nemo-mbridge-perf-moe-hardware-configs
Representative MoE training playbooks by hardware platform and model family. Summarizes rounded throughput bands, parallelism patterns, and common tuning stacks.
- License 许可
- Apache-2.0
- 最近更新
- 2026-07-18
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- nemo-mbridge-perf-moe-hardware-configs
- license
- Apache-2.0
- description
- Representative MoE training playbooks by hardware platform and model family. Summarizes rounded throughput bands, parallelism patterns, and common tuning stacks.
- when_to_use
- Hardware-specific MoE playbooks or throughput estimates; 'MoE on H100', 'GB200 config', 'expected throughput', 'MoE hardware playbook', 'parallelism for B200'.
03 SKILL.md 正文
MoE Hardware Configuration Reference
Stable docs: @docs/training/moe-optimization.md Card: @skills/nemo-mbridge-perf-moe-hardware-configs/card.yaml
Quick Platform Playbook
| Platform | Typical MoE strategy | What usually matters most |
|---|---|---|
| H100 | DeepEP + stronger PP + moderate TP | communication overlap and PP efficiency |
| B200 | DeepEP + MXFP8 + careful PP layout | container quality and tuned comm settings |
| GB200 | HybridEP + partial CUDA graphs + CPU cleanup | host overhead, topology-aware dispatch, memory headroom |
| GB300 | HybridEP + newer FP8 and kernel stack | same GB200 playbook, usually with a higher ceiling |
First Answer Checklist
For hardware playbook questions, answer from these canonical rows before adding throughput caveats:
| Workload | Hardware | Dispatcher | Layout |
|---|---|---|---|
| DSV3 | H100 | DeepEP | TP=2, EP=64, PP=8, VPP=4 |
| DSV3 | GB200/GB300 | HybridEP | TP=1, EP=64, PP=4, VPP=4 |
| Qwen3 235B | H100 | DeepEP | TP=2, EP=32, PP=8, VPP=4 |
| Qwen3 235B | GB200 | HybridEP | TP=1 or 2, EP=32-64, PP=4, VPP=unspecified |
For Qwen3 235B on GB200, explicitly say VPP=unspecified; do not invent or
extrapolate VPP=12 unless a measured row provides it. Include TE-scoped CUDA
graph scopes (attn, moe_router, moe_preprocess),
CUDA_DEVICE_MAX_CONNECTIONS selection,
PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True, NCCL_GRAPH_REGISTER=0,
GB200/GB300 CPU-side tuning, and the warning not to cargo-cult tracker rows.
Rounded Performance Bands
These are intentionally rounded so the document stays durable as the tracker moves. Treat them as planning ranges, not exact promises.
| Workload family | Hardware | Typical band | Representative shape |
|---|---|---|---|
| DSV3, large-scale | H100 | low-to-mid hundreds TFLOPS/GPU, high-teens MFU | TP2, EP64, PP8, DeepEP |
| DSV3, large-scale | B200 | high-hundreds TFLOPS/GPU, mid-teens MFU | TP1, EP32, PP8, DeepEP |
| DSV3, large-scale | GB200 | around 1K TFLOPS/GPU, low-20s MFU | TP1, EP64, PP4, HybridEP |
| DSV3, large-scale | GB300 | above the GB200 band, often mid-20s MFU | TP1, EP64, PP4, HybridEP |
| Qwen3 235B | H100 | low-300s TFLOPS/GPU, around 30% MFU | TP2, EP32, PP8, DeepEP |
| Qwen3 235B | GB200 | high-hundreds TFLOPS/GPU in tuned runs | TP1 or TP2, EP32-64, PP4, HybridEP |
| Qwen3 30B | H100 | low-200s TFLOPS/GPU | TP1, EP8, PP1, DeepEP |
| Qwen3-Next 80B | GB200 | low-300s TFLOPS/GPU in BF16-class runs | TP1, EP32, PP2, HybridEP |
Representative Config Families
DSV3 on H100
Dispatcher: DeepEP
TP=2 EP=64 PP=8 VPP=4
Routing: force balance
Recompute: light-to-moderate selective recompute
Priority: overlap communication and keep PP efficient
DSV3 on B200
Dispatcher: DeepEP
TP=1 EP=32 PP=8 VPP=2 or similar
Precision: MXFP8-class
Recompute: selective recompute around MLA up-projection and MLP-side modules
Priority: container quality, PP layout, and DeepEP SMS tuning
DSV3 on GB200 or GB300
Dispatcher: HybridEP
TP=1 EP=64 PP=4 VPP=4
Precision: MXFP8-class
CUDA Graph: attn + moe_router + moe_preprocess
Priority: HybridEP, CPU optimization, and graph-friendly static shapes
Qwen3 235B on H100
Dispatcher: DeepEP
TP=2 EP=32 PP=8 VPP=4
Recompute: norm and activation-side selective recompute
Priority: communication overlap and router-path cleanup
Qwen3 235B on GB200
Dispatcher: HybridEP
TP=1 or 2 EP=32 to 64 PP=4 VPP=unspecified unless measured
CUDA Graph: attn + moe_router + moe_preprocess
Recompute: moe_act, mlp, or norm depending on memory pressure
Priority: balance throughput against memory headroom
Qwen3-Next 80B on GB200
Dispatcher: HybridEP
TP=1 EP=32 PP=2 VPP around 4
CUDA Graph: attn + moe_router + moe_preprocess
Priority: pipeline layout and grouped GEMM quality
Cross-Cutting Patterns
PP layout
E= embeddingt= transformerm= MTPL= loss|= stage boundary
The biggest platform difference is usually not just the dispatcher. It is the combination of dispatcher, PP shape, and whether VPP keeps each stage balanced.
Recompute strategy
| Memory pressure | Starting point |
|---|---|
| low | none or a very narrow selective set |
| moderate | moe_act, mlp, norm, or similar selective modules |
| high | model-specific up-projection plus selective MoE and MLP modules |
| extreme or long-context | full recompute only if the selective path still does not fit |
Environment variables
CUDA_DEVICE_MAX_CONNECTIONS=1
CUDA_DEVICE_MAX_CONNECTIONS=32 # common when EP overlap and CUDA graphs are combined
PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
NCCL_GRAPH_REGISTER=0
CPU-side tuning
On GB200 and GB300, CPU affinity and general host-overhead cleanup can move the needle almost as much as a dispatcher swap. Treat them as first-class tuning work, not as afterthoughts.
Pitfalls
-
Do not cargo-cult a tracker row: the winning config usually depends on routing mode, container, and PP layout as much as on hardware name.
-
Container quality matters: large regressions can come from the software stack rather than the model recipe.
-
VPP must be intentional: a bad VPP split can erase the gain from a better dispatcher.
-
Compare absolute throughput, not only MFU: MFU can mislead when switching between BF16, FP8, and other precision modes.
-
Force-balance routing is the safer benchmark default: keep routing mode fixed when comparing hardware or dispatcher stacks.
04 同仓库更多 Skills
arxiv-search
Search arXiv physics, math, and computer science preprints using natural language queries. Powered by Valyu semantic search.
arxiv-search
Search arXiv physics, math, and computer science preprints using natural language queries. Powered by Valyu semantic search.
working-with-legacy-code
Safely change and test untested codebases using Feathers' "Working Effectively with Legacy Code". Use when the user mentions "legacy code", "no tests", "untested codebase", "how do I test this", "seams", "characterization tests", "golden master", "sprout method", "afraid to change this code", "monster method", "dependency breaking", or "inherited a messy codebase". Also trigger when changing code without tests safely, getting a class under test when constructors, statics, or singletons block it, adding features to tangled modules, or planning incremental test coverage for an old codebase. Covers the legacy-code change algorithm, seams, characterization tests, sprout/wrap, and dependency-breaking techniques. For refactoring code that already has tests, see refactoring-patterns. For day-to-day code quality, see clean-code.
working-with-legacy-code
Safely change and test untested codebases using Feathers' "Working Effectively with Legacy Code". Use when the user mentions "legacy code", "no tests", "untested codebase", "how do I test this", "seams", "characterization tests", "golden master", "sprout method", "afraid to change this code", "monster method", "dependency breaking", or "inherited a messy codebase". Also trigger when changing code without tests safely, getting a class under test when constructors, statics, or singletons block it, adding features to tangled modules, or planning incremental test coverage for an old codebase. Covers the legacy-code change algorithm, seams, characterization tests, sprout/wrap, and dependency-breaking techniques. For refactoring code that already has tests, see refactoring-patterns. For day-to-day code quality, see clean-code.
web-typography
Select, pair, and implement typefaces for web projects. Use when the user mentions "font pairing", "which typeface", "line height", "responsive typography", "web font loading", "type hierarchy", "variable fonts", "FOUT/FOIT", "typographic scale", or "the text is hard to read". Also trigger when choosing between system fonts and web fonts, optimizing font-loading performance, or designing readable long-form content. Covers readability evaluation, CSS implementation, and performance optimization. For overall UI design systems, see refactoring-ui. For dramatic typographic experiences, see top-design.
web-typography
Select, pair, and implement typefaces for web projects. Use when the user mentions "font pairing", "which typeface", "line height", "responsive typography", "web font loading", "type hierarchy", "variable fonts", "FOUT/FOIT", "typographic scale", or "the text is hard to read". Also trigger when choosing between system fonts and web fonts, optimizing font-loading performance, or designing readable long-form content. Covers readability evaluation, CSS implementation, and performance optimization. For overall UI design systems, see refactoring-ui. For dramatic typographic experiences, see top-design.