NVIDIA / NVIDIA/cccl

[BUG][STF] Enabling P2P accesses fails if the application already enabled it

Open
#4,412 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
C++
Stars
2.5k
Forks
486
Avg merge
2d 6h
Merged PRs (30d)
295

Description

### Is this a duplicate?

- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this bug and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Type of Bug

Runtime Error

### Component

CUDA Experimental (cudax)

### Describe the bug

Maybe this should not raise an error, but a warning, or no warning as the application must have enabled it on purpose.

We should check that there is no risk of disabling P2P before the completion of STF operations, or document that if it was enabled it is undefined behaviour to disable it again.

### How to Reproduce

Create a main application that enables P2P before creating a ctx on a multi-gpu machine

### Expected behavior

This should not crash during ctx startup

### Reproduction link

_No response_

### Operating System

_No response_

### nvidia-smi output

_No response_

### NVCC version

_No response_

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.