Login / Signup

GE-SpMM: general-purpose sparse matrix-matrix multiplication on GPUs for graph neural networks.

Guyue HuangGuohao DaiYu WangHuazhong Yang
Published in: SC (2020)
Keyphrases
  • general purpose
  • sparse matrix
  • matrix multiplication
  • neural network
  • graphics processing units
  • message passing
  • pattern recognition
  • random projections
  • matrix factorization
  • floating point
  • distributed memory