Probing the Efficacy of Hardware-Aware Weight Pruning to Optimize the SpMM Routine on Ampere GPUs.
Roberto L. CastroDiego AndradeBasilio B. FraguelaPublished in: PACT (2022)
Keyphrases
- graphics hardware
- graphics processing units
- graphics cards
- computational power
- hardware and software
- parallel architectures
- low cost
- real time
- graphics processors
- search space
- commodity hardware
- highly parallel
- hardware architecture
- computer systems
- parallel processing
- high end
- pruning algorithm
- weight assignment
- massively parallel
- pruning method
- hardware implementation
- quality assurance
- parallel programming
- general purpose
- image processing
- vlsi implementation
- pruning algorithms
- weighting scheme
- computing power
- data acquisition
- neural network
- heterogeneous computing
- blue gene
- gpu implementation
- multi core processors
- pruning strategy
- parallel computing
- computing systems
- personal computer