Communication-Avoiding Parallel Sparse-Dense Matrix-Matrix Multiplication.
Penporn KoanantakoolAriful AzadAydin BuluçDmitriy MorozovSang-Yun OhLeonid OlikerKatherine A. YelickPublished in: IPDPS (2016)
Keyphrases
- matrix multiplication
- distributed memory
- shared memory
- message passing
- parallel implementation
- dense stereo
- matrix factorization
- parallel algorithm
- motion field estimation
- coefficient matrix
- parallel processing
- high dimensional
- parallel machines
- optical flow
- dense optical flow
- sparse representation
- higher order
- special case
- dense motion estimation