perf-torch-cuda-graph-specialist
Subagent "perf-torch-cuda-graph-specialist" from NVIDIA/TensorRT-LLM: Expert in CUDA Graph capture, replay, and optimization for PyTorch.
Template by NVIDIA
Install
Works with Claude Code. Read the repository before you give any tool access to your data.
curl -fsSL https://raw.githubusercontent.com/NVIDIA/TensorRT-LLM/main/.claude/agents/perf-torch-cuda-graph-specialist.md -o perf-torch-cuda-graph-specialist.md
Read it and adapt it to your project before use.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: debugging, refactoring
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
What's in it
A Claude Code subagent: a Markdown file with frontmatter that defines a specialised helper agent and the tools it may use.
- For
- Claude Code
- Length
- 61 lines
- File
.claude/agents/perf-torch-cuda-graph-specialist.md- name
- perf-torch-cuda-graph-specialist
- description
- Expert in CUDA Graph capture, replay, and optimization for PyTorch. Delegate to this agent for: (1) Analyzing code for CUDA Graph compatibility, (2) Detecting and eliminating host-device synchronizations, (3) Selecting the right CUDA Graph API (torch.compile, make_graphed_callables, TE, MCore CudaGraphManager, FullCudaGraphWrapper, manual), (4) Implementing capture and replay, (5) Full-iteration CUDA Graph capture, (6) Troubleshooting capture failures and performance issues.
Sections
- Workflow
- Safe Modification Workflow
- Key Rules
The repository's license isn't open source, so the text isn't shown here. Read it on GitHub.
Signals
- Stars
- 14,937
- Forks
- 2,862
- Open issues
- 679
- Last commit (this folder)
- Apr 8, 2026 (6 months ago)
- Last release
- Apr 20, 2026 (6 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- NOASSERTION
- Repository
- github.com/NVIDIA/TensorRT-LLM
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from NVIDIA
- 0.745score
cdf-video-data-augmentation
NVIDIA
Run the CDF Video Data Augmentation Airflow workflow on Kubernetes or OSMO to create Cosmos video-to-video variants and auto-label both source and generated…
SkillAFreshness: A★ 3.7KClaude Code +9 - 0.745score
nvidia-skill-finder
NVIDIA
Use for NVIDIA-related requests where an NVIDIA skill might help, even if the user did not ask for a skill.
SkillAFreshness: A★ 3.7KClaude Code +9 - 0.745score
cdf-image-attribute-augmentation
NVIDIA
Run the CDF Orchestration Image Attribute Augmentation DAG on Kubernetes - person-crop clothing augmentation, attribute search, and augmented dataset…
SkillAFreshness: A★ 3.7KClaude Code +9 - 0.745score
cdf-orchestration-preflight
NVIDIA
Set up missing Linux prerequisites and audit whether a Cosmos Data Factory (CDF) Orchestration checkout and its Kubernetes or OSMO environment are ready to…
SkillAFreshness: A★ 3.7KClaude Code +9 - 0.745score
ethos-write
NVIDIA
Writes an AI agent's ethos.md — a durable snapshot of its purpose, intended behavior, boundaries, and success criteria — from the answers gathered by…
SkillAFreshness: A★ 3.7KClaude Code +9 - 0.745score
cdf-autonomous-vehicle-dataset-generation
NVIDIA
Prepare payloads, run and monitor the Autonomous Vehicle Dataset Generation Airflow workflow, and inspect results for driving-clip retrieval and Cosmos3-Nano…
SkillAFreshness: A★ 3.7KClaude Code +9
Cite this page
perf-torch-cuda-graph-specialist. Agent Mart. https://agentmart.dev/templates/nvidia-tensorrt-llm-perf-torch-cuda-graph-specialist (last updated 2026-10-24)