Automatic Fusions of CUDA-GPU Kernels for Parallel Map
| Název česky | Automatické fúze CUDA-GPU kernelů pro paralelní map |
|---|---|
| Autoři | |
| Rok publikování | 2011 |
| Druh | Článek v odborném periodiku |
| Časopis / Zdroj | Computer architecture news |
| Fakulta / Pracoviště MU | |
| Citace | |
| Obor | Informatika |
| Klíčová slova | GPU;CUDA;kernels fusion;map;mapped function |
| Popis | When implementing a function mapping on the contemporary GPU, several contradictory performance factors affecting distribution of computation into GPU kernels have to be balanced. A decomposition-fusion scheme suggests to decompose the computational problem to be sol\-ved by several simple functions implemented as standalone kernels and to fuse some of these functions later into more complex kernels to improve memory locality. In this paper, a prototype of source-to-source compiler automating the fusion phase is presented and the impact of fusions generated by the compiler as well as compiler efficiency is experimentally evaluated. |
| Související projekty: |