Tesla C2050 vs Geforce 480 GTX performance of the cuFFT

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi guys,

I’ve got a strange behaviour of the cuFFT on the Tesla C2050. This GPU is supposed to be x4 faster than the 480 GTX on double precision computation. Actually, the 480 GTX is 30% fatser than C2050.Any explanation on this, knowing that the price is x5 higher??

Many Thanks

2

Your problem might be memory bandwidth limited, not (double precision) compute limited. That happens more often than one might think, particularly as double precision numbers also take up twice the memory.