I was kinda lucky to have two 560Ti cards at my disposal. One straight from MSI for this review and one for a friends future build. Hence why the SLI tests were only done on air with unmodded cards ( read : no voltage modifications ). This to avoid the loss of a 190 euro expensive card and maybe a friendship :p.
So we only used the official MSI Afterburner and not the Extreme version ( which has got a higher voltage selection for the GPU ) locking the vGPU at 1125mV. Both cards were cooperating nicely at above 1000mhz for the GPU and +/-1200Mhz for the rams. As CPU we are using one of 2600K Sandy Bridge CPUs at 5600Mhz on a MSI Big Bang Marshall board. Sadly for our testing, the 2133Mhz ram divider was buggy and we kept the rams at the 1866 ram ratio divider. For real world applications the difference between 1866 and 2133Mhz ram speed is minimal, though when benching it can mean the difference between top 3 or top 10.
Here are the results :
3DMark 03 1st spot :
3DMark 05 3rd spot (missing some serious CPU ooomph)
3DMark06 5th spot :
3Dmark 011 1st spot :
Aquamark 1st spot :
Where's the 01 and Vantage score you might wonder. Well for 3DMark01 you really need a Gigabyte motherboard, which scores easily up to 3000-4000 points more with the same clocks. Secondly our Windows was getting a bit fubarred from the numerous lockups and Vantage wouldn't even run at stock clocks lol. So we called it a day... but really nice performance of air cooled out of the box reference cards. A small voltmod and some watercooling will make the above scores rapidly redundant...
AM3: 383149 1035/1200
3DMark01SE: 103431 1040/1200
3DM03: 92282 1010/1200
3DM05: 47079 1010/1200
3DM06: 35904 1000/1200
3DM011E: 9140 1000/1200
3DM011P: 5633 1000/1200
3DM011X: still to do
Vantage: still to do
Unigine DX9: 2937.27 1000/1200
Unigine DX11: still to do
Voltage is between 1.11V and 1.135V
Cooling is stock Twin Frozr II air cooler
Seems Albrecht's card can clock a bit better on air.