John Carmack Responds to the ATI Humas tweek

Status
Not open for further replies.
Joined
Aug 9, 2004
Messages
22
http://www.beyond3d.com/forum/viewtopic.php?t=14874&postdays=0&postorder=asc&start=220

I got a response from Carmack:

Quote:
Our specular function isn't a POW function, it matches the bias / scale calculations done on the NV10/NV20/R200 paths.

This of course explains why it doesn't look exactly the same. This is also a good thing. Now we don't even need to use a POW function, but we'll be fine with a MAD_SAT, which make it even faster. Got 21% boost in the timedemo with this. Now I still don't know exactly what exponent he's approximating, or if he's using different lookup tables for different materials, but I've tried this and it seems to look the same, but our fellow artifact hunters may want to verify that.

Code:
MAD_SAT R1, specular, 4.0, -3.0;

If that causes artifacts for someone, maybe approximating a lower exponent will do it:

Code:
MAD_SAT R1, specular, 3.0, -2.0;

------------------------------------------------------------------------------------------------------

http://www.beyond3d.com/forum/viewtopic.php?t=14874&postdays=0&postorder=asc&start=220

I've tuned it a bit more to try to get exactly the same output as the original shader, and come to the conclusion that this is pretty darn close even when comparing screenshots.

Code:
MAD_SAT R1, specular, 5.0, -4.0;

So I've put that into the tweak file:
http://esprit.campus.luth.se/~humus/temp/doom3PerformanceTweak.rar

Comparison shots:

With tweak:
http://esprit.campus.luth.se/~humus/temp/shot00017.jpg
Without tweak:
http://esprit.campus.luth.se/~humus/temp/shot00018.jpg

High-res .png files (1MB+):
http://esprit.campus.luth.se/~humus/temp/shot00017.png
http://esprit.campus.luth.se/~humus/temp/shot00018.png
 
im anxious to see what was in the "huge" e-mail that carmack sent to reverend. it will interesting how this turns out
 
they look pretty similar, but the humus tweak has a darker flashlight and less shiny metal.
 
I just finished trying this tweek on my system.
I got 7-8 FPS boot in the timedemo.

Nvidia A7N8X
Athlon 3000XP
Radeon 9800 Pro
1G Ram
 
and hopefully empty and burninggrave wont find it
Nup but tranCendenZ found it :p Like bulldogs to a packed lunch :DJ/K

But has anybody From here had a chance to try it yet?To verify the 10 to 11 FPS boost.I doubt ATI are gonna find much in Doom3.to optimized for The NV architecture.
 
@trapine said:
Nup but tranCendenZ found it :p Like bulldogs to a packed lunch :DJ/K

But has anybody From here had a chance to try it yet?To verify the 10 to 11 FPS boost.I doubt ATI are gonna find much in Doom3.to optimized for The NV architecture.

From what I've seen it varies from set-up to set-up. Generally, people seem to get a higher boost with AF enabled. When it's not artifacting of course.
 
From what I've seen it varies from set-up to set-up. Generally, people seem to get a higher boost with AF enabled. When it's not artifacting of course

There was talk about the artifacting could be because of most people having there X800Pros overclocked?Or does it also do it on X800XT?
 
The Batman said:
From what I've seen it varies from set-up to set-up. Generally, people seem to get a higher boost with AF enabled. When it's not artifacting of course.

so you're saying that if there is artifacting there isn't a higher boost with AF enabled?

seriously, what was the point of adding that last inflamatory statement? was it really needed? did you type it out of anger/jealousy/fear/!!!!!!sm? it seems totally unnecessary to me, i don't see how it added anything useful to this thread at all, except for more animosity between the two sides which is something this forum can surely do without.
 
Xtasy0 said:
seriously, what was the point of adding that last inflamatory statement?

For some people [in fact if you bothered to check the beyond3d thread] there were noticable problems, and not soley of the artifacting kind. Would you rather I say nothing about it and pretend all is preachy in wonderland. I imagine you would. Well you'll have to forgive me but I'm not being paid by ATIs PR Team. I'm going to give it to people straight, the good with the bad. Yes it helps performance, espeically with AF enabled. Yes some people have experienced problems with the hack.

was it really needed?

Yes.

did you type it out of anger/jealousy/fear/!!!!!!sm? it seems totally unnecessary to me,

It felt unneccessary to you? Some people feel that the Earth is still flat as well. It's a good thing I pay these people no heed. Next time I ask for your approval on anything...go look outside becuase Jesus will be singing Stairway to Heaven in a tutu.

i don't see how it added anything useful to this thread at all, except for more animosity between the two sides which is something this forum can surely do without.

Your delusions of a perceived forum war are of litte consequence to me. If you can't handle seeing both sides of the coin stated, I suggest you smash all the mirrors in your house. After all, they force you to see things as how they are, not how you would have them be.
 
The Batman said:
For some people [in fact if you bothered to check the beyond3d thread] there were noticable problems, and not soley of the artifacting kind. Would you rather I say nothing about it and pretend all is preachy in wonderland. I imagine you would. Well you'll have to forgive me but I'm not being paid by ATIs PR Team. I'm going to give it to people straight, the good with the bad. Yes it helps performance, espeically with AF enabled. Yes some people have experienced problems with the hack.



Yes.



It felt unneccessary to you? Some people feel that the Earth is still flat as well. It's a good thing I pay these people no heed. Next time I ask for your approval on anything...go look outside becuase Jesus will be singing Stairway to Heaven in a tutu.



Your delusions of a perceived forum war are of litte consequence to me. If you can't handle seeing both sides of the coin stated, I suggest you smash all the mirrors in your house. After all, they force you to see things as how they are, not how you would have them be.

the person you were replying to was asking for people who had personal experience with the hack, he was looking to verify the performance gain, your hearsay verifies shitall.

he obviously read the thread on b3d (or the thread here that linked to b3d and was trolled by people (who mostly owned cards created by a certain company) bitching about the artifacting) and most likely knew about the artifacting from said threads, But even if he didn't know about the artifacting he didn't mention that he was going to try the hack and there wasn't anything in his post suggesting that he would need to be informed of the artifacting. now assuming he said "hey guys im thinking about trying out the hack, can anyone here verify that they've had a performance increase? oh and whats the deal with the artifacting" then perhaps it would have been appropriate. but the way you stated it was totally unneccessary, especially in context of the post you were replying to.
 
Xtasy0 said:
so you're saying that if there is artifacting there isn't a higher boost with AF enabled?

seriously, what was the point of adding that last inflamatory statement? was it really needed? did you type it out of anger/jealousy/fear/!!!!!!sm? it seems totally unnecessary to me, i don't see how it added anything useful to this thread at all, except for more animosity between the two sides which is something this forum can surely do without.
What inflammatory statement sre you seeing in that post? Maybe you should e-mail those guys at beyond3d and tell them to keep quiet about possible artifacting. :rolleyes:
 
Xtasy0 said:
the person you were replying to was asking for people who had personal experience with the hack, he was looking to verify the performance gain, your hearsay verifies shitall.

he obviously read the thread on b3d (or the thread here that linked to b3d and was trolled by people (who mostly owned cards created by a certain company) bitching about the artifacting) and most likely knew about the artifacting from said threads, But even if he didn't know about the artifacting he didn't mention that he was going to try the hack and there wasn't anything in his post suggesting that he would need to be informed of the artifacting. now assuming he said "hey guys im thinking about trying out the hack, can anyone here verify that they've had a performance increase? oh and whats the deal with the artifacting" then perhaps it would have been appropriate. but the way you stated it was totally unneccessary, especially in context of the post you were replying to.

Righto then. In the future I shall run all my posts by you first for approval.
:rolleyes:
 
joemama said:
What inflammatory statement sre you seeing in that post? Maybe you should e-mail those guys at beyond3d and tell them to keep quiet about possible artifacting. :rolleyes:

even our special buddy the batman knew what i was talking about. don't tell me i need to make up illustrations and flash cards to explain it to you.

BTW the batman in his haste to create an inflamatory post and start an argument didn't even try to further help our good friend trapine when he askes a legitimate question about the artifacting. c`mon the batman, now is not the time to pretend all is peachy, someone needs your help!

@trapine said:
There was talk about the artifacting could be because of most people having there X800Pros overclocked?Or does it also do it on X800XT?
 
Xtasy0 said:
even our special buddy the batman knew what i was talking about. don't tell me i need to make up illustrations and flash cards to explain it to you.

BTW the batman in his haste to create an inflamatory post and start an argument didn't even try to further help our good friend trapine when he askes a legitimate question about the artifacting. c`mon the batman, now is not the time to pretend all is peachy, someone needs your help!

There was nothing inflamatory about my post. I'm sorry you see it as such.

And I did answer trapine's question. I don't see you doing anything productive.
 
The Batman said:
There was nothing inflamatory about my post. I'm sorry you see it as such.

And I did answer trapine's question. I don't see you doing anything productive.

Actually, you Didn't answer his question, at least not anywhere in this thread, not as far as i can see anyways, perhaps you were in such haste to argue you did read his question and think about replying, but the fact is (as anyone can see by looking up the page) you didn't answer his question specifically about the artifacting in x800 pro's and x800xt's.

see here is his question (which i'm trying to help him get an answer to):
@trapine said:
There was talk about the artifacting could be because of most people having there X800Pros overclocked?Or does it also do it on X800XT?

and where is your reply?....NOWHERE! you have let this forum down the batman, you should feel ashamed.
 
Newsflash: Artifacting can also occur with Nvidia cards when oc'ed too high or using modified game files.

OMFG!1 I just pissed off everyone who now or in the past has owned a Nvidia card and made their e-pen0s 1.3 inches shorter. :eek: :p
 
joemama said:
Newsflash: Artifacting can also occur with Nvidia cards when oc'ed too high or using modified game files.

OMFG!1 I just pissed off everyone who now or in the past has owned a Nvidia card and made their e-pen0s 1.3 inches shorter. :eek: :p

it wasn't WHAT he said it was HOW he said it and how it had nothing to do with the question he was apparently answering.

i dont care that the artifacting is being talked about, but he clearly didn't need to mention it in his post, especially not in the manner he did.
 
the other thread got out of hand. you can discuss the hack/tweak, but don't turn this into an ati/nvidia flame war or this one is getting locked down also.
 
fugu said:
the other thread got out of hand. you can discuss the hack/tweak, but don't turn this into an ati/nvidia flame war or this one is getting locked down also.

I second that.

Take heed, you get out of hand and Fugu will knock your ass down...

...then I'll kick ya!

:D
 
Please lets get back on topic. This is interesting.

Anybody else have results? I'm wondering if this will help my 9800pro.
 
Milenko said:
Please lets get back on topic. This is interesting.

Anybody else have results? I'm wondering if this will help my 9800pro.

Yeah, I want to see if others get better results (just curious, but I lack an r300 to test it out :().

Quick guide for those who don't want to read the full thread.

Download this file:
http://esprit.campus.luth.se/~humus/temp/doom3PerformanceTweak.rar

Extract so that the shader file goes under doom3\base\glprogs. This replaces a dependent texture read with equivalent math, which runs better on ATI cards, but seems to run slower on NV boards, so only apply this if you got an ATI card
 
Handed this tweak over to a friend of mine with very similar specs to myself sans the videocard (9800pro for him).

800x600 HQ = 54fps
800x600 HQ w/ tweak = 66fps

Hell of a jump, and he didn't notice any IQ degredation.
 
If you guys will go through the Beyond3D thread you will see that MOST people saw very little gain from the HACK. Some were lucky to get 2 fps increase.

There were several with X800XT PE's and they saw only marginal 2-5 fps gains as well. Some didn't even see that much.

Then you had your special cases where someone saw like a 10 fps increase or something.

Carmack will never endorse this sort of feature in an official patch because of the fact it changes the IQ for extremely minimal performance gains. It also hasn't been thoroughly tested to see what other effects it has.

I'm sure Carmack didn't appreciate Humus publicly screwing around with his code ither. Humus works for ATI and he should of contacted Carmack privately.

Lets not forget that it was also an ATI employee that leaked the DOOM 3 Alpha.
 
Jesus, let's at least post some truths here, shall we?

First, there is NO artifacting with this particular hack. There was with the original one, but it's been changed and tuned at least three times now. For the record, I never saw any artifacting, and the people that did offered no information about whether their machines were overclocked or not.

Second, there is NO loss in image quality. There was with the original one, but again, it's been changed and tuned. Further more, there is no longer even a DIFFERENCE in image quality. Several people have posted screenshots now that have no differences in image quality at all.

Lastly, this doesn't prevent you from playing multiplayer games if you do this properly (instead of changing the original file, just dropping it in base/glprogs). In multiplayer games, DOOM 3 will simply use the stock shader.

I'm surprised to see this thread sliding back into arguments already proven false in the other thread. Come on guys, why don't you stick with things that are TRUE, like how the nVidia hardware is still faster, if you want to pick a fight. This also has absolutely nothing to do with whether or not Carmack or ATI "approve" of this hack or not. The game was designed to be modifyable, and if they didn't want people messing with the shader code, they'd have locked it down, ot complied it into a file.

I'm sure Carmack didn't appreciate Humus publicly screwing around with his code ither. Humus works for ATI and he should of contacted Carmack privately.


There is absolutely nothing WRONG with how this was done, and Carmack has already gone on record as saying that he has no problem with this. There are more comments to come, but that we already know for sure.
 
Well, its good for singleplayer, which is supposed to take an average player 24 hours to play.

You can always rem out the line again when/if you decide to multiplayer, if the tweak affects it adversly for you (thats only an if)
 
Rob, just ignore him :p. Anyway, anyone trying the new patch with older ati hardware noticed any difference? I tried it and I saw a 3fps jump (38 -> 41fps) on my card (which is in my signature).

Edit: While messing around with the settings, it dawned upon me that this totally wouldn't help me at all since I'm not running the arb2 pathway. Kinda odd to see run differences jump so much though :(.
 
tranCendenZ said:
they look pretty similar, but the humus tweak has a darker flashlight and less shiny metal.
Just turn off specular lighting for that....
 
SuperRob said:
There is absolutely nothing WRONG with how this was done, and Carmack has already gone on record as saying that he has no problem with this. There are more comments to come, but that we already know for sure.

How many nVidia employee's do see you openly posting HACKS to game codes on forums?

How many nVidia employee's do you see openly posting information on upcoming performance increases?

NONE

Its all done through a private channel and professionally instead of out in the open on the forums. It would be one thing if Humus didn't work for ATI but he does.

And where is Carmacks response saying he has no problem with this?
 
burningrave101 said:
Carmack will never endorse this sort of feature in an official patch because of the fact it changes the IQ for extremely minimal performance gains. It also hasn't been thoroughly tested to see what other effects it has.
First thing, extremely minimal performance gains? Sure some people are getting small FPS increases but this is the HardForums, any kind of gain PERIOD is welcomed. Give me a break.

You're right that it might not be in a patch, but who cares? Other effects? Sure there might be some but a lot of people here LOVE to get really high FPS' so obviously this tweak is extremely welcomed here, and the mininal 'other' effects are irrelevent. A tweak is a tweak is a tweak. There are going to be pro's an cons to every tweak, doesn't mean it's not legitiment and worthy.

Sure, it might not help you, might not boost the FPS for an nVidia card, and might make the ATi vs nVidia race a close call with Doom 3, so of course you're negative about it!

Carmack getting pissed? Why would he care? Sure Humus works for ATi but all that says is he has the knowledge to legitimatly back up his tweak! He should contact Carmack first? Why the hell...he's a gamer just like the rest of us, should we contact Carmack if we find a tweak?

Be realistic.
 
I wish Reverend would hurry up and post the "mini-interview". :mad:
 
BakedGoods said:
Sure, it might not help you, might not boost the FPS for an nVidia card, and might make the ATi vs nVidia race a close call with Doom 3, so of course you're negative about it! A tweak is a tweak is a tweak. There are going to be pro's an cons to every tweak, doesn't mean it's not legitiment and worthy.

I'm not negative about it because of the nVidia vs ATI race because of the fact a tweak like this can't be used in an nVidia vs ATI review. Its a HACK. HACKS And Tweaks are NOT used in reviews. In reviews games only use default in-game settings and official patches.

This is nothing more then a game tweak like the following:

Go into your doomconfig.cfg file in the base directory look for the following


seta image_cacheMegs ---I changed mine to 2000 which is 2gigs. This is a
hard drive cache not ram cache.

seta image_cacheMINK--I changed mine to 20480

seta image_usecache--- I have 1 to enable it.


This seems to work for me. I have an AMD 2500 oc to 3200 with a GF 6800 GT and I can play on Ultra Quality with no skips at all.

Only difference is that tweak has no possible negative effects.
 
burningrave101 said:
I'm not negative about it because of the nVidia vs ATI race because of the fact a tweak like this can't be used in an nVidia vs ATI review. Its a HACK. HACKS And Tweaks are NOT used in reviews. In reviews games only use default in-game settings and official patches.
Yeah I understand that, but if I, who has an X800XTPE ( lets pretend :p ) and you have a 6800Ultra yet on the same settings I have a better FPS because of the hack doesn't that mean the X800XTPE is better then the 6800Ultra? Anyway, I'm not going to have a ATi vs nVidia arguement here, I'm just saying it's a tweak that will boost FPS with ATi cards. Carmack made it possible to edit the game so tweaks like this can be made! Who found the tweak, where he works, and what he eats for breakfast are all irrelevent.

If the tweak boosts preformance, good. If it doesn't, turn it off. Simple.

EDIT: Okay, you want to call it a hack? Sure, let me call it a 'Dancing Mexican'. So like I said: "If the Dancing Mexican boosts preformance, good. If it doesn't, turn it off. Simple."

Whatever you call it doesn't change anything.
 
Has anyone checked out the "plasma" lighting yet? Supposedly takes a huge performance hit, but some people are going ga-ga over it.
 
I saw screenshots on that post in beyond3d where IQ was lower (white artifacts poping over the place). Even humus admits that it won't work properly unless Carmak builds it into his engine. If you guys want to install a haked up version of Doom III which give you about 1 - 10 fps more then more power to you, but as for me, I'm fine with playing Doom III at 1280x1024 (my monitors max res otherwise it would be 1600x1200) with 4xFSAA and 16aniso on my 6800GT with plenty of power to spare.
 
For the people who say the hack is "identical" to the original specular highlights, here are the differences from Humus's own uncompressed png high res images (scaled down by 50% to keep the file sizes small):

hackvsnohack.jpg

hackvsnohacknegative.jpg


It's definitely not "identical," but it is somewhat similar. IOW, the hack works around JC's intentions of keeping the look the same between all rendering paths.
 
jon_k said:
I saw screenshots on that post in beyond3d where IQ was lower (white artifacts poping over the place).

That was the first hack, which was a wrong assumption on Humus' part, and was corrected with the second version of it, which doesn't have the white dots problem
 
I think someone said the white pixels can appear on overclocked or overheating cards and/or at medium texture quality.

It would make sense that a more efficient driver tweak/hack may have the card running hotter than normal (especially in the cases where some people seem to be running 40 percent faster) that may inadvertantly be causing some problems.

Surprisingly enough, adding higher levels of AF does not seem to affect FPS at all, whereas at default there was a pretty big hit.
 
jon_k said:
I saw screenshots on that post in beyond3d where IQ was lower (white artifacts poping over the place). Even humus admits that it won't work properly unless Carmak builds it into his engine. If you guys want to install a haked up version of Doom III which give you about 1 - 10 fps more then more power to you, but as for me, I'm fine with playing Doom III at 1280x1024 (my monitors max res otherwise it would be 1600x1200) with 4xFSAA and 16aniso on my 6800GT with plenty of power to spare.

That is the older method, he's since improved it and no one seems to be reporting any artifacts with the newer method. The newer one also increases fps much more than the first one so people are seeing better gains. With the newer method the artifacts seem to be gone and the performance is up, why you want to deny this to the ati users who benefit from it is beyond me.
 
Status
Not open for further replies.
Back
Top