Driver Heaven CSS benchmarks

[fsstrike]

I still have fun playing CSS on my GeForce3 Thought, I wish I had an x800 / 6800, Id be happy with either.

 

[CaiNaM]

Originally posted by: nitromullet

firingsquad does a much more complete analysis. Also, they have a MUCH faster cpu which seems to add ~10fps to each card at 1600 4xAA/8xAF. EXCEPT the x800pro, which seems to gain ~20fps.

it might mean that the x800s are FAR more cpu limited

frankly, i'm surprised with all the whining i see from the nv camp; we've all know for a long time nv > ati in doom3 and nv < ati in hl2. why is everyone so surprised by this?

I don't think that anyone is expexcting nV to beat ATi in HL2... The thing here is that there are some peculiar inconsistencies with this benchmark that were simply not present with the Doom3 tests. nV was faster across the board in Doom3 and evey website's results were roughly the same. With this test, we see completely different results wich are not scaling with different cpu's and are inconsistent with each other. IMO, the benchmark graphs from gamers-depot and FS should , at a glance, look somewhat the same. However, what you see is the that the 6800 Ultra solidly beats the X800 Pro at gamers-depot and it loses out at FS. At gamers-depot you also see the GT beating the Pro (slightly) in every benchmanrk. ...I really don't think that those extra 20MHz on the BFG 6800 GT OC are going to make that big of a difference. Then to top it all off, you get these wacky results from driver heaven that show the Ultra pushing 20 FPS at the same benchmark the X800 XTPE is pushing over 60!?! There are people on this forum that are getting (much) better FPS than that in the same benchmark with a slower cpu.

Obvioulsy, the XT PE is the card for HL2. That can not be disputed. However, the current benchmarks are not really conclusive about the performance of the other cards tested IMO... I guess I'll wait til AT benches these cards.

well, first of all, if you trust anything from GD.... anyone's whose "hardware" site looks worse than the average personal homepage does not instill confidence from me ;p also keep in mind one is using an intel platform and the other amd, which could account for some differences.

secondly, i wouldn't put too much into that 1 DH chart - it's just something ati fanboys use even tho it's comparing 2 different quality settings. stick to 4xAA/8xAF as that's the common denominator, not to mention the most common realworld settings. I also tend to put more credence in FS, as they've shown to be the most consistent (i think DH has has too much of an ati slant for some time). again, i'm not at all surprised ati has the edge here, nor was i when nv beat ati in doom, but doom played just fine on my x800pro, as i'm sure hl2 will play just fine on nv40 parts.

while i agree one should question some of these results, the passion at which it's either been defended or attacked by some people is really a sad indication of their self-worth or one's priorites in life

 

[Shamrock]

I know that this test can call for Tri-linear/Bi-linear filtering, but when it does, do the ATI drivers still use their "adaptive algorithm"? or does it use TRUE Tri-?

 

[BFG10K]

It could well be that ATi will be faster in HL2 but all of the leaked HL2 benchmarks I've seen so far show the 6800 series edging the X800 series, so I'll be waiting for official benchmarks before jumping to any conclusions about the matter.

 

[JBT]

Originally posted by: BFG10K
It could well be that ATi will be faster in HL2 but all of the leaked HL2 benchmark I've seen so far show the 6800 series edging the X800 series, so I'll be waiting for official benchmarks before jumping to any conclusions about the matter.

Agree 100% here. I'll wait till HL2 is out... Either way I'm sure it will play just fine on my rig.

 

[BFG10K]

...means 6xAA-16xAF on the Radeon and 8xSAA-16xAF on the Geforce...Again, at these enhanced settings the difference is huge.

Yeah, no sh!t.

 

[GeneralGrievous]

I think some of u havent noticed, but when the 6800U was using its max settings, it was using 8x supersampling, which we all know is poo, and it degrades performance totally for the 6800U, because its too much for it, and the fact that it wont run at 6x aa because it doesnt have that level to switch too, so its either 4x or 8x, so thats why in most other benchmarks u see it only going up to 4x, but FS does do the full 6x whack for the ATi tho

I don't think those benches are run in 8xAA mode. If they were the framerates would be in the 20s. Most likely they were done at 4x/16x.

 

[ronnn]

Originally posted by: Rage187

You only proved you were trolling.

You call that trolling? Not even any name calling, or references to shader days. I predict the real trolls will be here soon.



Edit: my bad they are already here. :disgust:

 

[Optimummind]

In my personal experience, that reviewer from Driverheaven is on some kind of crusade against nVIDIA. In the past, I found numerous shortcomings and insights on his part when it comes to hardware.

He tends to make news of all nVIDIA related shortcomings even before enough evidence is materialized (like the nVIDIA/Far Cry bashing he wrote up once). But when it comes to any situations that put ATI in a bad or lesser light, he doesn't write on those. For instance, where is the Doom 3 review using both cards but pretty much all major sites have already covered??

Also, that site is full of ATI fanatics which is why I don't visit there anymore except for the Driver Cleaner program.

 

[BFG10K]

I don't think those benches are run in 8xAA mode.

There was a set that was.

 

[ronnn]

DH was not being deceptive about running in 8 X AA mode. They just stated that in their opinion the
6 X AA gave the best image quality at playable frame rates. If you guys don't like DH, than don't visit. If you are lucky enough to own a 6800 ultra or gt, don't waste your time whining about DH - waste of a very good card.

 

[Optimummind]

Originally posted by: ronnn
If you are lucky enough to own a 6800 ultra or gt, don't waste your time whining about DH - waste of a very good card.

Sorry, I don't get it.

 

[ronnn]

Just saying the gt is a damm good card, no matter what DH says. Also would say the same thing about the x800, is a damm good card and at this point I would take either - if they were available at a realistic price in Canada. Besides I will always defend DH....... nice people there and you don't get flamed for the card you own if you ask a question. They have separate forums for Nvidia and ATI, keeps the rabble apart. And like most they do have their own bias.

 

[GeneralGrievous]

There was a set that was.

Sorry, I meant the 1600x1200 MAX bench at firingsquad.

 

[Regs]

I'm a little skeptical about those benches considering all other sites stating much better performance for Nvidia. Either way, I'll be glad to take 80 FPS with my GT 1600x1200 High Quality on a DX9 game any day. No one losses for whatever card they pick this year. And hopefully a driver update will fix AA/AF performance for the NV40, even though I hardly find a need for it at 16x12.

I'm just a little thrown off about how the x800 gets clobbered with AA/AF compared to the big brother, the x800PE. Must be the memory and pipes.

 

[Regs]

edit