AMD's Radeon HD 6990: The New Single Card King
by Ryan Smith on March 8, 2011 12:01 AM EST- Posted in
- AMD
- Radeon HD 6990
- GPUs
Compute Performance
Moving on from our look at gaming performance, we have our customary look at compute performance. With AMD’s architectural changes from the 5000 series to the 6000 series, focusing particularly on compute performance, this can help define the 6990 compared to the 5970. However at the same time, neither benchmark here benefits from the dual-GPU design of the 6990 very much.
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.
New as of Catalyst 11.4, AMD’s performance in our Civilization V DirectCompute benchmark now scales with CrossFire at least marginally. This leads to the 6990 leaping ahead of the 6970, however the Cayman architecture/compiler still looks to be sub-optimal for this test. The 5970 has a 10% lead even with its core clock disadvantage. This also lets NVIDIA and their Fermi architecture establish a solid lead over the 6990, even without the benefit of SLI scaling.
Our second GPU compute benchmark is SmallLuxGPU, the GPU ray tracing branch of the open source LuxRender renderer. While it’s still in beta, SmallLuxGPU recently hit a milestone by implementing a complete ray tracing engine in OpenCL, allowing them to fully offload the process to the GPU. It’s this ray tracing engine we’re testing.
There’s no CrossFire scaling to speak of in SmallLuxGPU, so this test is all about the performance of GPU1, and its shader/compute performance at that. At default clocks this leads to the 6990 slightly trailing the 6970, while overclocked this leads to perfect parity with it. Unfortunately for AMD this is a test where NVIDIA’s focus on compute performance has really paid off; coupled with the lack of CF scaling and even a $240 GTX 560 Ti can edge out the $700 6990.
Ultimately the take-away from this is that for most desktop GPU computing workloads, the benefit of multiple GPU cores is still unrealized. As a result the 6990 shines as a gaming card, but is out of its element as a GPU computing card unless you have an embarrassingly parallel task to feed it.
Facebook
Twitter
RSS
130 Comments
View All Comments
james.jwb - Tuesday, March 8, 2011 - link
I doubt it. That would be 80 dBA at ear level compared to whatever Ryan used. At ear level it's going to be a lot lower than 80.Still, that doesn't take away the fact that this card is insane...
Belard - Tuesday, March 8, 2011 - link
The card needs 3 slots to keep cool and such. They should have made a 2.5-slotted card, but with a bit of a twist.Channel the AIR from the front GPU chamber into a U-duct, then split into a Y that goes around the fan (which can still be made bigger. The ducts then exhaust out the back in a "3rd slot". Or a duct runs along the top of the card (out of spec a bit) to allow the fan more air space. It would add about $5 for more plastic.
Rather than blowing HOT air INTO the case (which would then recycle BACK into the card!
OR - blowing HOT air out the front and onto your foot or arm.
Noise is a deal killer for many people nowadays.
strikeback03 - Tuesday, March 8, 2011 - link
That ductwork would substantially reduce the airflow, making a sharp turn like that would be a large bottleneck.burner1980 - Tuesday, March 8, 2011 - link
I´m always wondering why reviews always neglect this topic. Can this card run 3 monitors @ 1920x1080p 120 HZ. 120 HZ monitors/beamer offer not only 3D but foremost smooth transitions and less screen tearing. Since this technique is available and getting more and more friends, I really would like to see it tested.Can anybody enlighten me ? (I know that Dual link is necessary for every display and that AMD had problems with 120 HZ+eyefinity) Did they improve?
silverblue - Tuesday, March 8, 2011 - link
...with two slightly downclocked 6950s. Alternatively, a 6890, albeit with the old VLIW5 shader setup. As we've seen, the 5970 can win a test or two thanks to this even with a substantial clock speed disparity.The 6990 is stunning but I couldn't ever imagine the effort required to set up a machine capable of adequately running one... and don't get me started on two.
Figaro56 - Tuesday, March 8, 2011 - link
I'd love to see Jessica Alba's beaver, but that aint going to happen either.qwertymac93 - Tuesday, March 8, 2011 - link
AUSUM SWITCHhttp://img202.imageshack.us/img202/8717/ausum.png
KaelynTheDove - Tuesday, March 8, 2011 - link
Could someone please confirm if this card supports 30-bit colour?Previously, only AMD's professional cards supported 30-bit colour, with the exception of the 5970. I will buy either the 6990 or Nvidia's Quadro based on this single feature.
(Because somebody will inevitably say that I don't need or want 30-bit colour, I have a completely hardware-calibrated workflow with a 30" display with 110% AdobeRGB, 30-bit IPS panel and the required DisplayPort cables. Using 24-bit colour with my 5870 CF I suffer from _very_ nasty posterisation when working with high-gamut photographs. Yes, my cameras have a colour space far above sRGB. Yes, my printers have it too.)
Gainward - Tuesday, March 8, 2011 - link
Just a heads up for anyone buying the card and wanting to remove the stock cooler.... There is a small screw on the back that is covered by two stickers with (its under the two stickers that look like a barcode). Well removing that you will then notice a void logo underneath it... I just wanted to point it out to you all...Didnt bother us too much here seen as ours is sample but I know to some droppin £550ish UK is quite a bit of cash and if all you are doing is having an inquisitive look it seems a shame to void your warranty :-S
mmsmsy - Tuesday, March 8, 2011 - link
I'd like to know how do you benchmark those cards in Civ V. I suppose it's the in-game benchmark, isn't it? Well, I read some tests on one site using their own test recorded after some time spent in the game using FRAPS and I'm wondering if using the in-game test is that really different scenario. According to the source, in the real world situation nVidia cards' performance show no improvement whatsoever over AMD's offerings. If you could investigate that matter it would be great.