diff --git a/orttraining/tools/amdgpu/Dockerfile.rocm4.0.pytorch b/orttraining/tools/amdgpu/Dockerfile.rocm4.0.pytorch index c4fb50aa56..c05888bfc4 100644 --- a/orttraining/tools/amdgpu/Dockerfile.rocm4.0.pytorch +++ b/orttraining/tools/amdgpu/Dockerfile.rocm4.0.pytorch @@ -190,6 +190,7 @@ ENV HSA_NO_SCRATCH_RECLAIM=1 # Distributed training related environment variables ENV HSA_FORCE_FINE_GRAIN_PCIE=1 ENV NCCL_DEBUG=INFO +ENV RCCL_ALLTOALL_KERNEL_DISABLE=1 # ENV NCCL_DEBUG_SUBSYS=INIT,COLL WORKDIR ${WORKSPACE_DIR}/script