Nemo Mbridge Perf Moe Comm Overlap Agent Skill

MoE Communication Overlap

For the higher-level overview, see:

• @docs/training/communication-overlap.md
• @skills/nemo-mbridge-perf-moe-comm-overlap/card.yaml

Quick Decision

Use MoE communication overlap when:

• EP > 1
• token dispatch or combine time is visible in the profile
• the run is already correct and you are now tuning throughput

Avoid turning it on as an early bring-up step. It is easier to validate after

the dispatcher, routing mode, and recompute plan are already stable.

Enablement

cfg.comm_overlap.overlap_moe_expert_parallel_comm = True

# Optional: delayed wgrad for additional overlap
cfg.comm_overlap.delay_wgrad_compute = True

# IMPORTANT: disable shared expert overlap when using dispatch overlap
cfg.model.moe_shared_expert_overlap = False

Prerequisites

• expert_model_parallel_size > 1
• num_moe_experts > 1
• moe_token_dispatcher_type must be "alltoall" or "flex"
• Precision: BF16 or FP16
• If PP is used, VPP (virtual_pipeline_model_parallel_size) must be set (non-None)

Flex dispatcher activation

Setting moe_flex_dispatcher_backend alone does not activate flex dispatch.

You must also set moe_token_dispatcher_type = "flex".

Recompute And CUDA Graph Interaction

• Full recompute is not a good companion for the overlap path.
• delay_wgrad_compute adds further constraints if CUDA-graph scopes include

attention or MoE-router work.

• In practice, selective recompute is the safer pairing when overlap is enabled.

Measured Short-Run Caveat

A 2026-05-18 current-main H100 x16 smoke on Qwen3 30B-A3B mock pretraining

used EP=16, alltoall, global batch size 1024, CUDA graphs disabled, and

moe_permute_fusion=false because the PyTorch 25.11 / TE / Triton stack failed

in Transformer Engine fused permutation in prior bring-up.

Results were directional rather than release-grade:

• no EP overlap: 41.25s steady-state mean over iterations 3-8
• EP overlap: 31.31s steady-state mean over iterations 3-8
• EP overlap plus delay_wgrad_compute: 31.20s steady-state mean over

iterations 3-8

Treat this as evidence that EP overlap can help an inter-node alltoall MoE

shape when communication is exposed. It is not proof that delayed wgrad is a

separate win, and it does not validate the fused permutation path. An earlier

2026-05-16 short smoke on the same shape showed the same pattern.

Code Anchors

• Overlap validation: src/megatron/bridge/training/comm_overlap.py
• Flex dispatcher backend: src/megatron/bridge/training/flex_dispatcher_backend.py
• Config: src/megatron/bridge/training/config.py
• Unit tests: tests/unit_tests/training/test_comm_overlap.py
• DeepEP tests: tests/unit_tests/training/test_deepep.py

Pitfalls

• Shared expert overlap conflict: moe_shared_expert_overlap and

overlap_moe_expert_parallel_comm can conflict. Disable shared expert

overlap when using the dispatch overlap path.

• PP without VPP: MoE overlap requires VPP when pipeline parallelism is

active. Without it, the overlap scheduling cannot interleave correctly.

• Flex != backend flag: moe_flex_dispatcher_backend="deepep" alone

does nothing if moe_token_dispatcher_type is still "alltoall".

• Conservative recipe defaults: Most public recipes leave MoE overlap

disabled. You need to explicitly enable it via overrides.

• Performance gains are workload-dependent: overlap helps most when dispatch

communication is already a visible slice of step time. It is not guaranteed

to help every small or lightly loaded EP run.

Verification

Look for overlap-related log messages during initialization. The comm overlap

validation in comm_overlap.py will raise if prerequisites are not met, so a

clean startup confirms the feature is active.

For a short performance-harness smoke, keep the command shape explicit and vary

only one overlap knob at a time:

uv run python scripts/performance/run_script.py \
  -m qwen \
  -mr qwen3_30b_a3b \
  --task pretrain \
  -g h100 \
  -c bf16 \
  -ng 16 \
  -gn 8 \
  --max_steps 8 \
  --cuda_graph_impl none \
  --moe_flex_dispatcher_backend None \
  --moe_a2a_overlap false \
  --tokenizer_type NullTokenizer \
  comm_overlap.overlap_moe_expert_parallel_comm=true \
  comm_overlap.delay_wgrad_compute=false \
  model.moe_shared_expert_overlap=false

If fused MoE permutation fails during bring-up, add

model.moe_permute_fusion=false to separate overlap timing from runtime-stack

validation, then retest with the matched production container.