NVidia driver 337.50 beta released!

[omeds]

So what are you saying, there is no gain for 780ti in BF4 under win8.1?

 

[Silverforce11]

No, I'm saying BF4 MP with Mantle destroys DX11 even with Win 8.1, and that benches that sum up BF4 using Singleplayer benches are unrepresentative and quite useless.

The other point is, from all these games tested by Computerbase.de with both latest drivers, a single 780ti is no longer faster than the R290X. That includes only BF4 SP benches and no Mantle results. Throw that in, R290X would end up faster on their summary chart. Surprising for such an uber driver from NV touting massively improved performance.

 

[omeds]

Mantle gets a 12% boost at 4k in your example from computerbase. Ti's are 11% in front when both under DX win7. That's 1% difference. It's reasonable to assume that given win8.1 produces large gains at lower resolutions for Nv hardware in BF4, it should at least give a slight boost at 4k too.

Now, how does that explain [H] results? It doesn't, but they're n00bs so who knows. I would expect 4k to be faster on 295x2 given ROP count and fillrate anyway, but I was just pointing the flaw in your logic based on computerbases results.

The other point is, from all these games tested by Computerbase.de with both latest drivers, a single 780ti is no longer faster than the R290X. That includes only BF4 SP benches and no Mantle results. Throw that in, R290X would end up faster on their summary chart. Surprising for such an uber driver from NV touting massively improved performance.

Again, they tested under win7.

 

[96Firebird]

Ah [H]'s BF4 MP testing. Where they had a 760 beating a 770...

 

[Silverforce11]

Mantle gets a 12% boost at 4k in your example from computerbase. Ti's are 11% in front when both under DX win7. That's 1% difference. It's reasonable to assume that given win8.1 produces large gains at lower resolutions for Nv hardware in BF4, it should at least give a slight boost at 4k too.

Now, how does that explain [H] results? It doesn't, but they're n00bs so who knows. I would expect 4k to be faster on 295x2 given ROP count and fillrate anyway, but I was just pointing the flaw in your logic based on computerbases results.



Again, they tested under win7.

You are confusing their SP benches vs MP and expecting a direct correlation between them. In actuality, Computerbase.de's own MP benches with CF R290X and 295 with DX11/Mantle line up very close to [H]'s results, note: 55.7fps @ 4K

Compare that to [H] result of 59fps. Pretty close for a MP bench from two different rigs.

Attacking a review site like [H] is the last refuge of someone who lost the argument. It's akin to attacking AnandTech or TPU as if they are some amateur reviewe site.

 

[omeds]

Oh ok, show me the MP benches with 780Ti's from computerbase because I don't know if the ones I'm looking at in the link are SP or MP.

 

[Silverforce11]

SP bench results:
http://www.computerbase.de/2014-04/...mark-test/6/#diagramm-battlefield-4-3840-2160

They did not test MP for 780ti in this review (hence their summary chart discarded 295 Mantle results), but their MP data for the R290X and 295 matches [H]'s results and there is a massive lead in MP with Mantle there.

I have no issue agreeing that the 780ti is faster it BF4 in Singleplayer.

 

[96Firebird]

Attacking a review site like [H] is the last refuge of someone who lost the argument. It's akin to attacking AnandTech or TPU as if they are some amateur reviewe site.

It's not so much attacking [H] as it is showing that BF4 MP results are unreliable. There is too much variance.

 

[omeds]

SP bench results:
http://www.computerbase.de/2014-04/amd-radeon-r9-295x2-benchmark-test/6/#diagramm-battlefield-4-3840-2160

They did not test MP for 780ti in this review (hence their summary chart discarded 295 Mantle results), but their MP data for the R290X and 295 matches [H]'s results and there is a massive lead in MP with Mantle there.

I have no issue agreeing that the 780ti is faster it BF4 in Singleplayer.

Well then you cant draw conclusions from computerbase results. There is no MP testing nor win8.1.

If you want to base it on [H] that's fine, I still think they're n00bs for other reasons but I believe their 4k results. I would expect 295x2 to be faster at such resolutions by virtue of ROP count and fill rate. However I don't think they reflect sub 1600p results when Ti's are not so fillrate bound.

 

[Sinthoras]

Mid-range gpu's seem to have only little or even no performance increase in most games. Or you must have a very weak CPU.
This site has been testing a GTX 760 in 3 different setups. One with an i7 at 4.2GHz, an A10 at stock and the i7 underclocked to 1.9GHz. Pretty interesting though...

http://blackholetec.com/drupal7/article/review-nvidia-geforce-33750-beta-driver

 

[krumme]

seems that you guys might have already figured this out but extreme tech had a article on the lack of promised advancements from this driver release

http://www.extremetech.com/gaming/1...-shouldnt-trust-manufacturer-provided-numbers

Here is a quote that imo seems to sum it up:

" Overall, the 337.50 driver mostly seems to reduce the bottleneck caused by SLI setups — a known issue with DirectX 11. If you have an SLI setup, you should get fairly reliable gains across multiple games with the 337.50 driver. If you have a single-GPU setup — be it mid- or high-range — you probably won’t notice much difference at all. If BlackHoleTec’s benchmarks are to be believed, there might even be a performance hit from using the new beta drivers.

And thus we arrive at the moral of the story: When trying to gauge the performance of a certain component, you should never trust first-party benchmarks. I get the nagging feeling that, rather than earnestly attempting to reduce CPU bottlenecking and increase game performance, Nvidia is instead trying to draw attention away from AMD’s Mantle. This isn’t surprising, but it’s something to bear in mind."

 

[Ed1]

here interesting tests with core count altered between 335 and 337beta .

http://www.computerbase.de/2014-04/geforce-337.50-cpu-skalierung-benchmarks/

Google English translation

http://translate.google.com/transla...cpu-skalierung-benchmarks/&edit-text=&act=url

 

[Attic]

seems that you guys might have already figured this out but extreme tech had a article on the lack of promised advancements from this driver release

http://www.extremetech.com/gaming/1...-shouldnt-trust-manufacturer-provided-numbers

I know a lot of folks were skeptical of what nVidia was putting forward regarding this driver prior to release, none more than me.

And Yea, that review is pretty hard hitting and it's true. That's the kind of honest and to the point review and commentary I like to see from review sites. There's good reason a lot of folks can't stand nVidia's PR games and deceptions that are constantly run against PC gamers and GPU enthusiasts. Though both companies do it, a lot of folks on forums believe whomever does is hurting PC gaming.

 

[Carfax83]

here interesting tests with core count altered between 335 and 337beta .

http://www.computerbase.de/2014-04/geforce-337.50-cpu-skalierung-benchmarks/

Google English translation

http://translate.google.com/transla...cpu-skalierung-benchmarks/&edit-text=&act=url

Thanks for the link. It apparently dismisses a forum rumor that NVidia got the gains from increasing the temperature boost clock limit :thumbsup:

People can say what they want about these drivers, but they are definitely legit and free performance. There's more to come apparently as well, with the WHQLs..

 

[Carfax83]

No, I'm saying BF4 MP with Mantle destroys DX11 even with Win 8.1, and that benches that sum up BF4 using Singleplayer benches are unrepresentative and quite useless.

And similarly, benches that DON'T use Windows 8/8.1 are similarly unrepresentative and quite useless..

 

[Silverforce11]

And similarly, benches that DON'T use Windows 8/8.1 are similarly unrepresentative and quite useless..

Sure, though lots of users still on Win7 would disagree about how unrepresentative it is. However, the [H] results were obtained with Win 8.1 AND a highly OC i7 CPU.

 

[VulgarDisplay]

Thanks for the link. It apparently dismisses a forum rumor that NVidia got the gains from increasing the temperature boost clock limit :thumbsup:

People can say what they want about these drivers, but they are definitely legit and free performance. There's more to come apparently as well, with the WHQLs..

What does the article say about it the translation is horrible?

 

[Deders]

Mid-range gpu's seem to have only little or even no performance increase in most games. Or you must have a very weak CPU.
This site has been testing a GTX 760 in 3 different setups. One with an i7 at 4.2GHz, an A10 at stock and the i7 underclocked to 1.9GHz. Pretty interesting though...

http://blackholetec.com/drupal7/article/review-nvidia-geforce-33750-beta-driver

Interesting. have any other sites done mid to low range CPU benchmarks?

 

[Carfax83]

Sure, though lots of users still on Win7 would disagree about how unrepresentative it is. However, the [H] results were obtained with Win 8.1 AND a highly OC i7 CPU.

If you want to make a fair comparison in the best light possible for both vendors, then Windows 8.1 is absolutely necessary..

And I think most NVidia users would admit that AMD has the edge right now at 4K resolution..

 

[Carfax83]

What does the article say about it the translation is horrible?

The last paragraph:

Only the theory arisen in forums that Nvidia should have raised with the GeForce 337.50 Temperature the target and thus the boost clock, we can refute: to this setting, there has been no adjustment.

 

[Ed1]

What does the article say about it the translation is horrible?

They say the driver didn't alter GPU clock or temp setting

It basically shows the drivers did optimize CPU utilization as lower core count gets best % increase , At least from there test points .

 

[Sinthoras]

Interesting. have any other sites done mid to low range CPU benchmarks?

Not that I know, I've only seen one site testing HT off vs on. But I forgot the name of that site.

Most bigger sites have been reviewing with a GTX 780 Ti in SLI and an i7 3930K, that's why I like the one on blackholetec testing with a GTX 760.

 

[Hitman928]

Just fyi, it isn't a "forum rumor," hardocp stated in their review that their 780ti was reading higher boost clocks and 3 degrees higher temps than previous drivers. Since I don't speak German, I can't confirm the actual translation, but I could shoot an email to h to confirm.

 

[Sinthoras]

Just fyi, it isn't a "forum rumor," hardocp stated in their review that their 780ti was reading higher boost clocks and 3 degrees higher temps than previous drivers. Since I don't speak German, I can't confirm the actual translation, but I could shoot an email to h to confirm.

Well, I live in Belgium, I speak German as my third language, so I'll read over the article this evening

 

[toyota]

my 780 is not clocking any higher and it's going to the same exact boost clock as always. and lets not be dramatic here as hard is reporting just a 13 mhz increase from when they originally reviewed that 780 ti. And I really doubt the temps are being reported in a perfectly controlled environment. Plus if the card is being utilized more with the new driver then temps would go up a degree or two.