Lossy all-to-all exchange for accelerating parallel 3-D FFTs on hybrid architectures with GPUs.
Sébastien CayrolsJiali LiGeorge BosilcaStanimire TomovAlan AyalaJack J. DongarraPublished in: CLUSTER (2022)
Keyphrases
- parallel programming
- parallel architectures
- multi core processors
- multicore processors
- parallel processing
- multi core systems
- graphics processing units
- single instruction multiple data
- highly parallel
- parallel computing
- massively parallel
- shared memory
- computing power
- data compression
- parallel algorithm
- parallel computers
- general purpose
- parallel computation
- high end
- graphics hardware
- cloud computing
- operating system
- processing elements
- hybrid learning
- basis functions
- computing systems
- distributed memory
- interconnection networks
- parallel genetic algorithm
- computational power
- neural network