Monday, April 7th 2014

NVIDIA to Launch GeForce 337.50 Beta Later Today

It's now NVIDIA's turn to come up with a 'wonder driver' that introduces large across-the-board performance improvements that could affect performance standings of GeForce GTX GPUs. Bearing the canonical version number GeForce 337.50 Beta, the driver will step up DirectX rendering performance thanks to some fundamental changes in the way the the driver and API handle graphics processing loads for the metal (GPU) to chew on. The difference that makes, according to a leaked NVIDIA press-deck, is that NVIDIA GPUs will offer higher frame-rates than Mantle, while staying in Direct3D mode. Put in numbers, NVIDIA is claiming performance improvements of up to 64% in Total War: Rome II, up to 21% in Star Swarm, and up to 10% in Metro: Last Light.
Sources: ComputerBase.de, VideoCardz
Add your own comment

48 Comments on NVIDIA to Launch GeForce 337.50 Beta Later Today

#1
burebista
I'm eagerly waiting for Wizz testing in house here at TPU.
Posted on Reply
#2
erixx
And these were not:
"335.23: 53,8 avg
337.50: 55,4 avg #smoother in battles, gained fps on campaign map"
This doesn't justify downloading it...
Feel free to get your free placebo pills!
Posted on Reply
#3
john_
So now that all those who where laughing at me defending the stupid and their stupidity have their new shiny drivers in their computers, tell me, DO YOU SEE WHAT NVIDIA WAS PROMISING?
...other than Rome of course where they fixed whatever problem they had there and Star Swarm where they seems to get excellent boost(maybe the driver is running Mantle code but we will never find out - yeah that's a stupid and crazy idea, you can laugh at that).

I just had a look at Anandtech and Toms Hardware and I don't see what the Nvidia slides where talking about. I might have a look at a few other sites and change my mind. But after a quick look I think you should go out and buy a 3930K at least and then keep laughing at me of course.
Posted on Reply
#4
W1zzard
by: burebista
I'm eagerly waiting for Wizz testing in house here at TPU.
I'm busy with the AMD side of things right now, but will try to get an article out for this driver after the AMD release.
Posted on Reply
#5
burebista
No problem. Take your time and do it right as usual. :toast:
Posted on Reply
#6
Hilux SSRG
by: burebista
Some user benchmarks.
Those BF4 benches look like a good improvement for a beta driver.
Posted on Reply
#7
HumanSmoke
by: john_
So now that all those who where laughing at me
Are still wondering why you started defending Mantle being optimized for CPU bound rather than GPU bound usage scenarios, then start bawling about an Nvidia driver (since you're an AMD board user I find hilarious) being shown in the same light.
Why do you expect an Nvidia driver to have wholesale improvements and live up to PR hype when Mantle is in the same boat?....in fact you seem hell bent on defending the latter.

BTW: Here are some benchmarks that are more CPU limited (Hitman and Crysis 3) versus some that aren't.

For me, if there is ANY performance or game quality gains to be had and a driver does not break compatibility with anything else I use, I'll use it.
You? Well, you should continue to not use an Nvidia driver with your not-Nvidia card.
Posted on Reply
#8
pr0n Inspector
by: john_
So now that all those who where laughing at me defending the stupid and their stupidity have their new shiny drivers in their computers, tell me, DO YOU SEE WHAT NVIDIA WAS PROMISING?
...other than Rome of course where they fixed whatever problem they had there and Star Swarm where they seems to get excellent boost(maybe the driver is running Mantle code but we will never find out - yeah that's a stupid and crazy idea, you can laugh at that).

I just had a look at Anandtech and Toms Hardware and I don't see what the Nvidia slides where talking about. I might have a look at a few other sites and change my mind. But after a quick look I think you should go out and buy a 3930K at least and then keep laughing at me of course.
"Mantle is good because some people won't pair a high end GPU with an appropriately high end CPU"

"Look! nVidia's attempt at solving the same problems but with only a driver update instead of a total rewrite of a game's render path doesn't give you massive boost in non-cpu-bound situations! Boo!"

Flawless logic strikes again.
Posted on Reply
#9
john_
by: HumanSmoke
Are still wondering why you started defending Mantle being optimized for CPU bound rather than GPU bound usage scenarios, then start bawling about an Nvidia driver (since you're an AMD board user I find hilarious) being shown in the same light.
Why do you expect an Nvidia driver to have wholesale improvements and live up to PR hype when Mantle is in the same boat?....in fact you seem hell bent on defending the latter.

BTW: Here are some benchmarks that are more CPU limited (Hitman and Crysis 3) versus some that aren't.

For me, if there is ANY performance or game quality gains to be had and a driver does not break compatibility with anything else I use, I'll use it.
You? Well, you should continue to not use an Nvidia driver with your not-Nvidia card.
LOL OK....
I am defending Mantle? Are you serious? Of course you are serious. Your fanboism just skyrocketed.
Nvidia was promising too much. They are using a 12 thread cpu and promise Mantle performance improvements in their big marketing campaign. 12 threads cpu. The perfect scenario where you don't need Mantle, where Mantle is useless.
Nvidia's driver is just like any other driver in the past. A few extra frames here, a few extra frames there and a couple of games where there are real performance improvements, like so many times before. Did you just bought a PC yesterday and these things are new for you? No you just play along with Nvidia's marketing campaign.

by: pr0n Inspector
"Mantle is good because some people won't pair a high end GPU with an appropriately high end CPU"

"Look! nVidia's attempt at solving the same problems but with only a driver update instead of a total rewrite of a game's render path doesn't give you massive boost in non-cpu-bound situations! Boo!"

Flawless logic strikes again.
And here comes the other green fanboy. Nvidia attempt in solving the same problem? Oh God.... Are you serious, of course you are like the other fanboy. CPU WITH 12 THREADS. Do you understand the number 12? They solved nothing. They did the typical optimizations like so many times before but they throw a huge campaign to create noise about how they made Mantle, or maybe I should say DirectX 12, unnecessary. And of course you are happy to believe them.
Funny...
Posted on Reply
#10
Fluffmeister
I gained about 58% performance running Star Swarm with this driver, on my ageing stock i7 920.
Posted on Reply
#13
HumanSmoke
by: john_
Remove the green glasses and check again. Then compare with Nvidia's charts.
So, you're having a big sad because Nvidia's slides show a best case scenario?
But aren't you the one who preferred best case scenario to real world usage? From the thread regarding AMD's Mantle press slide claims
by: john_
The key word here is "45%". It is way too big even for a marketing lie. I prefer an "up to 45%" in a title like BF4, than a guaranteed "15%", anytime.
:SMH:
Posted on Reply
#14
pr0n Inspector
by: john_
Remove the green glasses and check again. Then compare with Nvidia's charts.
No, you pull your head out of the sand, take off your welding goggles and look at how it improved performance for a guy with Q6600/GTX 670 and massively improved performance for SLI setups.

And where did you get the idea that it will provide massive boost for single-gpu at 2560 resolution? Despite what you've imagined in your head, nvidia's own benchmarks are not insane at all.
Posted on Reply
#15
john_
by: HumanSmoke
So, you're having a big sad because Nvidia's slides show a best case scenario?
But aren't you the one who preferred best case scenario to real world usage? From the thread regarding AMD's Mantle press slide claims


:SMH:
The rumors back then was a 10-15% max. So an "up to 45%" in a AAA title and not some benchmark or an old game title, was much better than expected, because it was leaving room for real performance improvements from Mantle. As I was saying back then and you quote me "It is way too big even for a marketing lie". A "guaranteed 15%" from a manufacture hardly is guaranteed. So, that 45% was really promising.

In Nvidia's case we have a marketing campaign about how they transformed DirectX 11 in DirectX 12. To do that of course they used a cpu that almost no one is using because it is just too expensive. By using that cpu they also eliminated whatever performance gains Mantle was going to have. Then they took the standard optimizations that we are all used to see from a new driver and presented them as something that no one has seen in the past. As something new.

Those two cases are different but never mind that. As you can see from what I was saying back then, even in the case of AMD I do keep in mind the possibility something to be just marketing lies. In your case you are happy to swallow the whole marketing campaign and be happy about that. Well, if this makes you happy, good for you. But me, I don't swallow what the marketing department is shouting.


by: pr0n Inspector
No, you pull your head out of the sand, take off your welding goggles and look at how it improved performance for a guy with Q6600/GTX 670 and massively improved performance for SLI setups.

And where did you get the idea that it will provide massive boost for single-gpu at 2560 resolution? Despite what you've imagined in your head, nvidia's own benchmarks are not insane at all.
You keep putting words in my mouth to make me look wrong. keep going.
Posted on Reply
#17
HumanSmoke
by: john_
In your case you are happy to swallow the whole marketing campaign and be happy about that.
Really? What I actually said was:
by: HumanSmoke
Why do you expect an Nvidia driver to have wholesale improvements and live up to PR hype when Mantle is in the same boat?.
Not really the same thing is it?
To quote you:
by: john_
You keep putting words in my mouth to make me look wrong
:SMH:

by: john_
So an "up to 45%" in a AAA title and not some benchmark or an old game title...As I was saying back then and you quote me "It is way too big even for a marketing lie". A "guaranteed 15%" from a manufacture hardly is guaranteed. So, that 45% was really promising
by: john_
But me, I don't swallow what the marketing department is shouting.
So, 45% said by AMD's PR guys is "too big" to be a lie and is "really promising"....but you don't really believe it :confused: :confused: :confused:
Posted on Reply
#18
john_
by: HumanSmoke
So, 45% said by AMD's PR guys is "too big" to be a lie and is "really promising"....but you don't really believe it :confused: :confused: :confused:
I always expect something that a company says to be possibly just the best case scenario of the best case scenario. But even someone who works in a marketing departments knows that if the performance gains are about 10% you don't shout 45% because it will backfire. So yes I don't really believe it, I don't take it for granted, I only hope this is true, I have to wait and see, but a 45% was really promising because it was too big to be a lie and if it was a lie it would have been too stupid from AMD's side.

In the case of Nvidia the "Mantle killer" is nothing more than typical optimizations and SLI fixes in specific games and we all know that when a company fixes SLI/CF performance, usually we see major double digit gains. They used a freaking 12 core cpu to present the Mantle killer and they also let everyone who was happy to believe them to think that this is a some kind of wonder driver that transforms DirectX 11 to DirectX 12. Not to mention that 780Ti is anyway not just more expensive, but also faster than 290X. So I pointed out at those things and what did I get? Comments about "Flawless logic", "High comedy" and the title of the one who is "defending the stupid".

In the other thread about these drivers, opinions are not even close to that this is a "wonder driver". It is just a typical driver and of course any new performance gains are as always welcomed. Nothing bad in it.

And NO. You can't put Mantle in the same boat with this driver. That's what Nvidia is trying to do with this marketing campaign. They presented a typical driver update as something equivalent with Mantle, using of course a 12 thread fast cpu of course which is another way to fight cpu bottlenecks (poor AMD haven't though of that solution, LOL).
In that phrase you are quoting from your post,
Why do you expect an Nvidia driver to have wholesale improvements and live up to PR hype when Mantle is in the same boat?.
you are playing happily that game. Marketing succeeded with you.
Posted on Reply
#19
HumanSmoke
by: john_
you are playing happily that game. Marketing succeeded with you.
Says the guy with an AMD card and 8 posts in a thread about Nvidia drivers :rolleyes:
Just to clarify, since my original point seems so far above your head it could be a satellite
"Why do you expect an Nvidia driver to have wholesale improvements and live up to PR hype when Mantle is in the same boat?" refers to the hype surrounding both Nvidia's driver and Mantle. I'll show you an example of this hype...unless you believe everything here
Posted on Reply
#20
john_
by: HumanSmoke
Says the guy with an AMD card
HD4890, HD5670 and HD6850
and
9600GT, 9800GT, 9800GX2, GT210, GT240, 550Ti, 560Ti, GT620

Those are the cards that passed from my hands the last few years.

But when we have nothing to say we start pointing at the other guys specs, not all the specs, we forget that GT620 there. Then we talk about our divine intelligence and throw a slide that it points out to the general advantages of a new API and then we expect the other guy to say that this is the same thing with a typical new driver update that it is marketed as a new API.
You are trolling.
Posted on Reply
#21
burebista
It's not a benchmark per se just a personal observation.
I'm playing a single game (War Thunder) and since December I'm a Closed beta tester for their tanks. My FPS's playing tanks were always in 30-40 interval.
Yesterday after installing this driver my FPS's went up into 50-60 interval. :twitch:

I don't know for sure if it's only a driver improvement or Gaijin did some wonders in yesterday patch but I can say for sure that in 3 months since I'm playing Ground Forces I never see 60 FPS on my rig. Never.

For example playing planes and I see no improvement, almost a steady 60 FPS (VSync on) as always but it's normal because with planes is almost nothing to do for CPU because battles are in a clear sky with few airplanes around.
In Ground Forces we have a lot of terrain/bushes/environment/smoke/fire to render and CPU is put to work and it seems that latest wonder driver is doing his job here.

My system is in Specs.
Posted on Reply
#22
john_
Specific game optimizations, or cases where the newer driver works better because the last one wasn't installed correctly, or cases where a patch fixes performance or other cases, are something that happens for MORE than a decade with Nvidia/AMD(ATI) drivers.
The question here is if this driver update is any different than other driver updates in the past. In my opinion the only different is the marketing campaign behind this driver release. Nothing more, nothing less. 10%, 20% even 50% performance improvements especially in SLi cases are not unheard.

Some NVIDIA R337.50 Driver Controversy | PC Perspective
Posted on Reply
#23
tokyoduong
by: john_
Specific game optimizations, or cases where the newer driver works better because the last one wasn't installed correctly, or cases where a patch fixes performance or other cases, are something that happens for MORE than a decade with Nvidia/AMD(ATI) drivers.
The question here is if this driver update is any different than other driver updates in the past. In my opinion the only different is the marketing campaign behind this driver release. Nothing more, nothing less. 10%, 20% even 50% performance improvements especially in SLi cases are not unheard.

Some NVIDIA R337.50 Driver Controversy | PC Perspective
Extremetech summed it up. http://www.extremetech.com/gaming/180088-nvidias-questionable-geforce-337-50-driver-or-why-you-shouldnt-trust-manufacturer-provided-numbers

the "up to" is all you need to know. For example, SLI systems gained 71% on Total War Rome II because SLI never worked before in that game. Like I said before, these kinds of claim only meant it was broken before or grossly selective representation. The real gains are 2-10% in SOME games. It's still a great driver but the hype machine was at full steam.
Posted on Reply
Add your own comment