Porr Scaling From 2x6970 to 3x6970?!?!

[pugheaven]

6512 With 2x6970's

V's

6959 With 3x6970's

??!!!

[Jarnis]

Crossfire with more than 2 GPUs has known problems (if you look carefully, you see some funky results out of Combined Test). We've told AMD. It is a driver issue.

Even without that the scaling wouldn't be 100% unless you have a very very fast CPU. Q6600... isn't. In fact, in your case the main issue is that you are CPU limited.

[footman]

Bet that third GPU all but eliminates any micro stutter though <grin>

[pugheaven]

Crossfire with more than 2 GPUs has known problems (if you look carefully, you see some funky results out of Combined Test). We've told AMD. It is a driver issue.

Even without that the scaling wouldn't be 100% unless you have a very very fast CPU. Q6600... isn't. In fact, in your case the main issue is that you are CPU limited.

The problem I always have with this argument with CPU bound, Heaven Benchmark gives 100% scaling with each GPU at 98+% all the time... so for 3DMark11 GPU tests not to scale, isn't too good... it's not as though there's CPU/GPU like one of the tests
we do... just very dissapointing as all other things scale nicely, BF3, Dirt 3... yet this benchy runs like a pile of poo with 3 cards. Hopefully that driver issue will sort out.

I do agree games like BF3 in MP mode are CPU bound, which then slows the GPU's down but still.

Zero microstutter on this system :lemo:

[Jarnis]

It does scale on modern CPUs. Q6600 just isn't quite cutting it any more with 3 cards.

[Meaker]

If you are running the P benchmark then that's a little silly. Try it on X.

[pugheaven]

If you are running the P benchmark then that's a little silly. Try it on X.

Did, weas getting 4k on X with 6990, when running 6990 and 6970 got 3500 lol