| Article ID: | iaor20115888 |
| Volume: | 39 |
| Issue: | 1 |
| Start Page Number: | 42 |
| End Page Number: | 47 |
| Publication Date: | Jan 2012 |
| Journal: | Computers and Operations Research |
| Authors: | Boyer V, El Baz D, Elkihel M |
| Keywords: | programming: dynamic |
A parallel implementation via CUDA of the dynamic programming method for the knapsack problem on NVIDIA GPU is presented. A GTX 260 card with 192 cores (1.4GHz) is used for computational tests and processing times obtained with the parallel code are compared to the sequential one on a CPU with an Intel Xeon 3.0GHz. The results show a speedup factor of 26 for large size problems. Furthermore, in order to limit the communication between the CPU and the GPU, a compression technique is presented which decreases significantly the memory occupancy.