Table 6.
Methods | Parallelization algorithm | Acceleration (CPU/GPU) |
---|---|---|
Parallelization of the direct method | Hybrid | 16 |
Memory access optimization | Hybrid | 50 |
Coarse-grained | 60 | |
Asynchronous data transfer | Coarse-grained | 90 |
Data compression | Coarse-grained | 130 |