Normalized Clocks: Separating Architecture & SMs from Clockspeed Increases

While we were doing our SLI benchmarking we got several requests for GTX 580 results with normalized clockspeeds in order to better separate what performance improvements were due to NVIDIA’s architectural changes and enabling the 16th SM, and what changes are due to the 10% higher clocks. So we’ve quickly run a GTX 580 at 2560 with GTX 480 clockspeeds (700Mhz core, 924Mhz memory) in order to capture this data. Games that benefit most from the clockspeed bump are going to be memory bandwidth or ROP limited, while games showing the biggest improvements in spite of the normalized clockspeeds are games that are shader/texture limited or benefit from the texture and/or Z-cull improvements.

We’ll put 2 charts here, one with the actual framerates and a second with all performance numbers normalized to the GTX 480’s performance.

Games showing the lowest improvement in performance with normalized clockspeeds are BattleForge, STALKER, and Civilization V (which is CPU limited anyhow). At the other end are HAWX, DIRT 2, and Metro 2033.

STALKER and BattleForge hold consistent with our theory that games that benefit the least when normalized are ROP or memory bandwidth limited, as both games only see a pickup in performance once we ramp up the clocks. And on the other end HAWX, DIRT 2, and Metro 2033 still benefit from the clockspeed boost on top of their already hefty boost thanks to architectural improvements and the extra SMs. Interestingly Crysis looks to be the paragon game for the average situation, as it benefits some from the arch/SM improvements, but not a ton.

A subset of our compute benchmarks is much more straightforward here; Folding@Home and SmallLuxGPU improve 6% and 7% respectively from the increase in SMs (theoretical improvement, 6.6%), and then after the clockspeed boost reach 15% faster. From this it’s a safe bet that when GF110 reaches Tesla cards that the performance improvement for Telsa won’t be as great as it was for GeForce since the architectural improvements were purely for gaming purposes. On the flip side with so many SMs currently disabled, if NVIDIA can get a 16 SM Tesla out, the performance increase should be massive.

GTX 580 SLI: Setting New Dual-GPU Records
Comments Locked

82 Comments

View All Comments

  • Oxford Guy - Friday, November 12, 2010 - link

    There is already more than once source.

    Anandtech's article shows the 480 beating the 580 in a minimum frame rate bench. I believe the game is Crysis, but I'm not sure.

    So, coupled with the review I posted from techgage -- there are already multiple sources.

    We need to see Anandtech show minimum frame rates for Unigine.
  • wtfbbqlol - Saturday, November 13, 2010 - link

    Hmm, don't see it. I scanned the Anandtech GTX580 review for it but couldn't find it.

    BTW I'm just getting a slight sense that you are more than a little worried about this...
  • Oxford Guy - Saturday, November 13, 2010 - link

    Anandtech only posted minimum frame rates for Crysis. In those posts, though, we see that the 480 SLI beat the 580 SLI at 1920 x 1200.

    http://images.anandtech.com/graphs/graph4008/33816...

    That is part of the pattern shown more obviously by the Unigine benchmark results I posted.

    Anandtech needs to post minimum frame rates for Unigine at least to resolve this issue.

    As for me being "worried"... I fail to see your point. If you don't think the 480 crushing the 580 in Unigine at lower than 2K res is worth investigating, that's your point of view.
  • B3an - Saturday, November 13, 2010 - link

    It's less than a 2 FPS difference... who cares?

    The 6870 and 5850 both beat the 5870 in that chart, which is also faster. Things like this are not that rare. It's nothing. Probably drivers.
  • wtfbbqlol - Saturday, November 13, 2010 - link

    Ok I see you were referring to the SLI results. I looked in the single-GPU review.

    Anyway, 2fps out of 55fps is kinda within a margin of error. And even in that Crysis minimum framerate graph, the single GPU GTX580 is faster than the single GPU GTX480. It looks like all the dual-GPU cards are started to get bottlenecked by the CPU/non-GPU parts of the system which is why the results are so clumped together.

    As for the Techgage results I am more than a little skeptical. As explained before there is no reason why a GTX480 can beat a GTX470 twofold, let alone GTX580. That super high minimum framerate may very well be a bug. It's the only video card in Techgage's graph that's far away faster than anything else. Doesn't make sense at all.

    And I believe you are doing a disservice by being so fixated on this minimum framerate issue without applying basic statistical analysis to the problem.
  • DigitalFreak - Wednesday, November 10, 2010 - link

    OMG ROFLMAO! You posted the same thing in the original 580 review thread and it's still funny as hell! :-/

    Retard
  • Oxford Guy - Friday, November 12, 2010 - link

    Do you want to post links to the minimum frame rate Unigine Heaven tests you've done?

    You, know... something productive?

    No. Then move along.
  • jewie27 - Tuesday, April 3, 2012 - link

    Wrong question. Can it play Battlefield 3 in 3D Vision Surround?
  • piroroadkill - Wednesday, November 10, 2010 - link

    No 5970 CF?
  • tech6 - Wednesday, November 10, 2010 - link

    It must be a conspiracy! ;-)

Log in

Don't have an account? Sign up now