NVIDIA-NeMo / NVIDIA-NeMo/RL

[INSTALL-5] nvidia-resiliency-ext build fails when CUDA_HOME and CPATH are not set

Open
#2,709 0 comments 0 reactions 1 assignee Claimed by @kajalj22 View on GitHub
Automation bug Documentation p1
Dominant language
Python
Stars
2k
Forks
561
Avg merge
4d 5h
Merged PRs (30d)
145

Description

## Problem

When running the `uv` install command for Transformer Engine, the build of `nvidia-resiliency-ext` fails with an error stating `cuda.h` could not be found. The fix requires manually setting `CUDA_HOME` and `CPATH`, which is not documented.

## Steps to reproduce

Run the `uv` install command from the bare-metal installation guide on a system where `CUDA_HOME` is not set. The build fails with a missing `cuda.h` error.

## Suggested fix

Document that `CUDA_HOME` and `CPATH` must be set before running the install command. Alternatively, detect and set these variables automatically in the install flow.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.