Is it normal that a CUDA application gives good results in debug mode and the release version appears to have syncronization problems? What can be the reason for that? I’m working with Visual Studio 2010 on a Tesla GPU. When I say debug mode, I mean C code debugging I’m not using paralell NSIGHT to debug de GPU, I’m only stopping the execution at the end, when everything is already done on the GPU.
Thank you in advance for your help.