Skip to content

#1281 numba mvc support #1282

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

Merged
merged 2 commits into from
Aug 14, 2023
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions Dockerfile.tmpl
Original file line number Diff line number Diff line change
Expand Up @@ -261,6 +261,10 @@ RUN pip install scipy \
apt-get install -y pandoc && \
pip install essentia

# #1281 Install numba MVC support:
Copy link
Contributor

@rosbo rosbo Aug 16, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@djherbis Do you have a bug for adding numba MVC support with where you found the instructions? Why it was necessary? etc.

Tests are failing with:

  File "/opt/conda/lib/python3.10/site-packages/numba/cuda/codegen.py", line 159, in get_cubin
    linker.add_ptx(ptx.encode())
  File "/opt/conda/lib/python3.10/site-packages/numba/cuda/cudadrv/driver.py", line 2710, in add_ptx
    compile_result = compile_ptx(ptx.decode(), self.ptx_compile_options)
  File "/opt/conda/lib/python3.10/site-packages/ptxcompiler/api.py", line 33, in compile_ptx
    raise RuntimeError(error_log)
RuntimeError: ptxas application ptx input, line 9; fatal   : Unsupported .version 8.0; current version is '7.8'

Which seems to indicate an incompatible version of ptx.

Also, this is installing cuda (gpu) specific libraries and should be inside a {{ if eq .Accelerator "gpu" }} ... {{ end }} block.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I found this guide which is probably where you got the instructions from: https://numba.readthedocs.io/en/stable/cuda/minor_version_compatibility.html#enabling-mvc-support

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@rosbo #1281 was the issue, numba was installing before but not actually working for some users.

Fixed so its only in the GPU image:
c195472

Note that the image was broken before these changes too, we were getting numba failures from @psbang PR, so I merged mine hoping it would fix theirs (it did change the error).

RUN pip install ptxcompiler-cu11 cubinlinker-cu11 --extra-index-url=https://pypi.nvidia.com
ENV NUMBA_CUDA_ENABLE_MINOR_VERSION_COMPATIBILITY=1

RUN apt-get install -y git-lfs && \
/tmp/clean-layer.sh

Expand Down