Hi Andrea, On Thu, Jul 19, 2012 at 2:39 AM, Andrea Cesari <[email protected]> wrote: > Hi, this is my code that, keep a 3d array, and for each pixel of the matrix > find the minimum and put it to the corresponding pixel of a matrix b. Then > compare the result with the cpu. > Obviously, with grid=(1,1) the max dimensions of the matrix can be 32x32 > (1024 threads per block; GTX 580). > So, because i want a matrix of (500,500) , i I thought to declare a > block=(32,32,1) and grid=(16,16). But it doesn't work. I noted that it work > only with square matrix and with dimensions minor of 32.
Could you elaborate on that? What exactly does not work? The code you provided produces correct results. _______________________________________________ PyCUDA mailing list [email protected] http://lists.tiker.net/listinfo/pycuda
