An asynchronous and parallel row-wise compressed SpMV kernel on heterogeneous CPU-GPU architectures.
Huachen TanPublished in: Int. J. Embed. Syst. (2022)
Keyphrases
- heterogeneous computing
- graphics processing units
- compute intensive
- general purpose
- parallel computing
- parallel architectures
- parallel processing
- parallel implementation
- parallel programming
- single instruction multiple data
- parallel computation
- gpu implementation
- real time
- massively parallel
- parallel computers
- computing systems
- memory bandwidth
- graphics hardware
- multi core processors
- pc cluster
- asynchronous cellular automata
- heterogeneous systems
- compute unified device architecture
- graphics processors
- grid computing
- shared memory
- floating point
- pairwise
- kernel function
- data structure
- multicore processors
- processing units
- computer architecture
- multithreading
- data transfer
- level parallelism
- distributed memory
- support vector
- efficient implementation
- kernel methods
- feature space
- parallel execution
- cluster of workstations
- multi threaded
- parallel hardware
- parallel algorithm
- parallel processors
- multiple kernel learning
- high performance computing