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
Shadowmaster625 - Tuesday, January 31, 2012 - link
Well some of us do care about noise and power consumption. That sapphire card runs very cool and quiet, nearly silent in fact. It uses 30 less watts at typical idle and nearly 100 watts less than a 580 while gaming. That's about a penny per hour of gaming. Power savings could easily reach $40 over two years of gaming, plus another $30 if you leave your pc on 24/7, for a total of $70 saved over two years.Thats is about what one year of technological advancement is worth.
SlyNine - Tuesday, January 31, 2012 - link
Because in part the 480/580 are dogs in regards to power consumption. My 5870 idles at nearly the same and only uses a bit more under load. It is also 2 1/2 years older.AssBall - Tuesday, January 31, 2012 - link
You do realize that the parts you have "execs jumping out of windows" about account for maybe 1/200th of the companies income? Guess not.chizow - Tuesday, January 31, 2012 - link
Actually I think these parts will make up an overwhelming % of the company's income given their graphics division is one of the few within AMD that consistently turns a profit. But I guess I actually know what "income" is and how its calculated. Given these parts sell for 50% higher price than any single-ASIC SKU (CPU or GPU) in the last 5 years, I think there's a good chance these parts will make up the bulk of their profit for the quarter.You really think AMD is turning some amazing profit on their free AR Bulldozer promos or happy meal priced APUs when they struggle to make a profit on a quarterly basis? There's break-even (90-95% of their revenue) and then there's parts like this that actually sell for a profit. They don't need to sell high volume on these parts when the ASP and margins are so healthy.
repoman27 - Tuesday, January 31, 2012 - link
Actually, you should try reading AMD's Q4 earnings report which was just posted here: http://www.anandtech.com/show/5465/amd-q411-fy-201...The graphics division only accounted for 8.4% of AMD's operating income for fiscal 2011, if you ignore the accounting monkey business regarding GloFo.
Of that 8.4%, about half is likely attributable to FirePro because the gross margins for professional and compute solutions are easily double that of consumer GPUs.
While the ASP's and gross margins on enthusiast cards are great compared to the high volume consumer stuff, they still represent a niche market, and realistically less than 3% of AMD's 2011 operating income.
chizow - Wednesday, February 1, 2012 - link
I have read it and it obviously doesn't include any significant Tahiti or 28nm parts, since they only started to ship the first few units for revenue for that Jan 7th 7970 launch.As stated in the article:
"while the first true 7000 series part (Tahiti) did not launch until 2012 and only started shipping for revenue very late into 2011. Still, it was enough to have a significant impact on AMD’s GPU ASP, increasing it over 2010’s ASP even with the limited number of new products."
Obviously projected revenue will be much higher with a full quarter at these prices.
But yes the rest of your post confirms my point, if you want to point to FirePro I can already guarantee you they will sell more 7970 consumer desktop parts than any FirePro parts using the same ASIC, we're not talking about Quadro here lol.
At $450-600, that's already a HUGE markup and increase in their gross margins compared to any previous single-ASIC desktop SKU, so its obviously going to result in a huge increase in profits.
And no you can't just discount that "GloFo" business, if the rest of the company loses money because of poor supplier/pricing decisions and selling prices and the other consistently makes money and represents the bulk of the company's income.
repoman27 - Wednesday, February 1, 2012 - link
Seriously, put down the crack pipe.Have you ever looked at the relative prices of consumer desktop parts vs. FirePro parts based on the same chip? Yes, they will sell a lot more of the consumer parts, and yet they will still make more profit on the fewer FirePro parts sold. So when AMD wholesales chips to partners so the partners can produce cards that end up retailing for $450-600, you believe their gross margins are higher than when they sell a FirePro V9800 directly to a retailer who turns around and sells it for $2800?
Why are you so convinced that the difference in retail release prices for the 7 series and 6 series is solely attributable to AMD expanding their gross margin? The 7 series is being offered for 50% more than the 6 series was, and yet the transistor count increased by over 62%. Not to mention that these are the first parts produced on the new 28nm node, so yields are probably less than stellar at this point, despite what TSMC might have us believe.
Furthermore, the GloFo business is germane to both the CPU and GPU divisions. Graphics accounted for less than a quarter of AMD's overall revenue for 2011 and only 8.4% of their income, no matter how you want to look at it. All I omitted was the one time charges that AMD took for the decrease in value of their stake in GloFo and the costs associated with their corporate restructuring.
No matter what you believe, 8.4 is a larger number than 91.6. Hang it up, you're flat out wrong.
chizow - Thursday, February 2, 2012 - link
No, you need to put down the "crack pipe."You're using past performance to define FUTURE revenue forecasts, which is what my statement was about. That's the whole point, the GPU division WAS disappointing in 2011, but ASP and profit increased year over year in Q4 based on the high ASP of the few Tahiti parts that shipped in December.
Obviously Tahiti parts will have a bigger impact on both numbers in Q1 '12 with a full 3 quarters and volume shipment at these high ASPs which is exactly what I stated.
As for why I think the difference in 7 and 6 series pricing was the difference in segment margins? Because I know how to read financial statements and MD&A. 11/12 months with DECLINING revenue and margins selling only 6 series and specifically cited by AMD caused by declining sales and ASP on their desktop and shortages of their Llano mobile parts. Then in Q4, despite declining revenue sequentially and y2y, their ASP, margins and profit increases. Do you think this is a result of their 6 series which spent the whole year declining in sales and margin, or the 7 series that sell for 50-100% more than their previous products? Hmmm.....
Now what do you think is going to happen with a full 3 quarters at the same prices for the graphics division? What do you think is going to happen to the company's financials given they are predicting a gloomy outlook with predicted 8% decrease in revenue sequentially? I can already guarantee you the graphics division will be profitable, maybe ~$100M, but the CPU division will at best break even if they don't post a loss. Which brings us full circle to the comment you and others apparently had issue with:
Actually I think these parts will make up an overwhelming % of the company's income given their graphics division is one of the few within AMD that consistently turns a profit.
repoman27 - Wednesday, February 1, 2012 - link
And clearly I meant to say that 8.4 is NOT a larger number than 91.6.Sunburn74 - Tuesday, January 31, 2012 - link
I think the power.consumption reductions are quite significant and eventually may be passed onto the mobile space.