Using hybrid GPU/CPU kernel splitting to accelerate spherical convolutions.
Paul M. SutterBenjamin D. WandeltFranz ElsnerPublished in: Astron. Comput. (2015)
Keyphrases
- graphics processing units
- convolution kernel
- graphics processors
- gpu implementation
- kernel function
- heterogeneous computing
- real time
- support vector
- general purpose
- kernel methods
- parallel computing
- graphics hardware
- data transfer
- feature space
- memory bandwidth
- gpu accelerated
- spherical harmonics
- unit sphere
- kernel regression
- parallel computation
- kernel matrix
- hybrid learning
- parallel implementation
- computer systems
- feature extraction