Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[BUG]: Editable builds of cuda_cccl and cuda_parallel broke after merge of pull/3597 #3761

Closed
1 task done
oleksandr-pavlyk opened this issue Feb 10, 2025 · 0 comments · Fixed by #3762
Closed
1 task done
Assignees
Labels
bug Something isn't working right.

Comments

@oleksandr-pavlyk
Copy link
Contributor

oleksandr-pavlyk commented Feb 10, 2025

Is this a duplicate?

Type of Bug

Silent Failure

Component

cuda.parallel (Python)

Describe the bug

gh-3597 switched build systems from setuptools to scikit-build-core.

They differ in their methods of producing editable installs. Setuptools places install artifacts into the source tree, while scikit-build-core uses redirect method and does not pollute the source tree, but places artifacts into site-packages.

The logic in cuda_cccl and in cuda_parallel to find header files by cuda_cccl or the shared library by cuda_parallel assumed that artifacts were installed in the source tree.

With the merge if gh-3597 that changed and editable installs stopped working correctly,

How to Reproduce

Install cuda-parallel using editable install CXX=g++ CUDACXX=/usr/local/cuda/bin/nvcc CUDAARCHS=86 pip install -e . and run tests pytest -sv --maxfail=1 tests

Expected behavior

Editable install should work as before.

Reproduction link

No response

Operating System

No response

nvidia-smi output

-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.86.10              Driver Version: 570.86.10      CUDA Version: 12.8     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA RTX A400                Off |   00000000:01:00.0  On |                  N/A |
| 30%   35C    P5             N/A /   50W |    2075MiB /   4094MiB |     28%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                         

NVCC version

(build-py312) opavlyk@ee09c48-lcedt:~/repos/cccl/python/cuda_parallel$ /usr/local/cuda/bin/nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2025 NVIDIA Corporation
Built on Wed_Jan_15_19:20:09_PST_2025
Cuda compilation tools, release 12.8, V12.8.61
Build cuda_12.8.r12.8/compiler.35404655_0
(build-py312) opavlyk@ee09c48-lcedt:~/repos/cccl/python/cuda_parallel$ python --version
Python 3.12.8
@oleksandr-pavlyk oleksandr-pavlyk added the bug Something isn't working right. label Feb 10, 2025
@github-project-automation github-project-automation bot moved this to Todo in CCCL Feb 10, 2025
@oleksandr-pavlyk oleksandr-pavlyk self-assigned this Feb 10, 2025
@cccl-authenticator-app cccl-authenticator-app bot moved this from Todo to In Review in CCCL Feb 10, 2025
@github-project-automation github-project-automation bot moved this from In Review to Done in CCCL Feb 11, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working right.
Projects
Status: Done
Development

Successfully merging a pull request may close this issue.

1 participant