cuda-graphs
OfficialSpeed Megatron Bridge with CUDA-graphs.
System Documentation
What problem does it solve?
CUDA graphs enable stable capture and replay of GPU operations to reduce host overhead during Megatron Bridge training.
Core Features & Use Cases
It supports two implementations: local full-iteration graphs via MCore and Transformer Engine scoped graphs for per-layer capture of attention, MLP, and MoE components, making it suitable for dense, MoE, and large-model training scenarios. Use cases include accelerating pretraining and fine-tuning by reducing iteration time and increasing throughput.
Quick Start
Enable CUDA graphs by selecting a graph implementation and scope, then start training as usual.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: cuda-graphs Download link: https://github.com/NVIDIA/skills/archive/main.zip#cuda-graphs Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.