• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Nvidia responds to the Det 40 drivers performance

X

Xernex

Senior member
Looks like we may have an explanation for why 3dmark scores jump and not game performance. Here is what they had to say.

"A lot of the improvements in performance are due to improved efficiency in the driver code that effect vertex and pixel shaders. That is why the jump in Nature. A big jump can be seen in Aquanox, too, another app that uses vertex shaders and pixel shaders. Apps that use pixel and vertex shaders also just happen to be the ones that are popular for benchmarking. Games that do not use vertex and pixel shaders may not see as significant of an improvement as games that do."

Brian Burke
NVIDIA Corp.

Source http://www.3dchipset.com/news.shtml#newsitem1026190468,5309

This explains why we haven't seen performance increases in games some of you have been benchmarking. It also explains why cards that don?t support things, such as the nature scene(my gf2 gts doesn?t support it) haven?t seen an improvement but Geforce 3s and 4s have.
 
There was also more in Brian's statement, which explains why those with GF4's had slow-downs especially when fiddling with the anisotropic settings:

The default setting on the new driver and the old driver are both the same. The problem that is being identified is a bug that occurs when the AF settings are changed. It seems that once the AF slider is moved away from 0 to another setting, then back to 0 again, it triggers a bug that causes the control panel to misread the register setting. That setting is not point sampling, but it is something less than 0. The only way to "reset" back to true 0 is by reinstalling the driver or deleting the registry setting.

We are going to fix this bug and others and submit the new BETA driver to WHQL, before the driver moves from BETA to "official".
Click to expand...

So that may explain some of the weirdness people have been experiencing. These drivers have been great on my GF3 Ti200.

Greg
 
Viable reasoning, and what I imagine most of us have already concluded. For the most part even games that do utilize vertex/pixel shaders heavily don't seem to benefit much however...
Morrowind and Comanche come to mind as two games that utilize shaders, and neither seem to benefit. Indeed, Morrowind seems to exhbity increasingly reduced performance relative to prior driver revisions as resolution increases.

Still any gains are good, even if they arent quite the well rounded boosts many were expecting

I wish he'd have commented on the rampant complaints of reduced image quality in many populars games however, and the seemingly darker gamma under OpenGL.
 
Sounds nice and all, but in the end the charts are still showing 3dMark2001 faster, QuakeIII (still the world's most popular benchmark game) faster and everything else the same or slower.

I'm not buying it so far.
 
I still can?t see how you can think there is still something fishy, it explains it perfectly, I think. Nvidia would have learned from ATI's stunt with Q3 last year, not to try and cheat drivers, because people WILL work it out. Nvidia would not want to risk loosing face like that, especially when they have the 9700 to contend with, until NV30 is out.
 
Originally posted by: Xernex
I still can?t see how you can think there is still something fishy, it explains it perfectly, I think. Nvidia would have learned from ATI's stunt with Q3 last year, not to try and cheat drivers, because people WILL work it out. Nvidia would not want to risk loosing face like that, especially when they have the 9700 to contend with, until NV30 is out.
Click to expand...

Well, as I recall, ATI "explained" the quack incident quite well, too. The fact is, its only faster in benchmarks, and many people have complained of worsened IQ. Also, the fact that NVIDIA has *officially* released buggy beta drivers for the first time in ages, and possibly the first time ever.
 
cough* Infinite Loop *cough*
Click to expand...

Have not had that since 12.11 drivers ,Pete behave yourself or I`ll come round and sabotage your ATi card 😉.

So far 40.41 drivers are not bad for me on stability,only had one game that does not like 2x anisoptropic filtering in (D3D mode) .
 
Originally posted by: Ben50
Morrowind and comanche probably don't see any improvement because they are limited by cpu speed.
Click to expand...

Comanche I'd agree to some degree as it tends to be pre-dominantly processor dependent and hasnt shown itself to benefit substantially from hardware vertex shading in the past, though I'd have hoped for at least a minor boost. Morrowind is strongly graphics card limited however, even with the most powerful boards available, yet it actually drops in performance with the latest drivers and by a rather large degree at higher resolutions.


I still can?t see how you can think there is still something fishy, it explains it perfectly, I think.
Click to expand...

I don't think they cheated at all, I just think the drivers were massively over-hyped by the public that expected huge performance increases.
I think nVidia would have been better served naming these the 31.XX drivers rather then the 40.XX as the improvements arent exactly substantial.

As for whether nVidia would cheat, well I still find the issues of a few months ago with nVidia boards dropping in performance when the splash screens were disabled in 3DMark to be dubious. Some underhanded benchmark specific optimizations were likely going on there.

I think ATi's Quake/Quack issue and the public reaction to that has ensured vendors will be a little more circumspect in the future however.
As long as we don't see a repeat of the 3DWinBench 98 days in which nVidia began the cheating, and ATi, 3dfx, Matrox and everyone else quickly followed suit.
 
DAOC hates the 40.xx drivers, i cant enable texture sharpening or Aniso (any seeting but 0x) without getting severe texture corruption
 
Originally posted by: Mem
cough* Infinite Loop *cough*
Click to expand...

Have not had that since 12.11 drivers ,Pete behave yourself or I`ll come round and sabotage your ATi card 😉.

So far 40.41 drivers are not bad for me on stability,only had one game that does not like 2x anisoptropic filtering in (D3D mode) .
Click to expand...


😛
 
I had the infinite loop problem for about 3 months. KT266a chipset, geforce 3ti200...blue screens....blue screens...under xp...idling at the desktop....bluescreens....opengl....reboot....
believe me. It's no myth.
 
Originally posted by: Xernex
Looks like we may have an explanation for why 3dmark scores jump and not game performance. Here is what they had to say.

"A lot of the improvements in performance are due to improved efficiency in the driver code that effect vertex and pixel shaders. That is why the jump in Nature. A big jump can be seen in Aquanox, too, another app that uses vertex shaders and pixel shaders. Apps that use pixel and vertex shaders also just happen to be the ones that are popular for benchmarking. Games that do not use vertex and pixel shaders may not see as significant of an improvement as games that do."

Brian Burke
NVIDIA Corp.

Source http://www.3dchipset.com/news.shtml#newsitem1026190468,5309

This explains why we haven't seen performance increases in games some of you have been benchmarking. It also explains why cards that don?t support things, such as the nature scene(my gf2 gts doesn?t support it) haven?t seen an improvement but Geforce 3s and 4s have.
Click to expand...

So, did you really expect Nvidia to tell the truth; that they had to do something to make it appear the GF4 was still a "competitor" to the 9700? That they feared their customers "jumping ship" if the 3DMark2001 scores were too lopsided in ATI's favor?

rolleye.gif


. . . pure marketing-driven decision to rush-release the beta Det 40s. 😉

EDIT: From AcesHardware's front page . . . Interestingly . . .
Interestingly, these drivers also seem to be a platform for emulating the upcoming NV30. From this page, you can download a program to enable NV30 emulation in the new drivers, as well as download the NV30 OpenGL Specification document and accompanying slides in PDF format.
Click to expand...
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top