A W-cycle algorithm for efficient batched SVD on GPUs.
Junmin XiaoQing XueHui MaXiaoyang ZhangGuangming TanPublished in: PPoPP (2022)
Keyphrases
- times faster
- preprocessing
- dynamic programming
- optimization algorithm
- experimental evaluation
- np hard
- optimal solution
- cost function
- k means
- search space
- computational cost
- single pass
- linear programming
- highly efficient
- memory requirements
- high accuracy
- singular value decomposition
- detection algorithm
- expectation maximization
- input data
- worst case
- probabilistic model
- significant improvement
- learning algorithm
- graphics hardware
- pruning strategy
- gpu implementation
- parallel implementation
- computing systems
- improved algorithm
- convergence rate
- matching algorithm
- segmentation algorithm
- theoretical analysis
- objective function