Originally Posted by anothergeek
6000 vs 5000 memory, less throttling, and twice the VRAm. Should mean a lot.
The VRAM amount won't have a big effect on average FPS scores at all (might affect minimum FPS/frametimes).
As for the memory overclock; the 295X2 uses two full Hawaii cores with a CLC. Those extra shader cores balance out memory clock speed advantages with the more cut-down 390.
As for throttling, the 295X2 is using a CLC whereas two 390s would be using open-air cooling designs each expelling the same amount of heat as a single 980 Ti in a confined area. It's unlikely they would throttle less. Speaking as the owner of multiple open-air-cooled 290Xs, a 780, and a 980 Ti, two such power-hungry GPUs in such close proximity absolutely will throttle and/or get extremely hot. It's a lot of heat to expel into a case.
Also, that is an overall graph. An actual case of 90% efficiency like Firestrike? The crossfire would beat the living daylights out of the 980 ti.
Of course it's an average/overall graph (so is the OP list) since an average better reflects across the board performance allowing a better judge of bang-for-buck. It's only the 4K average graph too, which actually shows the 295X2 in its best light. It's actually behind in the other resolutions overall from that same site.
Are you directly saying we should cherrypick and falsely generalize a best-case scenario based on a synthetic benchmark?
And speaking of that synthetic benchmark (Firestrike), here are various 980 Tis all taking a dump on crossfire 290Xs and the 295X2:
Crossfire 390s absolutely do not stand a chance of doing what you claim.Edited by Serandur - 10/17/15 at 12:45pm