20 likes | 206 Views
Nodari Sitchinava KIT. Volker Weichert Goethe University. Provably Efficient GPU Algorithms. Block 0. Block B-1. Warp 0. Warp 0. Warp 1. Warp 1. Warp C-1. Warp C-1. Block 0. Block C-1. Threads. Threads. Shared memory. Shared memory. Global Memory. CPU. Kernel 0. Kernel 1.
E N D
Nodari SitchinavaKIT Volker WeichertGoethe University Provably Efficient GPU Algorithms Block 0 Block B-1 Warp 0 Warp 0 Warp 1 Warp 1 Warp C-1 Warp C-1 Block 0 Block C-1 Threads Threads Sharedmemory Sharedmemory Global Memory CPU Kernel 0 Kernel 1 Kernel R-1 A[0] A[S] A[2S] Memory Bank 0 A[1] A[S+1] A[2S+1] Memory Bank 1 A[S-1] A[2S-1] A[3S-1] Memory Bank M-1 MADALGO – Center for Massive Data Algorithmics, a Center of the Danish National Research Foundation