A graph-theory-based method for parallelizing the multiple-flow-direction algorithm on CUDA compatible graphics processing units.
Lijun ZhanChengzhi QinPublished in: ICSDM (2011)
Keyphrases
- parallel implementation
- gpu implementation
- graph theory
- dynamic programming
- graph theoretical
- computational complexity
- matching algorithm
- objective function
- cost function
- graphics processing units
- energy function
- combinatorial optimization
- compute unified device architecture
- general purpose
- times faster
- minimum spanning tree
- cpu implementation
- image matching
- search space
- search algorithm