MPI-GDS: High Performance MPI Designs with GPUDirect-aSync for CPU-GPU Control Flow Decoupling.
Akshay VenkateshKhaled HamidoucheSreeram PotluriDavide RossettiChing-Hsiang ChuDhabaleswar K. PandaPublished in: ICPP (2017)
Keyphrases
- graphics processing units
- control flow
- parallel implementation
- parallel computing
- high performance computing
- general purpose
- parallel programming
- message passing interface
- data flow
- parallel computers
- scientific computing
- parallel algorithm
- shared memory
- massively parallel
- message passing
- parallel computation
- distributed memory
- multithreading
- gpu implementation
- memory hierarchy
- graphics processors
- computing systems
- parallel processing
- data transfer
- parallel architectures
- transition systems
- software testing
- processing units
- process model
- real time
- memory bandwidth
- heterogeneous computing
- parallel machines
- modeling language
- test cases
- knowledge representation
- database systems