NCCL only multi-gpu multi-node training without MPI #1114

Sign in to view logs

Triggered via pull request June 18, 2024 05:48

chinthysl

synchronize #426

ahrefs:multinode-slurm

Status Success

Total duration 7m 57s

Artifacts –

ci.yml

on: pull_request

build-cuda-windows

build-ubuntu20-04

build-cuda-fp32

build-cuda-bf16

build-cuda-fp16

build-cuda-kernels

Matrix: build-and-test-cpu