NVIDIA's GeForce GTX 580: The SLI Update
by Ryan Smith on November 10, 2010 10:00 AM ESTNormalized 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.
Facebook
Twitter
RSS
82 Comments
View All Comments
maverick7614 - Wednesday, November 10, 2010 - link
Hi guys,Thanks for the great review of the GTX580 from yesterday!
I know this is a bit off topic regarding GTX but as someone that just received as a gift a GTX480 recently, and has no space in the case to use a 3 slot air cooling solution, I was wondering if it would be possible to fit the new GTX580 cooling solution on a GTX480.
If this would be possible it would be great if you guys could retest the temperatures, noise and power draw for such a modded GTX480.
Regards,
Alex
slickr - Wednesday, November 10, 2010 - link
I see no reason why you would want to keep CIV5 in the benches, since its obviously not reliable to test graphic card speeds.I would like you to add another strategy game in its place, something like Total war series, you know with large scale combat and all that good stuff.
And why isn't there no Starcraft 2 benchmark there, its basically one of two PC only titles and yet you don't include it?
smookyolo - Wednesday, November 10, 2010 - link
I say they benchmark Age of Empires.The first one.
8D
Exodite - Wednesday, November 10, 2010 - link
While it's clear from this review, as well as others, that the GTX 580 is a significant improvement over the GTX 480 in power usage, temps and noise I realized that what I've actually taken away from reading it is how totally awesome the Radeon 6870 is. In CrossFireX especially.I reckon that's not quite what Nvidia had in mind but meh, it seems the absolute extreme-end cards still aren't a winning proposition.
silverblue - Wednesday, November 10, 2010 - link
Aside of the odd issue with minimum frame rates, Barts is a notable improvement over Cypress for Crossfire. If I was to paraphrase a quote taken from the 460 launch, this would be Evergreen done "right", at least in terms of multiple graphics card setups - better scaling and better power usage essentially means less waste, and it's not as if Evergreen was hungry to start with. Just a shame that AMD has effectively limited Barts to two cards at once.As for the 580, very well done, however I think we're going to have to wait for another redesign before power usage is properly tackled.
spiked_mistborn - Wednesday, November 10, 2010 - link
It would be nice if you could include 5770 CF scores in the future. I have a Sapphire Vapor-x running at 960 / 1350 and it's very quiet and performs ok, and I could get another one for about $140 on Newegg. The 5770 has similar or faster tesselation performance to that of 5870 (due to faster clock speed) and in CF has similar shading / texturing abilities. It would be nice to see how 2 of these at stock speeds compare to the other cards. Overall this is a great article as always!Folterknecht - Wednesday, November 10, 2010 - link
2x 5770 ~ 1x 5870Neomis - Wednesday, November 10, 2010 - link
What are the pc specs on this test system? it would be nice to know what processor/Clock limited some of these tests...shaggart5446 - Wednesday, November 10, 2010 - link
look how quick he do an sli and the 580 just release yesterday and when the 6870 and 6850 was release there wasnt any xfire the next day but he could also do an over clock 460 sometimes i wonder why amd even send u guys card to reviewhaukionkannel - Wednesday, November 10, 2010 - link
Well, most simple explanation between these different heat result is that the review part was cherry picked (like they normally are, so that you can get better overclockin results etc.) and the Asus one is normal version, that you can normally expect to get.When the production run gets better, I think that we will see more chips that are more like the early cherry picked one...