Tuning basic Linear Algebra Routines for Hybrid CPU+GPU Platforms.
Gregorio BernabéJavier CuencaLuis-Pedro GarcíaDomingo GiménezPublished in: ICCS (2014)
Keyphrases
- linear algebra
- graphics processing units
- graphics processors
- gpu implementation
- computer architecture
- image processing
- singular value decomposition
- sparse matrices
- data transfer
- image segmentation
- parallel implementation
- parallel processing
- numerical methods
- parallel computing
- high performance computing
- parallel algorithm
- image reconstruction
- least squares
- preprocessing