OpenMP Code Offloading: Splitting GPU Kernels, Pipelining Communication and Computation, and Selecting Better Grid Geometries.
Artem ChikinTyler GobranJosé Nelson AmaralPublished in: WACCPD@SC (2018)
Keyphrases
- parallel computation
- parallel programming
- parallel processing
- parallel computing
- fine grain
- graphics processing units
- shared memory
- source code
- support vector
- real time
- parallel implementation
- kernel function
- communication cost
- multiple kernel learning
- general purpose
- feature space
- communication technologies
- grid environment
- graphics hardware
- grid points
- computation intensive