HyperSLI vs SLI certified configuration:
Since I've been using Hypersli since version 0.6, I decided to test my performance loss/gain because there must be some sort of frames per second loss due to the 'hack'.
My basic system configuration:
i5 2500K - OC 4.5GHz
8GB DDR2000
Gainward 560ti Phantom 2 x 2
My friend's configuration:
i7 3770K - OC 4.0GHz
8GB DDR1600
Geforce 570 x 2 (Gainward + Gigabyte)
These 2 systems is close enough to make a comparison.
We have only tested Heaven DX11 Benchmark 2.5 and 3DMark 2003 (yes it's not SLI compatible).
The 2 x Geforce 570's outperformed my 2 x Geforce 560ti's (off course)
We swapped his 570's with my 560ti's and ran the benchmarks again.
SOMETHING INTERESTING:
DX11 Benchmark 2.5 is SLI compatible, so the 2x570's outperformed my 560ti's
3DMark 2003 is NOT SLI compatible and my 560ti's outperformed his 570's by FAR! And I mean really far.
When we checked MSI afterburner, my PC running HyperSLI used both GPU's during benchmark, where his genuine SLI setup only used 1 GPU.
Conclusion:
There is a +-10-12% performance loss using HyperSLI (or its just because his genuine SLI motherboard has a 8x PCIe slot for the second GPU, where my board only have a x4 PCIe slot)
But with HyperSLI my PC utilize both cards with non-SLI games/applications (like 3DMark 2003, etc)
And before everybody ask me to post figures/screenshots, it was done a couple months ago...so it will be hard to find
