High-Performance Tensor Contractions for GPUs.
Ahmad AbdelfattahMarc BaboulinVeselin DobrevJack J. DongarraChristopher W. EarlJoel FalcouAzzam HaidarIan KarlinTzanio V. KolevIan MasliahStanimire TomovPublished in: ICCS (2016)
Keyphrases
- graphics processing units
- high order
- general purpose
- higher order
- compute intensive
- scientific computing
- high reliability
- highly parallel
- parallel processing
- parallel programming
- computational power
- cost effective
- gpu implementation
- tensor decomposition
- real time
- operating system
- projection matrices
- dt mri
- diffusion tensor images
- pairwise