AMD "Never Settle" 12.11 Driver - benchmarks are in!

[BFG10K]

Maybe some good news for the red team for a change. Promises optimizations for the GCN architecture across several games.

Slides: http://www.techpowerup.com/174145/AMD-Announces-Never-Settle-Catalyst-12.11-Driver-Update.html

Game bundles: http://www.techpowerup.com/174146/AMD-Also-Announces-Never-Settle-Game-Bundles.html

Benchmarks:

http://www.techpowerup.com/reviews/AMD/Catalyst_12.11_Performance/1.html

http://www.anandtech.com/show/6393/amds-holiday-plans-cat1211-new-bundle

http://hexus.net/tech/reviews/graph...211-benchmarked-surprising-performance-gains/

http://www.guru3d.com/articles_pages/amd_catalyst_12_11_driver_performance.html

http://www.hardwarecanucks.com/foru...md-12-11-never-settle-driver-performance.html

http://www.computerbase.de/artikel/grafikkarten/2012/bericht-amd-catalyst-12.11-beta/

http://www.pcper.com/reviews/Graphics-Cards/AMD-Never-Settle-1211-Beta-Driver-Performance-Testing

 

[RussianSensation]

Wow, we posted 2 minutes apart

This is arguably the best game bundle I've ever see. All those should be pretty solid titles and since they are new prob. can be sold at fairly high prices for those who don't want them. HD7000 series were already offering better price/performance and overclocking at nearly every price level below $500 and this deal completely seals the deal for value. The question is how in the world can AMD afford to bundle 3 high profile $50 games and 20% off another major title?

 

[iMacmatician]

Bar graphs with arrows on the bars… haven't seen those in quite a while.

 

[blastingcap]

sheesh you guys camping out on the same newsfeed lol. competition is good.

edit: if this resulted from their going off monthly driver updates, i'm all for fewer but better updates!!

 

[RussianSensation]

I am going to hold off for professional reviews to validate those driver increases. They often tend to be overly optimistic. At the same time, if this driver is so big for AMD, it would force the review websites to redo all their testing for all HD7000 series which isn't a bad thing to reflect current performance. I've seen some websites still using Cats 12.4-12.6.

 

[blastingcap]

I've seen some websites still using Cats 12.4-12.6.

name 'em and shame 'em!

 

[BD231]

How's image quality on AMD cards lately? 7970s are getting cheap, but does amd still sacrafice image quality to squeeze more performance out of popular titles?

 

[BFG10K]

If someone sends me a GCN part, I’ll be happy to evaluate them.

 

[blastingcap]

How's image quality on AMD cards lately? 7970s are getting cheap, but does amd still sacrafice image quality to squeeze more performance out of popular titles?

BFG10K can answer that, and he's in this thread and led the charge against AMD for setting defaults too low or whatever.

Edit to add: This is what Anandtech's review stated:

Finally, these seem to be clean optimizations, as image quality has been held constant in most of our games (e.g.: Battlefield 3). The sole exception is Skyrim, and this is something we’re certain is an unrelated bug. On our 7970 we appear to be missing a lighting pass, but only on our 7970. On our 7870 with the same exact settings everything is being rendered correctly. AMD has told us that they haven’t seen this issue in-house (it would admittedly be hard to miss) so this may be some esoteric issue; due to the short preview window and our own testing time constraints AMD hasn’t had time to look into it further.

And HWC:

AMD has assured us that every optimization they’ve found hasn’t come at the expense of image quality and we have to agree. After using this driver extensively for the better part of a week, we haven’t been able to find a single instance where visual fidelity may have suffered. 12.11 is simply allowing an architecture that’s been on the market since January another chance to shine again. - http://www.hardwarecanucks.com/foru...md-12-11-never-settle-driver-performance.html

I am going to hold off for professional reviews to validate those driver increases. They often tend to be overly optimistic. At the same time, if this driver is so big for AMD, it would force the review websites to redo all their testing for all HD7000 series which isn't a bad thing to reflect current performance. I've seen some websites still using Cats 12.4-12.6.

You asked. They delivered? http://www.techpowerup.com/reviews/AMD/Catalyst_12.11_Performance/23.html

The graphs on this page show a combined performance summary of all tests and resolutions from previous pages. Each graph shows the tested card as 100% and all other cards' performance as relative to it. A sixth graph summarizes all tests in all resolutions to calculate the total relative performance of the review sample.

The relative improvements for each card are:
HD 7750: +4%
HD 7770: +5%
HD 7850: +4%
HD 7870: +10%
HD 7950: +7%
HD 7970: +7%
HD 7970 GHz: +7%

After exhaustive testing of AMD's new Catalyst 12.11 driver, we can confirm that it brings significant performance improvements with every single game we tested. The differences vary by game. The highlights were Battlefield 3, Sleeping Dogs, Star Craft II, and World of Warcraft. This driver also fixes a bug that caused low performance in Dragon Age 2. I didn't expect the low performance in Dragon Age 2 to be fixed at all given that it is a relatively old title. It shows that AMD did take a real close look at game performance.
I also find it surprising that benchmarks like 3DMark 11 and Unigine Heaven saw good gains - something that typically does not happen during "normal" driver updates.

Our testing also confirms that these performance benefits are only for Radeon HD 7000 cards since the HD 6970 in our test group saw no improvements.

Actual performance improvements vary between cards. Improvements start with 4% for the HD 7750 and reach an amazing 10% for the HD 7870. The HD 7950, HD 7970 and HD 7970 GHz Edition settle in at 7%. Single digit improvements may not sound like much, but these results are the biggest performance boosts we've ever seen through a driver update. Especially the 10% on the HD 7870 made it feel like a completely different card while gaming.

AMD should also be commended for bundling their graphics cards up with high-profile games and not just games that nobody wants to play anymore. If you do not plan on buying these games, selling these coupons will provide you with some more money for your next purchase, which could mean a faster card despite a limited budget.

Check out BF3, btw, AMD obviously focused on that title: http://www.techpowerup.com/reviews/AMD/Catalyst_12.11_Performance/5.html

 

[BallaTheFeared]

Wow graphs starting at 90%, that's new for me.

 

[RussianSensation]

How's image quality on AMD cards lately? 7970s are getting cheap, but does amd still sacrafice image quality to squeeze more performance out of popular titles?

"Still sacrifice"? Do you have proof of this, that AMD uses a practice of sacrificing image quality regularly? What are you saying that HD5800 and 6900 series had worse image quality than GTX400/500 series and that somehow HD7000 series has worse image quality than NV's 600 series? Both companies have differences in anti-aliasing methods they use (FXAA vs. MLAA, Edge Detect CFAA vs. TRSAA for NV) but that has been the case for a long time now. I think there was a texture issue with 1 driver release earlier in the year that affected only HD7000 series when Tom's hardware tested it but they later acknowledged it was fixed with the next version. Your comment not only made it general as if AMD purposely cripples image quality but you referred to sacrificing image quality in popular titles, I am guessing you know something the rest of the gaming community does not? And which popular titles showed a major difference in image quality in 2012?

"As a final prelude to the benchmarks, we spent a little time in comparing the image-quality attributes of the Catalyst 12.8 and 12.11 drivers. Both have the same default IQ parameters in the control panel, and looking at side-by-side screenshots showed no image-quality 'cheating' in the newer set - something that would aid performance at the expense of IQ." ~ Hexus

You asked. They delivered? http://www.techpowerup.com/reviews/AMD/Catalyst_12.11_Performance/23.html

Thanks blastincap! That was fast on TPU's part. :thumbsup:

 

[Jaydip]

Wow graphs starting at 90%, that's new for me.

I remember NV used the same for their slides :biggrin: Its silly but well whatever.

 

[BFG10K]

I've added the benchmarks to the OP. :thumbsup:

 

[DarkKnightDude]

There's some great performance increases, damn BF3 especially. And Sleeping Dogs!

 

[blastingcap]

I've added the benchmarks to the OP. :thumbsup:

Great, feel free to add more as the NDAs expire. For instance:

http://www.anandtech.com/show/6393/amds-holiday-plans-cat1211-new-bundle

But it’s Battlefield 3 that really takes the cake: the performance improvement from Catalyst 12.11 ranges from 13% for the 7770 at 1680 to a whopping 29% for the 7970 at 1920. This makes Catalyst 12.11 a very special driver for AMD – not only are performance improvements over 20% particularly rare, but Battlefield 3 has long been a thorn in AMD’s side. NVIDIA’s hardware has until now always outperformed AMD’s equivalent hardware here, and as BF3 has remained an extremely popular MP game it’s been one of the most important games for high-end video card buyers. In other words it has always been the game AMD could least afford to lose at.

With the 12.11 drivers AMD has completely eradicated their performance defecit in BF3, with the 7970, 7870, and 7770 being made performance competitive with (if not a hair faster than) their respective NVIDIA GTX 600 counterparts in our BF3 benchmark. AMD has told us that the specific performance benefits are map-dependent with our results appearing at the high-end of their guidance, so while not every map will see the same 20%+ performance gains, some of them will while others will be in the 10% range. Much like our overall performance averages, the largest gains are at 1920 with FXAA, while 2560 with FXAA and 1920 with MSAA will see smaller gains, once again hinting that AMD’s optimizations are on the shader/texture side rather than ROP/memory.

http://hexus.net/tech/reviews/graph...chmarked-surprising-performance-gains/?page=5

We wanted to see if the extra performance increased system-wide power-draw. The answer seems to be no, going by our near-on identical figures derived from Batman: Arkham City.

http://www.legitreviews.com/article/2059/10/ and http://www.legitreviews.com/article/2059/11/

Power Consumption Results: The system with the Sapphire Radeon HD 7970 Vapor-X used the same power at idle reguardless of the driver, but at load there was a clear difference. We were seeing about 14 Watts higher power use on average in game titles. We asked AMD about this power increase and they said it was because the new driver more effectively uses the GPU cores, which means they could possibly mean more power being used. AMD wasn't 100% sure and we won't begin to speculate, but there is a power increase that can be observed.

and

The Catalyst 12.11 Never Settle driver didn't impact the cards overclockability, which is great news. We found that taking a Radeon HD 7970 and using the 12.11 driver along with a little overclocking created one heck of a gaming graphics card!

http://www.hardwarecanucks.com/foru...md-12-11-never-settle-driver-performance.html

 

[badb0y]

Sweet, good for me

 

[RussianSensation]

Crysis 1 - HD7970 GE - first stock single-GPU to reach 60 fps at 1080P. It finally happened 5 years after the game launched.

Overall, at 1600P HD7970 GE is now 48% and 65% faster than GTX580 and HD6970, respectively. I guess this generation didn't turn out that bad.

 

[raghu78]

http://www.anandtech.com/show/6393/amds-holiday-plans-cat1211-new-bundle

"
With the 12.11 drivers AMD has completely eradicated their performance defecit in BF3, with the 7970, 7870, and 7770 being made performance competitive with (if not a hair faster than) their respective NVIDIA GTX 600 counterparts in our BF3 benchmark. AMD has told us that the specific performance benefits are map-dependent with our results appearing at the high-end of their guidance, so while not every map will see the same 20%+ performance gains, some of them will while others will be in the 10% range. Much like our overall performance averages, the largest gains are at 1920 with FXAA, while 2560 with FXAA and 1920 with MSAA will see smaller gains, once again hinting that AMD’s optimizations are on the shader/texture side rather than ROP/memory.

For our part we have long theorized that the Frostbite 2 engine’s heavy use of deferred rendering techniques – particularly its massive G-buffer – was the factor that AMD was struggling with. While these results don’t really further validate or invalidate that theory, what is clear is that AMD has fixed their Frostbite 2 performance problem. Given the fact that Frostbite 2 will be used in at least a couple more games, including the AMD Gaming Evolved title Medal of Honor Warfighter, this was an important engine for AMD to finally conquer."

Seriously with this driver AMD has clearly addressed the one issue which mattered most. Frostbite 2 performance. really matters as MOH Warfighter is out tomorrow. bravo AMD. :thumbsup:

http://www.techpowerup.com/reviews/AMD/Catalyst_12.11_Performance/23.html

TPU has confirmed the performance improvements. BF3 at close to 10%. Sleeping dogs, Max Payne 3 see more than 5% performance gain. Surprisingly world of warcraft sees a 10% performance improvement.

AMD has clearly leapfrogged the competition not just in performance but also with their aggressive holiday bundle. AMD has the strongest holiday product stack ever.

 

[raghu78]

Crysis 1 - HD7970 GE - first stock single-GPU to reach 60 fps at 1080P. It finally happened 5 years after the game launched.

Overall, at 1600P HD7970 GE is now 48% and 65% faster than GTX580 and HD6970, respectively. I guess this generation didn't turn out that bad.

Yeah this proves the learning on a new architecture takes time. I am excited for HD 8900 series.

 

[lavaheadache]

Hurry up and gimme this driver!!!!

 

[Meekers]

In a funny way I guess all the AMD bashers were right, the drivers did suck!

 

[skipsneeky2]

These new numbers from a single 7970 have ended what i can consider to have been the most difficult choice as far as which card to get,the gtx680 or the 7970...

Wonder if i picked up a simple $400 7970 and clocked it to ge specs,would it perform nearly identical if not the same as the review numbers suggest for the ge with 12.11?

 

[thilanliyan]

When are these game bundles being released?

 

[DarkKnightDude]

When are these game bundles being released?

I'm guessing later this month or early November, we'll start seeing the bundles pop up.

AMD hasn’t set a hard date on when the bundle will go live, and since their marketing department is typically ahead of their promotions department by a week or two we’d expect to see these bundles finally become active in November.

 

[RussianSensation]

In a funny way I guess all the AMD bashers were right, the drivers did suck!

That's like saying Fermi drivers sucked before their 10% performance increase. New GPU architectures designed from the ground-up such as Fermi or GCN are expected to have a larger performance increase with drivers than mature architecture. The difference was GCN didn't just have driver issues, but AMD introduced the driver recovery feature that would black screen your monitor if your overclock was not stable, CF drivers were poor and performance was slower in BF3 a very popular game that worked well on NV right away. In general though, GCN was always criticized this way but when Fermi got faster drivers over, people expected that to happen and never really said that Fermi's drivers sucked at launch. Double standard. Kepler was not a real new architecture compared to Fermi, which is why NV had more mature drivers out of the gate. At the same time that meant they had very little room to to improve driver performance in games and we see this as AMD continues to extend its lead in games where it was leading already, while NV hasn't closed those gaps (other than Shogun 2, which was a good job by NV). Not that it matters since HD7970 GE was already faster on average than 680 was, all the way in June when Cats 12.7 betas came out.