Tetris: Accelerating Sparse Convolution by Exploiting Memory Reuse on GPU.
Xiaoyan LiuXuegui ZhengHailong YangZhongzhi LuanDepei QianPublished in: PPoPP (2024)
Keyphrases
- high dimensional
- real time
- graphics processors
- data transfer
- memory space
- parallel computing
- software reuse
- image processing
- graphics hardware
- parallel implementation
- limited memory
- multi threaded
- memory usage
- sparse data
- computing power
- compressed sensing
- parallel hardware
- computational power
- sparse representation
- multiscale
- random access
- parallel computation
- graphics processing units
- sparse coding
- general purpose
- e learning
- linear filtering
- neural network