Skip to content

rccl-2.7.9 for ROCm 3.10.0

Compare
Choose a tag to compare
@saadrahim saadrahim released this 30 Nov 17:02

New Features

  • Added experimental support for clique-based kernels (opt in with RCCL_ENABLE_CLIQUE=1)
    • Clique-based kernels may offer better performance for smaller input sizes
    • Clique-based kernels are currently only enabled for AllReduce under a certain byte limit (controlled via RCCL_CLIQUE_ALL_REDUCE_BYTE_LIMIT)
  • Performance improvements for Rome-based systems

Known Issues

  • Clique-based kernels are currently experimental and have not been fully tested on all topologies. By default clique-based kernels are disabled if the topology is not supported (override with RCCL_FORCE_ENABLE_CLIQUE)
  • Clique-based kernels may hang if there are differences between environment variables set across ranks