Hyper-Q aware intranode MPI collectives on the GPU.
Iman FarajiAhmad AfsahiPublished in: ESPM@SC (2015)
Keyphrases
- parallel implementation
- parallel computing
- parallel programming
- message passing interface
- general purpose
- shared memory
- graphics processing units
- graphics hardware
- parallel algorithm
- real time
- collective intelligence
- parallel computation
- massively parallel
- high performance computing
- computing systems
- message passing
- gpu implementation
- gpu accelerated
- graphic processing unit
- learning algorithm
- parallel processing
- parallelization strategy
- markov random field
- parallel computers
- multithreading
- data structure
- information systems
- data sets