More Efficient Reduction Algorithms for Non-Power-of-Two Number of Processors in Message-Passing Parallel Systems.
Rolf RabenseifnerJesper Larsson TräffPublished in: PVM/MPI (2004)
Keyphrases
- message passing
- shared memory
- parallel architectures
- distributed systems
- interconnection networks
- message passing interface
- distributed shared memory
- single processor
- matrix multiplication
- distributed memory
- sum product algorithm
- computational complexity
- belief propagation
- parallel algorithm
- parallel execution
- parallel programming
- inference in graphical models
- high end
- parallel computing
- parallel computers
- multithreading
- probabilistic inference
- parallel processing
- graphics processing units
- factor graphs
- parallel machines
- multiprocessor systems
- processing units
- massively parallel
- parallel implementation
- approximate inference
- multi view
- bayesian networks