Nvidia Kepler performance driver fix released.

[nurturedhate]

Yea, noticing 10-15% improvement from this latest driver with a 780ti. So...

1) This was just completely overlooked, Nvidia forgot Kepler or they "broke" something and it took about a week for a 10-15% improvement....

or

2) This was on purpose and the backlash was too much to handle....

This really isn't hard to figure out. If it is option 1 then it is complete incompetence and if it is option 2 then.. well.. option 2.

To the legion of Nvidia people that posts here, a question for you... is your company incompetent or evil?

 

[xthetenth]

To the legion of Nvidia people that posts here, a question for you... is your company incompetent or evil?

Untrustworthy.

 

[Face2Face]

Here is some more info on the newest drivers 353.06

Kepler Performance Optimizations

Following end user reports of lower-than-expected performance in The Witcher 3: Wild Hunt when using GeForce GTX 600 and 700 Series Kepler GPUs, we have identified and fixed three bugs that were limiting performance not only in The Witcher 3, but also Far Cry 4 and Project Cars. With the new GeForce Game Ready drivers installed, frame rates are increased in each title, improving and optimizing your experience.

We would like to thank the community for their feedback on this issue. These reports and benchmarks helped greatly in rapidly identifying and resolving the issue. If you have any further feedback following the installation of the new driver, please post it in this thread.

What some users are saying on Guru3D

This is sort of amazing. My frame rate increase to at least 10 to 20 fps at least. Some scenes the increase is close to 40 fps.

My frame rate before in the open world with everything on ultra and on, I was getting 60 to 70 fps.

Now i am getting 79 to 85 fps. Close up scenes to Geralts hair would drop down to 55 now Its at 65 70. The most intense scenes where it would drop to 45 fps, stays at 60 fps.

This is confirmation that Kepler was getting screwed, for all those naysayers. Now i am getting a frame rate comparable to 980s in sli. This is using the "old tech", you know the tech that loses to a 970. Please. So many stupid people posting crap.

http://forums.guru3d.com/showpost.php?p=5085759&postcount=82

I can confirm, I've very impressed:

Before driver update in Witcher 3 -
55 max FPS, with really bad dips constantly to 30-40 and averaging around 45-50 fps.
after driver update in Witcher 3 -
60 almost stable, occasional dips to 55 but much more stable almost locked at 60.

So it may work for more heavy games in the future, I bloody hope it does considering how close to unplayable my Witcher 3 was before this update.

http://forums.guru3d.com/showpost.php?p=5085716&postcount=67


Some additional testing -
http://www.computerbase.de/2015-06/geforce-353.06-leistung-kepler-witcher-3/

 

[Face2Face]

Loaded up The Witcher 3 tonight. Here are my results

352.86 Driver

50.2 FPS

353.06 Driver

54.7 FPS

353.06 Driver- PhysX on CPU

56.5 FPS

From the little testing I've done thus far, the new 353.06 drivers are a little better, but stutter and choppiness is still prevalent. With the PhysX moved to the CPU, the game losses the stutter and feels really smooth. It looks like moving PhysX to the CPU was a must for my setup.

Game settings:

1440p Full screen
Ultra settings - With exceptions below
Shadow Quality - Medium
Terrain Quality - High
Grass density - Medium
Foliage range - Medium

All post processing enabled, except for Vignette and AA.
Sharpening on Low
HBAO+ enabled

NO HairWorks enabled

EDIT: I upped Terrain quality to Ultra and Grass and Foliage up to High and I'm seeing on average around 50 FPS in-game. The game feels so much smoother now and is much more playable. I'm glad NVIDIA took the anchor off Kepler with this driver update.

 

[SolMiester]

It shouldn't matter what the reason is. As a consumer, if one spends $699 for a flagship card in November 2013, he/she should expect it to mop the floor with a $399 R9 290 in 95% of games and remain so for the next 2 years. Today, $650 780 is getting killed by a $400 R9 290, the card it was meant to compete with, while 780Ti is slower than R9 290X. It's so bad that 980Ti at $649 is 75-80% faster than a 780, which I guess is a good upgrade for 780 owners!

Emmmm, you could say that for the flagship 79xx owners that went CF and had broken drivers for what was it 6mths?, not to mention the 2yr wait for runt frame etc...
Resources or incompetence?

 

[Flapdrol1337]

The game feels so much smoother now and is much more playable. I'm glad NVIDIA took the anchor off Kepler with this driver update.

It does feel smoother. Even if it's just 10%, maybe consistency has improved more.

 

[Face2Face]

I've had a chance to do a bit more testing, so here's what I've come up with.

All tests where done @ 1440p with the GTX 780 running at stock speed of (1124/1502) Core clock speed never fluctuated while testing.

Shadow of Mordor: 353.06 - 54.79 FPS 352.86 - 54.35 FPS
Bioshock Infinite: 353.06 - 70.47 FPS 352.86 - 70.23 FPS
Thief 353.06 - 45.9 FPS 352.86 - 46.9 FPS
Witcher 3 353.06 - 34.8 FPS 352.86 - 31.65 FPS
GTA V 353.06 - 75 FPS 352.86 - 63.38 FPS

As you can see, older games aren't seeing any benefit from the newer drivers, but GTA V and The Witcher 3 are seeing some gains. I will say, while the The Witcher 3 isn't showing huge improvements on the FPS end, the game is much smoother and less jittery with the 353.06 driver.

Moving on to GTA V the FPS gains are significant. Not only does the game run faster, but again like Witcher 3, the games is much smoother.

Overall, I feel the Witcher 3 is the major game tested that really pounds Kepler. The Witcher was run with Ultra settings, AA - HBAO+, and No Hairworks. Moving the foliage distance and shadows down to high and disabling AA, ups the frame rate to the 40-50 FPS range - Overclocked to nearly 1300Mhz, ups the frame rate to the - 50-60 FPS range.

 

[Deders]

I've had a chance to do a bit more testing, so here's what I've come up with.

All tests where done @ 1440p with the GTX 780 running at stock speed of (1124/1502) Core clock speed never fluctuated while testing.

Shadow of Mordor: 353.06 - 54.79 FPS 352.86 - 54.35 FPS
Bioshock Infinite: 353.06 - 70.47 FPS 352.86 - 70.23 FPS
Thief 353.06 - 45.9 FPS 352.86 - 46.9 FPS
Witcher 3 353.06 - 34.8 FPS 352.86 - 31.65 FPS
GTA V 353.06 - 75 FPS 352.86 - 63.38 FPS

As you can see, older games aren't seeing any benefit from the newer drivers, but GTA V and The Witcher 3 are seeing some gains. I will say, while the The Witcher 3 isn't showing huge improvements on the FPS end, the game is much smoother and less jittery with the 353.06 driver.

Moving on to GTA V the FPS gains are significant. Not only does the game run faster, but again like Witcher 3, the games is much smoother.

Overall, I feel the Witcher 3 is the major game tested that really pounds Kepler. The Witcher was run with Ultra settings, AA - HBAO+, and No Hairworks. Moving the foliage distance and shadows down to high and disabling AA, ups the frame rate to the 40-50 FPS range - Overclocked to nearly 1300Mhz, ups the frame rate to the - 50-60 FPS range.

It's only really demanding games that were released after the 970 and 980 were made available that seem to suffer from poor (or no) driver optimisation. It's a consistent pattern that makes me feel like a 2nd teir customer, even though I bought my 780 2 months before Maxwell was released.

 

[YBS1]

Ya, I am starting to think the business model of 2-year-GPU architecture life-cycles suits NV very well. It's like the smartphone 2-year contract model, without the official contract. Either way, looks like NV is converting an awful lot of gamers to Maxwell. If they cripple Kepler cards, they know that their gamers will just move on to Maxwell and not even consider any R9 200 series cards. The types of gamers who pay attention to how fast a GPU architecture falls apart or how the driver optimizations have stopped are a tiny minority of the entire GPU buying market. This could be very dangerous for the GPU industry since a lot of those gamers will just keep rolling over to the next generation of NV cards which means AMD could be stuck in the 20% market share or less for the next 5 years.

Knowing how competitive Maxwell cards are, it gives NV almost no incentive at all to optimize the drivers for Kepler. What's the point when there is so little risk that your loyal customers will abandon you? It's a win-win for NV since they make more $ crippling older generations and once their loyal customers upgrade and spread word-of-mouth experience of how amazing Kepler is, it gets brand agnostic gamers to upgrade too.

While I agree with about everything you've said here, what is the alternative? Buying the other brand where the card doesn't perform up to snuff for the first 6-12 months of it's lifespan? Lagging support for new games right out of the gate, sometimes ages for proper multi-card support, and now having to deal with GameDon'tWorks bs? Don't write me off as you're so apt to do lately as a close minded, only buys nVidia no matter what customer. I want AMD to succeed, we need AMD to succeed, but they have to do the legwork. I haven't bought a 980 Ti yet because I want to see what AMD brings, I want to see another 9700 Pro, I want to see nVidia humbled. I was impressed with the R9 290(X) generation, particularly the Crossfire scaling as compared to SLI, but that means nothing if I've already stopped playing the game by the time AMD gets it sorted. This is what AMD is up against.

I don't think there are as many true loyal buyers as you think there are. If AMD could put their foot on nVidia's throat (it may well take more than one generation of doing it though), you might get to see that. This generation may be lost in any case though, being so late to the market eliminates many who might have been potential buyers if given the choice. The timing happens to work well for me as I was waiting on the Classifieds/HOFs/Lightnings anyway, so I get to see Fury X as well, but many have already jumped.

 

[escrow4]

Finished V before the driver update and the Witcher III was maybe a touch smoother. I'm not upgrading because there is nothing to upgrade for. And I don't do smartphones. Tried 2 went back to a trusty Nokia dumbphone. I do believe I'll be switching to AMD permanently though. Nvidia is stupid shifty.

 

[amenx]

I would ditch Nvidia in a heartbeat if it ever became common practice for them to neglect driver support for older gen cards. They have a history of supporting much older cards than kepler and I often find fixes for fermi or older cards in release notes of drivers out over last few months. The issue with kepler appears to have been a temporary hiccup in driver support but with no similar history to establish a pattern of any sort.

 

[Spanners]

Emmmm, you could say that for the flagship 79xx owners that went CF and had broken drivers for what was it 6mths?, not to mention the 2yr wait for runt frame etc...
Resources or incompetence?

How is that relevant? You could bring up instances where either company has had missteps till the cows come home. Doesn't make Kepler's recent performance any different. AMD was taken to task at the time for everything you mentioned. Tu quoque FTL.

 

[nurturedhate]

How is that relevant? You could bring up instances where either company has had missteps till the cows come home. Doesn't make Kepler's recent performance any different. AMD was taken to task at the time for everything you mentioned. Tu quoque FTL.

Exactly. Though it seems to be pretty common. Make a thread about nvidia drivers and be reminded every three posts that AMD is so very much worse..... sure... Both companies have their issues, but I don't remember any AMD drivers killing cards.

 

[dark zero]

I have a GTX 740M... I'll test this driver on how this driver improve my chip.

I wonder if this will get some notable gains on GTA V and DOTA 2

 

[rgallant]

I would ditch Nvidia in a heartbeat if it ever became common practice for them to neglect driver support for older gen cards. They have a history of supporting much older cards than kepler and I often find fixes for fermi or older cards in release notes of drivers out over last few months. The issue with kepler appears to have been a temporary hiccup in driver support but with no similar history to establish a pattern of any sort.

just saying
my take on nv statements is that they found the issues only after testing kelper in those games.

so nv did not test kelper in those GE games ,when they helped the dev's to add features.

so not a temporary driver issue if they did not test kepler, nor react to it on their own forums until they had to.

imo

 

[hawtdawg]

Emmmm, you could say that for the flagship 79xx owners that went CF and had broken drivers for what was it 6mths?, not to mention the 2yr wait for runt frame etc...
Resources or incompetence?

Just not as good of marketing. Now that Maxwell has terrible SLI frametimes, instead of fixing the issue, they just pay their mouthpeices to keep it quiet (Pcper)

 

[x3sphere]

Just not as good of marketing. Now that Maxwell has terrible SLI frametimes, instead of fixing the issue, they just pay their mouthpeices to keep it quiet (Pcper)

SLI or CF, frametimes still suck compared to single GPU. SLI isn't any worse with Maxwell it's just that AMD got a lot better. The two are pretty much dead even now, it just depends on the game as to which is ahead.

 

[happy medium]

Now that Maxwell has terrible SLI frametimes, instead of fixing the issue, they just pay their mouthpeices to keep it quiet (Pcper)

Mabe you should go to the source instead of just spreading BS.

quoted from pcper reviewer:
"Why do I keep having to repeat this statement. I use frame times on every GPU review I have done since I started using Frame Rating software!! Let me demonstrate:"
http://hardforum.com/showpost.php?p=1041664517&postcount=113
"
GTX 980 Ti Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX Titan X Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX 960 Overclocking Testing: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX 970 Memory Issue SLI Testing: http://www.pcper.com/files/imagecach...160_PLOT_0.png

GTX 970 Memory Issue Single GPU Testing: http://www.pcper.com/files/imagecach...440_PLOT_0.png

GTX 960 Review: http://www.pcper.com/files/imagecach...x1080_PLOT.png

MSI GTX 970 Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

NVIDIA MFAA Testing: http://www.pcper.com/files/imagecach...0x1440_PER.png

Civ Beyond Earth DX11/Mantle Testing: http://www.pcper.com/files/imagecach...0-23/be-4k.png

GTX 980 3-Way and 4-Way SLI: http://www.pcper.com/files/imagecach...x2160_PLOT.png

Shadow of Mordor Testing: http://www.pcper.com/files/imagecach...440_PLOT_0.png

GTX 980 and GTX 970 Single / SLI Testing: http://www.pcper.com/files/imagecach...x2160_PLOT.png

Radeon R9 285 Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

If you are going to make claims against me at least pick something that isn't so obviously proven wrong."

 

[3DVagabond]

quoted from pcper reviewer:
"Why do I keep having to repeat this statement. I use frame times on every GPU review I have done since I started using Frame Rating software!! Let me demonstrate:"
http://hardforum.com/showpost.php?p=1041664517&postcount=113
"
GTX 980 Ti Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX Titan X Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX 960 Overclocking Testing: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX 970 Memory Issue SLI Testing: http://www.pcper.com/files/imagecach...160_PLOT_0.png

GTX 970 Memory Issue Single GPU Testing: http://www.pcper.com/files/imagecach...440_PLOT_0.png

GTX 960 Review: http://www.pcper.com/files/imagecach...x1080_PLOT.png

MSI GTX 970 Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

NVIDIA MFAA Testing: http://www.pcper.com/files/imagecach...0x1440_PER.png

Civ Beyond Earth DX11/Mantle Testing: http://www.pcper.com/files/imagecach...0-23/be-4k.png

GTX 980 3-Way and 4-Way SLI: http://www.pcper.com/files/imagecach...x2160_PLOT.png

Shadow of Mordor Testing: http://www.pcper.com/files/imagecach...440_PLOT_0.png

GTX 980 and GTX 970 Single / SLI Testing: http://www.pcper.com/files/imagecach...x2160_PLOT.png

Radeon R9 285 Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

If you are going to make claims against me at least pick something that isn't so obviously proven wrong."

Getting 404 message on all of the links.

Went to the source links and got them.

 

[happy medium]

Getting 404 message on all of the links.

Went to the source links and got them.

Well thanks, but I made my point.

Source for the rest of you guys.
http://hardforum.com/showpost.php?p=1041664517&postcount=113

 

[3DVagabond]

Well thanks, but I made my point.

Source for the rest of you guys.
http://hardforum.com/showpost.php?p=1041664517&postcount=113

No inference. I was simply letting you know so you could fix them.

 

[AnandThenMan]

Those frame time graphs are nearly useless they need to be much larger and the time scale stretched out.

 

[SolMiester]

How is that relevant? You could bring up instances where either company has had missteps till the cows come home. Doesn't make Kepler's recent performance any different. AMD was taken to task at the time for everything you mentioned. Tu quoque FTL.

Revelance is one was an oversight, the other was inablilty to correct a feature for damn near the life of the product.

 

[ocre]

Mabe you should go to the source instead of just spreading BS.

quoted from pcper reviewer:
"Why do I keep having to repeat this statement. I use frame times on every GPU review I have done since I started using Frame Rating software!! Let me demonstrate:"
http://hardforum.com/showpost.php?p=1041664517&postcount=113
"
GTX 980 Ti Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX Titan X Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX 960 Overclocking Testing: http://www.pcper.com/files/imagecach...x1440_PLOT.png

GTX 970 Memory Issue SLI Testing: http://www.pcper.com/files/imagecach...160_PLOT_0.png

GTX 970 Memory Issue Single GPU Testing: http://www.pcper.com/files/imagecach...440_PLOT_0.png

GTX 960 Review: http://www.pcper.com/files/imagecach...x1080_PLOT.png

MSI GTX 970 Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

NVIDIA MFAA Testing: http://www.pcper.com/files/imagecach...0x1440_PER.png

Civ Beyond Earth DX11/Mantle Testing: http://www.pcper.com/files/imagecach...0-23/be-4k.png

GTX 980 3-Way and 4-Way SLI: http://www.pcper.com/files/imagecach...x2160_PLOT.png

Shadow of Mordor Testing: http://www.pcper.com/files/imagecach...440_PLOT_0.png

GTX 980 and GTX 970 Single / SLI Testing: http://www.pcper.com/files/imagecach...x2160_PLOT.png

Radeon R9 285 Review: http://www.pcper.com/files/imagecach...x1440_PLOT.png

If you are going to make claims against me at least pick something that isn't so obviously proven wrong."

I just am starting to wonder why almost the same things just keep getting repeated from one site to the next. It will be different posters but almost the exact same thing over and over. It is not a jab at the guy who said this, perhaps he read it and started saying the same thing here. But it seems to happen a lot these days.

We have few deep or original discussion anymore. Just the same stuff spilling over in every thread.

 

[skipsneeky2]

These recent drivers offer any improvement for a gtx650 with maybe GTA 5?Running 350.12 and revisiting the game on the 650 and i3 3220 it feels dramatically different then it did in April.I figured patches are responsible as i have heard complaints about stutter recently.

I rarely update from something that works,newest of drivers appear very iffy and GTA 5 is the most recent of games i play outside of BF4.What do you guys think i should do?Take the plunge?