Solving knapsack problems on GPU |
| |
| Authors: | V Boyer D El Baz M Elkihel |
| |
| Affiliation: | a CNRS, LAAS, 7 avenue du Colonel Roche, F-31077 Toulouse, France
b Université de Toulouse, UPS, INSA, INP, ISAE, LAAS, F-31077 Toulouse, France |
| |
| Abstract: | 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 (1.4 GHz) 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 3.0 GHz. 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. |
| |
| Keywords: | Combinatorial optimization problems Dense dynamic programming Parallel computing GPU computing CUDA |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
