122
edits
Changes
→V1 -- One array
[[File:pyonearray.PNG]]
int* d_pixs; cudaMalloc((void**)&d_pixs, N * N * 3 * sizeof(int)); kernel_tray << <nblocks, ntpb >> >(pix_col, N, d_pixs); cudaMemcpy(h_pixs, d_pixs, N * N * 3 * sizeof(int), cudaMemcpyDeviceToHost);
=== V2 -- Occupancy ===
