[AT] A Look At Triple-GPU Performance And Multi-GPU Scaling

[Gloomy]

For those who don't look at the main page :awe:

http://www.anandtech.com/show/4254/triplegpu-performance-multigpu-scaling-part1/1

Nah, just kidding. Discuss?

 

[Arkadrel]

In terms of average FPS gains for 2 GPUs, AMD has the advantage here. It’s not much of an advantage at under 10%, but it is mostly consistent. The same can be said for 3 GPU setups, where the average gain for a 3 GPU setup versus a 2 GPU setup nets AMD a 127% gain versus 121% for NVIDIA.

Minimum framerates with 3 GPUs does give us a reason to pause for a moment and ponder some things. For the games we do collect minimum framerate data for – Crysis and Battlefield: Bad Company 2 – AMD has a massive lead in minimum framerates.

looks good AMD CF has gotten much better than it used to be, only a year or two ago.

Man... reguardless of if you have 580x3 or 6970x3..... the power bills alone.... '

 

[Mistwalker]

AMD's 6000-series CF scaling continues to impress, while Nvidia is no slouch either.

I was a little disappointed at the lack of multi-monitor resolutions, so really glad to see they plan on a follow-up later this month. Also, if you're going to benchmark lower resolutions with so much GPU power then SSAA is the perfect way to do it, nice choice there.

 

[1h4x4s3x]

Nice :thumbsup:
No multi-monitor :thumbsdown:
Promised multi-monitor (3d) reviews in the future :thumbsup:

 

[AnandThenMan]

Interesting article, but without a multi-monitor test environment, the tests are not very useful. Using a tri-card system with a single monitor is pointless.

 

[Mopetar]

Interesting article, but without a multi-monitor test environment, the tests are not very useful. Using a tri-card system with a single monitor is pointless.

Looks like they're already on top of that (from the end of the article):

"Ultimately triple-GPU performance and scaling cannot be evaluated solely on a single monitor, which is why we won’t be stopping here. Later this month we’ll be looking at triple-GPU performance in a 3x1 multi-monitor configuration, which should allow us to put more than enough load on these setups to see what flies, what cracks under the pressure, and whether multi-GPU scaling can keep pace with such high resolutions. So until then, stay tuned."

 

[badb0y]

AMD has really improved their Crossfire scaling, that's for sure. Waiting on 28 nm now.

 

[nemesismk2]

interesting article, can't wait to read part 2

 

[n0x1ous]

enjoyed it as well. Would have liked to see a better i7 that could have been at 4ghz or so to better eliminate cpu bottlenek. Ryan mentioned the 920 they had on hand could only hit 3.3 ghz

 

[nanaki333]

i like how the tri-fire comes out on top a few times and it's half the price.

 

[Silverforce11]

enjoyed it as well. Would have liked to see a better i7 that could have been at 4ghz or so to better eliminate cpu bottlenek. Ryan mentioned the 920 they had on hand could only hit 3.3 ghz

Agreed. CPU needs to be 4ghz+ for these tripple + setups.

AT should also discard some of the older games and move to newer titles. Such as Shogun 2 (an amazing AAA game that punishes GPUs) and DA2.

Games like Mass Effect/HAWX is rather pointless, midrange cards don't even break a sweat. But, i did appreciate the use of SSAA on those titles.

 

[Carfax83]

This review is pure fail.. 3.3ghz isn't enough to remove the CPU bottleneck from dual, let alone tripple card set ups....especially as Fermi relies heavily on CPU speed for max performance.


And don't tell me that one of the largest computer enthusiast websites on the internet couldn't find a CPU that could go north of 4ghz

If you can't do it properly, may as well not do it at all..

 

[WelshBloke]

Agreed. CPU needs to be 4ghz+ for these tripple + setups.

AT should also discard some of the older games and move to newer titles. Such as Shogun 2 (an amazing AAA game that punishes GPUs) and DA2.

Games like Mass Effect/HAWX is rather pointless, midrange cards don't even break a sweat. But, i did appreciate the use of SSAA on those titles.

Dude, too obvious

 

[SlowSpyder]

This review is pure fail.. 3.3ghz isn't enough to remove the CPU bottleneck from dual, let alone tripple card set ups....especially as Fermi relies heavily on CPU speed for max performance.


And don't tell me that one of the largest computer enthusiast websites on the internet couldn't find a CPU that could go north of 4ghz

If you can't do it properly, may as well not do it at all..

I still see scaling with more GPU's/faster GPU's so I don't know how much the CPU is really holding the cards back.

 

[blastingcap]

Interesting article, but without a multi-monitor test environment, the tests are not very useful. Using a tri-card system with a single monitor is pointless.

Well, to be fair, 2560x1600 at 120Hz/3D is stressful too, but yeah, I'd like to see 3x1050p and 3x1080p tests, but if Crysis is already running into the memory wall with one monitor (in tri-SLI), then that doesn't bode well for 1GB or 1.5GB cards. 2GB cards may do somewhat better.

 

[rgallant]

NVIDIA ForceWare 266.58 -I don't get it