NVidia driver 337.50 beta released!

[tweakboy]

nVidia beta is like AMD final release.

 

[Grooveriding]

Anyone want to try them? http://www.geforce.com/drivers/results/74735

I personally stay away from nVidia beta drivers, having been burned several times in the past.

I went for a comparison in the game nvidia claimed the largest performance increase in :

Up to 64% in Total War: Rome II

Installed the latest WHQL drivers first for my first run and then used these new betas.

I got nothing performance wise. Nothing, zip, zilcho... no performance gains... Well .3 fps increase...

335.23 Driver

337.50 Driver

Nothing. Maybe the claimed performance gains come with the usual caveat that you have to run them at a specific resolution, with certain settings disabled (mine are all maxed), in a specific part of the game and while standing on one leg.

Would of expected to at least see some improvement with such huge gains claimed.

 

[FalseChristian]

I was gonna try 'em but only the Chinese version is available as of yet.

 

[Grooveriding]

I was gonna try 'em but only the Chinese version is available as of yet.

English only:

x32:
http://us.download.nvidia.com/Window...glish-beta.exe
x64:
http://us.download.nvidia.com/Window...glish-beta.exe

I found the official links at guru3d forums.

Users there are saying the same thing, no real gains as claimed. Some of them tested multiple games. I'm too lazy and just did the one. Must be very narrow situations where these huge gains are seen.

 

[sioux]

Does it improve some performance on GTX 660 ?
Thank you

 

[desprado]

So far i have benched only Sleeping dogs and Metro last light
Sleeping Dogs
335.23 121fps
337.50 144fps

Metro last light
335.23 54 fps
337.25 58.1 fps

 

[tviceman]

So far i have benched only Sleeping dogs and Metro last light
Sleeping Dogs
335.23 121fps
337.50 144fps

Metro last light
335.23 54 fps
337.25 58.1 fps

Decent gains there.

 

[tviceman]

English only:

x32:
http://us.download.nvidia.com/Window...glish-beta.exe
x64:
http://us.download.nvidia.com/Window...glish-beta.exe

I found the official links at guru3d forums.

Users there are saying the same thing, no real gains as claimed. Some of them tested multiple games. I'm too lazy and just did the one. Must be very narrow situations where these huge gains are seen.

Maybe your massive GPU overclock, for the most part, cancels out the CPU driver overhead?

 

[Grooveriding]

Maybe your massive GPU overclock, for the most part, cancels out the CPU driver overhead?

Eh, usually the more GPU power you give, the more CPU power you require to take full use of it. In the Valley benchmark I gain 5-10FPS for every 100Mhz I add over 4.6Ghz on my CPU.

70% gains claimed in Rome 2 and I got less than 1% of a gain...

I never expected these drivers to deliver what was claimed. It's just not realistic to expect those sorts of gains out of just a driver update. Even Mantle, which is an entirely distinct and custom API baked in on a game by game basis, only gives about 10% more performance on high end setups. There was no way a simple driver update was going to be able to do anything like that in games that have already received driver optimizations.

Maybe someone can try something like BF4 on all low settings @ 1080P on a powerful GPU setup and see if there are the kinds of 50%+ gains Mantle gives in those CPU limited situations.

The driver must do something for all the PR about it. Maybe it just needs to be tested in whatever specific scenarios it actually makes a difference under.

 

[BFG10K]

I was wondering what the hell was going on as I've been getting those pauses as well. I've been playing online lately and thought it was some sort of lag or problem with my router. I didn't think it could of been a driver issue.

Have you had any pauses with these new beta drivers?

 

[BFG10K]

A good write-up here: http://www.geforce.com/whats-new/articles/nvidia-geforce-337-50-beta-performance-driver

They list the settings used.

 

[tviceman]

Eh, usually the more GPU power you give, the more CPU power you require to take full use
The driver must do something for all the PR about it. Maybe it just needs to be tested in whatever specific scenarios it actually makes a difference under.

Yeah it's probably at 1080p low / medium settings where the most gains will be had - just like with Mantle BF4.

 

[parvadomus]

Nice PR stuff. Link Looks like the typical specific-game driver like always :awe:

 

[Carfax83]

Guru3d ran some benches and got a massive gain in Hitman Absolution and a decent increase Crysis 3 in SLI..

 

[96Firebird]

Groove, looks like Ryan Smith got a 70% boost in Rome 2 with his 780 Ti SLI setup, wonder why you can't get the same?

What are your reference drivers? Ryan is using the fairly old 331...

 

[BFG10K]

http://www.tomshardware.com/news/nvidia-geforce-337.50-driver-benchmarks,26473.html

 

[Spjut]

PCgameshardware.de had pretty disappointing results except in Crysis 3. They used a QX9650 and i7 3770k

http://www.pcgameshardware.de/Nvidi...-sGeforce-Treiber-33750-gegen-Mantle-1116527/

 

[Pandora's Box]

PCgameshardware.de had pretty disappointing results except in Crysis 3. They used a QX9650 and i7 3770k

http://www.pcgameshardware.de/Nvidi...-sGeforce-Treiber-33750-gegen-Mantle-1116527/

Oh look, typical Nvidia PR machine in full swing :awe:

 

[Grooveriding]

Groove, looks like Ryan Smith got a 70% boost in Rome 2 with his 780 Ti SLI setup, wonder why you can't get the same?

What are your reference drivers? Ryan is using the fairly old 331...

Look at the single card gains with a 780ti, non-existent like mine were. The 780ti SLI gain is because they just added an SLI profile to this driver, previously SLI didn't work in the game.

 

[Grooveriding]

Have you had any pauses with these new beta drivers?

Yes they seem to be fixed. I was getting the pauses in Diablo 3 about every 10-15 minutes. Game would just freeze, still had sound, then continue in a few seconds. Hasn't been happening with these drivers in the hour or so I played this morning.

 

[parvadomus]

Haha the QX9650 results shows how well it removed the CPU bottleneck in games. Its MILES away from Mantle, the Nvidia PR was at best disgusting this time.

 

[3DVagabond]

It's not often you see 10% improvement with single cards (Not in everything, but still...). Multi GPU sure, but not single. Some of the claimed gains understandably had people excited. It was supposed to be the Mantle killer. The hype and PR were just over the top, which is too bad because it detracts from what is an excellent performance driver.

 

[96Firebird]

Look at the single card gains with a 780ti, non-existent like mine were. The 780ti SLI gain is because they just added an SLI profile to this driver, previously SLI didn't work in the game.

Oh I see, so the results were only because an SLI profile were added... I just looked at your sig and saw 780 Ti SLI, so those are the results I looked at.

 

[AdamK47]

Rome II has improved quite a bit for me. All four Titans are being used during the forest benchmark.

Settings @ 2560x1600:

Results:

 

[el etro]

Seems that bring performance gains to all games, and in some cases hit the "up to" performance gains expectations.