skills.rest
yo-steven
yo-stevenCOMMUNITY
Explore all skills by this author →
GitHub

exec-local-compile

Compile TensorRT-LLM from source into a Python package inside a Docker container.

Updated May 23, 2026
Search Tags:#docker#build automation#python packaging#incremental build#nvtx#gpu compilation#tensorRT-LLM
One-click install
npx skills add https://github.com/yo-steven/skills-exploration-20260522 --skill exec-local-compile-yo-steven
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: exec-local-compile
Source: https://github.com/yo-steven/skills-exploration-20260522/tree/main/skills/TensorRT-LLM/exec-local-compile
Command: npx skills add https://github.com/yo-steven/skills-exploration-20260522 --skill exec-local-compile-yo-steven

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you compile TensorRT-LLM from source reliably when you have direct GPU access on a compute node, avoiding the slow and constrained workflow of building on non-GPU machines.

Core Features & Use Cases

  • GPU-aware verification: Confirms the environment is a compute node with visible GPUs by checking nvidia-smi before compiling.
  • Targeted build configuration: Builds with architecture selection, NVTX support, and incremental compilation defaults for faster iteration.
  • Developer-friendly install & validation: Installs the editable development package and verifies the installed tensorrt_llm version import.

Quick Start

Run nvidia-smi to confirm GPUs are visible, then compile in-place with ./scripts/build_wheel.py --trt_root /usr/local/tensorrt --benchmarks --use_ccache -a "<arch>" -f --nvtx and install with pip install -e .[devel].

Frequently Asked Questions about exec-local-compile

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I compile TensorRT-LLM from source on a GPU compute node?▼

To compile TensorRT-LLM from source on a GPU compute node, run the provided build_wheel.py script with your TensorRT root path, enable ccache for incremental builds, and specify your target GPU architecture. This process builds the package inside a Docker container with visible NVIDIA GPUs.

Why should I build TensorRT-LLM directly on a GPU machine instead of a non-GPU build server?▼

Building TensorRT-LLM directly on a GPU machine avoids the slow and constrained workflow of compiling on non-GPU machines. This approach leverages direct GPU access on a compute node to reliably compile the source code into a working Python package.

What prerequisites are needed to compile TensorRT-LLM locally with Docker?▼

Compiling TensorRT-LLM locally requires a working Docker container environment with visible NVIDIA GPUs verified via nvidia-smi, and a valid TensorRT installation located at /usr/local/tensorrt. You also need to specify the target GPU architecture during the build process.

Can I enable NVTX support and incremental compilation when building TensorRT-LLM?▼

Yes, you can enable NVTX support and incremental compilation when building TensorRT-LLM by passing the appropriate flags to the build_wheel.py script. The build process uses ccache by default for faster iterative local builds and supports configurable GPU architecture selection.

How do I install and validate the compiled TensorRT-LLM Python package?▼

To install and validate the compiled TensorRT-LLM Python package, use pip install -e .[devel] to set up the editable development package. After installation, verify the build by importing the tensorrt_llm module and checking the reported version.

Related Skills

Explore More Software Engineering Skills→
ovachieverovachiever
52

tensorrt-llm

Optimize LLM inference on NVIDIA GPUs with TensorRT-LLM.

Community
Advanced
BellaBeBellaBe
37

verify-maps

Verify Python code maps against generated symbol tables before code generation.

Community
Intermediate
yo-stevenyo-steven

exec-slurm-compile

Compile TensorRT-LLM on a SLURM cluster via enroot container batch jobs.

Community
Intermediate
NVIDIANVIDIA
2.8k

exec-local-compile

Build and install TensorRT-LLM from source in a Docker container with visible GPUs.

Official
Intermediate