Performance comparison of Eulerian kinetic Vlasov code between flat-MPI parallelism and hybrid parallelism on Fujitsu FX100 supercomputer.
Takayuki UmedaKeiichiro FukazawaPublished in: EuroMPI (2016)
Keyphrases
- massively parallel
- shared memory
- message passing interface
- parallel computing
- parallel computers
- parallel computation
- parallel algorithm
- high performance computing
- message passing
- parallel implementation
- data parallelism
- parallel execution
- parallel processing
- address space
- distributed memory
- parallel architectures
- fine grain
- fine grained
- parallel machines
- building blocks
- parallel programming
- processing elements
- genetic algorithm
- level parallelism
- hybrid learning
- computational power
- data flow
- source code
- open source
- image processing
- learning algorithm