Rename fear.exe and gain 30% on X1000 cards.

[Ubermann]

Info can be found at Beyond3D

1600x1200
everything high w/ volumetric lights on medium
X4AA X8AF
timedemo avg

fear.exe 27fps
fear1.exe 37fps

Seems like current drivers has a bug in AI Rage3D

[perkam]

Oh my...a conspiracy in the works ??? Wonder which side is the culprit...

Someone should try this.

Perkam

[Turok]

hmmm...

1) If a CrossFire, SIS or any other mobo that's not nVidia gives more performance on ATi cards, it could be true that nVidia's NForce drivers reduce performance on ATi cards

2) If it happens only with F.E.A.R., and there is no improvement on nVidia by tweaking F.E.A.R., could it be that nVidia paid Monolith to tweak the game so it loses performance on ATi cards? and could be the reason why nVidia ran better on the demos compared to ATi? Back of my F.E.A.R. retail box has the nVidia logo, so they are sponsoring this game.

3) It could also be that we have yet to see the full performance of the x1xxx, and future ATi drivers will show a huge boost in performance to catch nVidia by surprise when they release the GTX 512mb. It doesnt look likely, tho :P:

I think #2 is the most probable situation



BTW, is there a tweak for the x8xx and older series too?

[kromosto]

2. is the impossible situation because if nvidia paid monolith to tweak fear for not to perform good at x1xxx series ati cards then this performance gain cant happen with changing the name of the game, it can be done by changing the name of the card. because running exe is same so the code determines the ati card and making performance worse is still runnig there it is not changed. i think it is a driver issue.

[perkam]

Just read all three pages of Beyond3d, this "fix" works on all cards dating back to the 9xxx series, so this'll be a pretty big release a little while from now.

I'm thinking that with the things the 4.12 cats did, this'll most likely be the 5.12 cats.

Perkam

[DilTech]

This has nothing to do with NVidia. For NVidia to be the problem, NVidia drivers would be required, as LOE said.

If it was monolith, then changing the filename wouldn't make a difference, it'd perform bad unless you use 3danalyze and told it you have a 7 series instead of a x1000. Remember HL2 and FX cards?

It'd be ATi's mix up, sounds like they simply put in a wrong driver optimizations for when fear is ran, and it ran better without said optimization...

[Ubermann]

Ati replied in thread:
"It’s a "bug" in that we simply got an IF statement backwards Thanks for pointing this out to us, you have just helped us get a big performance gain. Of course there is no difference in the rendering, it’s just a CATALYST AI game specific optimization that was good for the demo version, but backfired in the final version. We will get it sorted out in a future Catalyst (not 5.11 which is being posted tomorrow by the way)."

[trakslacker]

^ yeah, just read that over at B3D. Looks like the 5.12's will indeed have some nice little improvements for us.

[Mr. Tinker]

Ati replied in thread:
"It’s a "bug" in that we simply got an IF statement backwards Thanks for pointing this out to us, you have just helped us get a big performance gain. Of course there is no difference in the rendering, it’s just a CATALYST AI game specific optimization that was good for the demo version, but backfired in the final version. We will get it sorted out in a future Catalyst (not 5.11 which is being posted tomorrow by the way)."

DilTech called it. All hail DilTech.

[DilTech]

DilTech called it. All hail DilTech.

[Cybercat]

Ati replied in thread:
"Thanks for pointing this out to us, you have just helped us get a big performance gain."

anybody else find that statement funny?

[Ubermann]

ati has more issues in drivers than usefull stuff rushed into demo optimizations, then forgot to double check it.. sounds lame to me

then why don't change it in the 5.11? After all all they need is to edit the fear profile, won't take more than 5 mins.... I am missing the point...

Maybe because the 5.11 are already completed.

Cybercat It was funny, nothing i would expect from a catalyst programmer =)

[cooter]

ati has more issues in drivers than usefull stuff rushed into demo optimizations, then forgot to double check it.. sounds lame to me

then why don't change it in the 5.11? After all all they need is to edit the fear profile, won't take more than 5 mins.... I am missing the point...

Another ATI bashing comment. To be honest with you i have had more issues with my Nvidia card then my past ATI cards. It does not sound like this caused any problems with the drivers, it simply did not get the performance that could be had. I do not look at that as an issue.

[Entity_Razer]

10FPS is quite the mixup. Good PR right now but I reckon they did loose some clients over this....

[DilTech]

10fps is definitely an oopsey, nice to see it's being fixed though.

As for the complaints about drivers, BOTH companies have had their fair share of driver issues..

NVidia, remember the temp bug causing temps to read 10C higher than they were in idle?

ATi, remember the bug originally with the X800's where it only used half of it's ram?

This is commonplace, there's just so much to do in a driver set that mistakes are easily made, so lets keep the "Oh ****'s drivers are T3h Suck" to yourselves...

[Turok]

Ati replied in thread:
"Thanks for pointing this out to us, you have just helped us get a big performance gain."

anybody else find that statement funny?

What I find funny about it is that ATi respondes to this bug so fast, and dont notice that they have'nt fixed the low perforamnce on OpenGL games.
I would sure like a 20% boost in Quake4.
Someone should remind ATi that most ATi users dont have a x1xxx series and should release a OpenGL fix for the older cards too.

[Starscream]

@Turok i wouldnt trust onit.
when the Xxxx series came people with 9xxx cards were left in the dark.
like when a boost came for the x800 the older 9xxx didnt get a boost.
if a fix was made for the Xxxx cards in the newest driver the old 9xxx had to wait 2 more driver versiosn for that fix.

so i wouldnt expect a opengl boost for older ATI cards.

[Revv23]

plus it sounds like most of the open gl gains were made because of the new memcontroller