pinhodecarlos
11-25-12, 02:47 PM
1) Can someone enlighten me how much work is done on a GPU wu ( POEM++ OpenCL version) when compared to the CPU ones (POEM++)? I would like a response like GPU wu is X times larger than CPU wu or 1 GPU wu is equal to X CPU wu. X is the variable.
2) How many cores are necessary to feed a nvidia GTX 680 and an ATI HD 7970 to keep them at least at ~95 %? How many GPU wu's must run in parallel to reach 95 % of GPU usage?
3) Timings for a default GTX 680 and an ATI HD 7970 running a POEM++ OpenCL wu.
Thank you in advance,
Carlos
2) How many cores are necessary to feed a nvidia GTX 680 and an ATI HD 7970 to keep them at least at ~95 %? How many GPU wu's must run in parallel to reach 95 % of GPU usage?
3) Timings for a default GTX 680 and an ATI HD 7970 running a POEM++ OpenCL wu.
Thank you in advance,
Carlos