: 53.7 fps

Catalyst 4.9 beta (will try and test 4.8 later)

So there is definitely an improvement. I will also try and give some numbers with shadows on later.

And from reading this entire damned thread, all i see is people saying it like he was pushed by his employment at ATI to do this. He wasn't the only one looking for tweaks you know, there are other shader modifications floating around. Just because he had the knowhow and means doesn't mean he was pushed by something like this. He has been doing opengl and d3d demos for some time, helps lots of people out with shader and c++ 3d code on some other forums i'm on. IMO if it really does help this much, i don't see why Carmack couldn't have made different render paths. He knows the cards are very different in respects, why would he strive to push only one render path when doing 2 separate ones can give ALL users a greater experience, than just 50% of them. He had plenty of time to do it too. I know he wanted it to have the same reactions on all platforms, I guess he did it mainly for compatibility. But this one line of code changed makes a 22% difference for me. That's disgustingly better. I don't know about anyone else. But for me to go from being able to run only 800 x 600 high up to 1024 x 768 high is a big difference for me...

Edit: Have the numbers for shadows (only with newest tweak vs no tweak though) Same settings 1024 x 768 high detail (shadows)

Without tweak first run: 34.8
Without tweak second run: 38.4

With tweak first run: 37.6
With tweak second run: 42.2

So it does make a difference, at least for me it does. I noted some spots where it would drop to 17 without the tweak where it dropped to only 24 fps afterwards. Also some other spots that were dropping to about 22 fps were up to 30 fps. It definitely makes a difference.

 

[Manny Calavera]

There was no need to go public with this. Did you see Nvidia post the Shader Model 3.0 patch on Beyond3D before it was officially available from FarCry even though it could have given boosts to Nvidia users? No. Does Humus work for a graphics company? Yes. Could he have done this the proper way through the proper private channels instead of publically busting open carmacks code and "fixing" it his way since he works for ATI? Yes.

So the question is, why did he go through the public channel?

Could be any number of reasons:
A) Because he didn't want to wait to do it the official way and just wanted to get the tweak out there after he discovered it, and didn't think it would be a big deal. Totally innocent, just wanting to help, just never thought about the negative impacts of what he did and said and how they may reflect on Carmack and his engine being that he is an ATI representative, not to mention the community reaction.

B) To attempt to lessen Nvidia's Doom3 performance lead with an unofficial release that might not translate into an official patch, similar to how he did with the sm2.0b Dynamic Branching Demo & Nvidia's Shader Model 3.0 when the FarCry patch came out... because of personal negative feelings towards nvidia who didn't hire him; to quote humus when he released the dynamic branching demo: "nvidia can consider themselves pwned!"

C) To attempt to lessen Nvidia's Doom3 performance lead with an unofficial release that might not translate into an official patch, similar to how he did with the sm2.0b Dynamic Branching Demo & Nvidia's Shader Model 3.0 when the FarCry patch came out... under the supervision of ATI, as something like this could never be officially released by ATI themselves. One can say Humus did it on personal time all one wants, but in the end he does work for ATI; plus, 2 out of the 3 "personal" demos he programmed since he was hired by ATI used ATI-specific technology designed to show that ATI is as good or better than Nvidia in the areas the demos showcased, and his Doom3 tweak only improves performance on ATI cards, not NV cards.

D) Because he has a big ego and wanted to feed it, including publically examining and changing Carmack's code to what he felt was better.

Could be any of these reasons, but those are some ideas.

Should what Humus have done be a model graphic companies should follow? Should Nvidia get a rep in beyond3d that can bust open half life 2's shaders if they suspect optimization can be done, then tell gabe how he didn't optimize properly and provide a optimized fix? Should all new games just be busted open and reoptimized by representatives of the respective companies in public forums? It's a slippery slope.

I would go with C and D,btw,best post so far.

 

[lopoetve]

It is not the same. One is a game, the other is a benchmark only. Giving people the option to get more frames, while retaining virtually all of the same IQ is not a bad idea. If people do not like it, they do not have to use it.

How this can be a bad thing, is beyond me. Besides, a "Graphic card company" didnt change the code. Humus did this without any endorsement from ATi, just on his own. Many people are glad he did it.

Um, because it ISN'T the same IQ?