Validate and use MoE expert-parallel communication overlap in Megatron-Bridge, including overlap_moe_expert_parallel_comm, delay_wgrad_compute, and flex dispatcher backends such as DeepEP and HybridEP.
Install
npx skills add NVIDIA/skills --skill nemo-mbridge-perf-expert-parallel-overlapnpx skills add NVIDIA/skills --skill nemo-mbridge-perf-expert-parallel-overlap -a claude-code -gnpx skills add NVIDIA/skills --skill nemo-mbridge-perf-expert-parallel-overlap -a codex -ggemini skills install https://github.com/NVIDIA/skills.git --path skills/nemo-mbridge-perf-expert-parallel-overlapgh skill install NVIDIA/skills nemo-mbridge-perf-expert-parallel-overlapgit clone --depth 1 https://github.com/NVIDIA/skills /tmp/NVIDIA__skills
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/NVIDIA__skills/skills/nemo-mbridge-perf-expert-parallel-overlap ~/.claude/skills/nemo-mbridge-perf-expert-parallel-overlap # Claude Code
cp -R /tmp/NVIDIA__skills/skills/nemo-mbridge-perf-expert-parallel-overlap ~/.agents/skills/nemo-mbridge-perf-expert-parallel-overlap # Codex, Cursor, Gemini CLI, Copilot, OpenCode