Litcius/Paper detail

Shfl-BW

Guyue Huang, Haoran Li, Minghai Qin, Fei Sun, Yufei Ding, Yuan Xie

2022Proceedings of the 59th ACM/IEEE Design Automation Conference16 citationsDOIOpen Access PDF

Abstract

Weight pruning in deep neural networks (DNNs) can reduce storage and computation cost, but struggles to bring practical speedup to the model inference time. Tensor-cores can significantly boost the throughput of GPUs on dense computation, but exploiting tensor-cores for sparse DNNs is very challenging. Compared to existing CUDA-cores, tensor-cores require higher data reuse and matrix-shaped instruction granularity, both difficult to yield from sparse DNN kernels. Existing pruning approaches fail to balance the demands of accuracy and efficiency: random sparsity preserves the model quality well but prohibits tensor-core acceleration, while highly-structured block-wise sparsity can exploit tensor-cores but suffers from severe accuracy loss.

Topics & Concepts

SpeedupPruningComputer scienceGranularityComputationTensor (intrinsic definition)CUDAInferenceSparse matrixParallel computingBlock (permutation group theory)ExploitReuseVoronoi diagramAlgorithmComputational scienceArtificial intelligenceMathematicsPure mathematicsOperating systemEcologyGaussianQuantum mechanicsBiologyAgronomyPhysicsGeometryComputer securityAdvanced Neural Network ApplicationsStochastic Gradient Optimization TechniquesParallel Computing and Optimization Techniques
Shfl-BW | Litcius