Hello,
I noticed, that the total speed varies widely from one simulation to another. I
have a Geforce GTX 770 and if I change the phantom or choose different
source/detector settings my speed varies from about 1 million x-ray/s to 13
million x-rays/s.
Could you explain this? Could I do any changes (e.g. in the Input-File) to
optimize the simulation speed? I've tried to change the "GPU-Threads per
Cuda-Block" and the "simulated histories per GPU-Thread", but I can only make
it slower.
Best regards
Jenny
Original issue reported on code.google.com by
jenny.ko...@t-online.deon 27 Jan 2015 at 5:56