Compute & Tessellation Performance

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 benchmark.

While Turks and the 6670/6570 are quite a bit faster than the 6450, both are still so-so as compute cards. Although the unit of measurement (fps) is effectively arbitrary, what isn’t arbitrary is that the 5750 is nearly 33% ahead of the 6670. Juniper (5750/5770) seems to be the floor for good compute performance on AMD cards.

Our second compute benchmark of the day 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. AMD in turn recently hit a milestone by offering the OpenCL runtime with their drivers, so SmallLuxGPU will now run out of the box without the AMD Compute SDK.

At 3200K rays/sec, the 6670 actually beats the Radeon HD 4870 by about 10%, showcasing just how far AMD’s compute performance has come compared to the 4000 series. The 6670 also enjoys a similar lead over the 5670 thanks to its additional shaders, but the 720SP 5750 looms large above it.

At the other end of the spectrum from GPU computing performance is GPU tessellation performance, used exclusively for graphical purposes. Barts’ tessellation improvements are inherited by Turks, which should give it an edge over the 5670.

What ended up surprising us the most with our DX11 Detail Tessellation Sample test wasn’t so much the raw performance of the 6670, it was the performance relative to some other AMD cards. At a high tessellation factor the Radeon HD 5870 isn’t all that far ahead, and meanwhile the 5750 is edged out by the 6670 in all cases. Given that our most tessellation-intensive games struggled at 1680x1050 on the 6670 it may not be powerful enough to really use its tessellation abilities in full on top of a heavy graphical workload, but the basis for respectable tessellation performance is there.

Wolfenstein Power, Temperature, and Noise
Comments Locked

53 Comments

View All Comments

  • ClagMaster - Tuesday, April 19, 2011 - link

    The HD6670 is a nice upgrade for a 9600GT, both of which are 65W cards.

    I beleive, based on 240GT performance, that the HD6670 has about 40-50% more performace over the 9600GT.

    This card has plenty of performance for games published 4 years ago.
  • arthur449 - Tuesday, April 19, 2011 - link

    Last page, middle of fourth paragraph:

    "... the Radeon HD 5770 and GeForce GTS 450 are both regularly on sale for under $100 and are easily 30% faster than the 6770."
  • Shadowmaster625 - Tuesday, April 19, 2011 - link

    Newegg has a sapphire 5750 for $103 AR

    http://www.newegg.com/Product/Product.aspx?Item=N8...

    Best part is it can apparently be flashed into a 5770.

    A 5770 curbstomps a 6670, by more than 50%.
  • Jasker - Tuesday, April 19, 2011 - link

    Finally an actual comparison using an older card. I've been using my 8800 GTS 512 for years and yet to really see a game I could determine for sure was GPU bound (I have an old dual Operaton CPU). Thanks to this article I know I can almost definately see some games with todays mid range cards. Thanks!
  • mosox - Tuesday, April 19, 2011 - link

    you don't include the HD video quality benchmark but a load of TWIMTBP games instead.
    Read Tomshardware for a decent review.
  • Spivonious - Tuesday, April 19, 2011 - link

    Agreed. These cards are not marketed towards gamers. They will more likely be used for media playback and casual gaming, not running Crysis.
  • SlyNine1 - Wednesday, April 20, 2011 - link

    No, I wouldn't buy a 6670 for someone thats just going to use it for media playback. There are much better options for that.
  • casteve - Tuesday, April 19, 2011 - link

    "The 5570 was the ultimate HTPC card, but the 6450 has dethroned it. "

    Except Ganesh is still in hiding.....so, you can't say it definatively.

    I was going to say, these aren't gaming cards. Then I wandered over to the March 2011 Steam Hardware survey and found ~29% of users had 1280x1024 or less monitors. Yow.

    The other humorous factoid was how the 6670 was the equivalent of the venerable 8800 GT; using 30W less in idle and half the power under load. Sort of a nice data point for progress.

    But, the biggest missing piece to this article is the analysis for HTPC. How well do the 65xx/66xx score versus the 55xx/56xx/etc? How well do they implement 24fps?

    Finally, Adobe is using gpu acceleration for some Photoshop/Lightroom/etc features. It would be nice (instead of Adobe's vague 'you can use your gpu card to accelerate' premise) to have some benchmarks for these lower tier cards to see what they actually do and whether a bottom end card is all you truly need for these apps.
  • Belard - Tuesday, April 19, 2011 - link

    Agreed.

    Thing is, these new cards are quite usable for people on a budget - for gaming, or to do some gaming in an HTPC, which the 6450 cannot do. I still play games on my rather old 4670.

    The video quality and feature set is what makes the 6000 series better than the 5000 and of course the GeForce cards which have severe HTPC / video output issues.
  • Spivonious - Tuesday, April 19, 2011 - link

    Can the charts with noise levels please include the noise floor of the room, or of the system without the video card installed (which I'm assuming is around 41dBA in this case)? Also, can the distance of measurement be included? I would love to see measurements taken at two distances - one next to the case and one about 3-5 feet away to get the level at a normal seating distance.

Log in

Don't have an account? Sign up now