POSTER: Cache-Oblivious MPI All-to-All Communications on Many-Core Architectures.
Shigang LiYunquan ZhangTorsten HoeflerPublished in: PPOPP (2017)
Keyphrases
- memory hierarchy
- general purpose
- message passing
- query processing
- parallel implementation
- data access
- parallelization strategy
- parallel algorithm
- prefetching
- communication systems
- shared memory
- interconnection networks
- multithreading
- communication networks
- graphical models
- parallel computers
- high performance computing
- parallel computing
- parallel programming
- computational power
- hit rate
- data management
- message passing interface
- fine grained
- neural network