AMD Radeon HD 7950 Review Feat. Sapphire & XFX: Sewing Up The High-End Market
by Ryan Smith on January 31, 2012 9:02 AM ESTCompute Performance
Moving on from our look at gaming performance, we have our customary look at compute performance. Since compute performance is by definition shader bound, the 7950 is at a bit of a disadvantage here compared to gaming performance. Whereas ROP performance scales with the core clock, shader performance is hit by both the reduction in the core clock and the disabled CU array.
Our first compute benchmark comes from Civilization V, which uses DirectCompute to decompress textures on the fly. Civ V includes a sub-benchmark that exclusively tests the speed of their texture decompression algorithm by repeatedly decompressing the textures required for one of the game’s leader scenes. Note that this is a DX11 DirectCompute benchmark.
AMD’s greatly improved compute performance continues to shine here, though in the case of Civilization V it’s largely consumed by just closing the previously large gap between the GTX 500 series and the Radeon HD 6000 series. As a result the 7950 falls ever so short of the GTX 580, while the factory overclocked Sapphire and XFX cards give the 7950 enough of a push to come within 5% of the 7970.
Our next benchmark is SmallLuxGPU, the GPU ray tracing branch of the open source LuxRender renderer. We’re now using a development build from the version 2.0 branch, and we’ve moved on to a more complex scene that hopefully will provide a greater challenge to our GPUs.
Under SmallLuxGPU the 7970 enjoyed a large lead over the GTX 580, and this continues with the 7950. Even though the 7950 is well behind the 7970—to the tune of 24%—it’s still 33% ahead of the GTX 580 and the lead only grows from there. Meanwhile the XFX and Sapphire cards can catch up to the 7970 somewhat, but as this is truly a shader-bound test, you can’t make up for the lack of shaders units on the 7950.
For our next benchmark we’re looking at AESEncryptDecrypt, an OpenCL AES encryption routine that AES encrypts/decrypts an 8K x 8K pixel square image file. The results of this benchmark are the average time to encrypt the image over a number of iterations of the AES cypher.
In spite of being a compute benchmark, AESEncryptDecrypt is not particularly sensitive to GPU performance, showcasing the impact that setup times can have. The 7950 trails the 7970 by 10%, and overclocking doesn’t change this much. Unfortunately for AMD NVIDIA is still the leader here, showing that AMD’s compute performance still has room to grow.
Finally, our last benchmark is once again looking at compute shader performance, this time through the Fluid simulation sample in the DirectX SDK. This program simulates the motion and interactions of a 16k particle fluid using a compute shader, with a choice of several different algorithms. In this case we’re using an (O)n^2 nearest neighbor method that is optimized by using shared memory to cache data.
With the compute shader fluid simulation we once again shift back into a compute task that’s much more shader-bound. The 7950 only reaches 80% of the performance of the 7970, once more proving the real impact of losing a CU array. This is still enough to handily surpass the GTX 580 however, with the 7950 taking a 15% lead.
Facebook
Twitter
RSS
259 Comments
View All Comments
chizow - Sunday, February 5, 2012 - link
No its not my standard, its the standard for what the market will bear using historical data points as my evidence.Buts its fine, its clear illogical and irrational people such as yourself don't have any standard to determine buy decisions, which is fine too.
Ignorance is bliss.
Galidou - Sunday, February 5, 2012 - link
LOL ultimate knowledge is crazyness...Galidou - Sunday, February 5, 2012 - link
I'm not saying at ANY time that the pricing is super right, that it is the right thing to do, I'm just saying from the beginning that it isn'T the worse that ever happened while you're making a freaking case of it.Galidou - Sunday, February 5, 2012 - link
And btw, the freaking card is around 75-90% faster than a 6950 which isn't bad. Nothing amazing but... That gtx 280 was around 65% to 100% faster than a 8800gt(2 gen below gtx2xx series).Galidou - Sunday, February 5, 2012 - link
Whenever any card came out there was already a part that was an x2 card more powerful and cheaper that what was actually out there, what'S new today?JNo - Thursday, February 2, 2012 - link
@chizow,You keep writing like AMD couldn't adjust their prices after Kepler's launch. Also, you do realise that AMD isn't competing with nvidia's 580 from a year ago right? They are competing with it *right now*.
Well done if you bought the 580 a year ago but which card is better value today if you're a buyer? Right now the 7970 looks to be a better price performance proposition. If AMD's pricing makes the 580 look poor, nvidia are free to adjust their pricing but I'd go for the 7950 personally as it is right now.
And no point pre-judging AMD pricing based on a for 'after the Kepler launch' argmument. Just because nvidia haven't adjusted pricing downwards, doesn't mean AMD won't.
I'm not an AMD fan (I buy both camps) but your arguments don't make sense.
chizow - Thursday, February 2, 2012 - link
@JNoI'm not ignoring that possibility at all, I've actually alluded to the possibility on numerous occasions with my "when Kepler launches" comments. It hinges greatly on what Nvidia does of course and how Kepler performs but I don't think ANYONE expects Nvidia to introduce "next-gen" parts at last-gen performance levels because that's the ONLY way AMD's current pricing on these Tahiti parts will make sense. Why? Because they're basing next-gen pricing on last-gen performance.
Instead, what's most likely to happen based on historical pricing and performance metrics, Nvidia will release a new line-up that will completely shift the current market that effectively makes last-gen price/performance obsolete and establishing a new metric that will offer roughly +50% performance at the same price points. Again, mountains of historical evidence from both Nvidia and AMD back my point. This is what is expected from "Next-Gen" architectures on "Next-Gen" fabrication processes.
What AMD is doing here is cashing in short-term profits but ignoring long-term repercussions. As I stated in another comment, the people most likely to buy this product are AMD's most devout and loyal fans. IF they have to drop the pricing on these Tahiti parts because they were forced to so shortly after launch as a result of Nvidia's Kepler price/performance, how do you think these early adopters are going to feel? Their biggest fans are going to feel the biggest burn.
There is precedence for this with the GTX 280 launch. Nvidia did right by their customers by issuing rebate checks for $100-150 per card. Do you think AMD is willing to do the same? Just something to consider.
Sabresiberian - Tuesday, January 31, 2012 - link
People like you love to look at benchmark results that support their statements, and ignore the rest of them.Your statement is no more accurate than the statements of those that say the 7970 is barely faster than the GTX 580.
;)
swx2 - Thursday, February 2, 2012 - link
Are you listening to your self? did you just say that a overclocked 7970 (current gen card) is JUST NOW competitive with a last gen card? And you think that AMD has done well with this accomplishment?...what-is-this-i-don't-even...
Iketh - Tuesday, January 31, 2012 - link
Ladies and gentlemen, drugs are bad for you.