160 likes | 252 Views
This study explores GPU acceleration for matrix calculations with Quad CPU optimization for enhanced performance. Analysis includes Grid and Block size configurations for different GPUs and CPUs. Matlab version 2009b used.
E N D
0 1 1 2 3 2 3 0 3 2 4 5 N+1 Nmax Nmax b 1 3 4 0 2 A c 7 block1 1 1 0 X 3 2 block2 X 3 0 Jc Ir Pr
a cs b Thread1 c1 × Block c2 Thread2 × sum_1 cn Threadn × Thread1 c1 × Block Thread2 Result c2 sum_2 × cn Threadn × Block × ck Threadk sum_n ... ×
c b A c1 X c2 c3 c4 c5 c6 c7 ...
0 1 2 3 4 0 1 1 2 2 3 3 3 0 4 2 5 N+1 Nmax Nmax b A c 7 BLOCK1 1 1 0 X 3 2 BLOCK2 X 3 0 Jc Pr Ir
Dreieckförmige Summation Iterationsschritte
(ms) Matrix: 100000x100000 1 Diagonale GPU:GTX260 Grid size:1024
Matrix: 100000x100000 32 Diagonale GPU:GTX260 Grid size:1024
Matrix:5000x5000 Quad CPU: Q6700@2.66GHZ RAM:3.25GB GPU:GTX260 Grid size:1024 Block size: 16x16 Matlab version: 2009b
Matrix:5000x5000 Quad CPU: Q6700@2.66GHZ RAM:3.25GB GPU:GTX260 Grid size:1024 Matlab: v. 2009b
BlkY\BlkX 1 16 32 64 128 256 512 1 2.41 2.714 3.77 6.98 14.56 39.8 16 0.3 0.93 2.45 32 0.3 1.33 64 0.36 128 0.65 256 0.67 512 0.71 (ms) Matrix: 100000x100000 1 Diagonale GPU:GTX260 Grid size:1024
Matrix: 100000x100000 32 Diagonale GPU:GTX260 Grid size:1024
Quad CPU: Q6700@2.66GHZ RAM:3.25GB GPU:GTX260 Grid size:1024 Block size: 8x64
Quad CPU: Q6700@2.66GHZ RAM:3.25GB GPU:GTX260 Grid size:1024 Block size: 8x64