Skip to content

Introduce cuda_p2p based fused_all_gather_matmul and fused_matmul_reduce_scatter #208233

Introduce cuda_p2p based fused_all_gather_matmul and fused_matmul_reduce_scatter

Introduce cuda_p2p based fused_all_gather_matmul and fused_matmul_reduce_scatter #208233

linux-jammy-cuda11.8-cudnn8-py3.8-clang12  /  build

succeeded May 24, 2024 in 45m 58s