Python issue when performing new/clean install Ubuntu 22.04

Install, Configure and Update
1

Python issue when installing, this is output from the install.sh.
This is from packages downloaded today, so most current version.

Any assistance is working through out build issue would be appreciated.
thanks - Brian

  creating build/lib.linux-x86_64-cpython-38/pycuda/compyte
  copying pycuda/compyte/array.py -> build/lib.linux-x86_64-cpython-38/pycuda/compyte
  copying pycuda/compyte/dtypes.py -> build/lib.linux-x86_64-cpython-38/pycuda/compyte
  copying pycuda/compyte/__init__.py -> build/lib.linux-x86_64-cpython-38/pycuda/compyte
  running egg_info
  writing pycuda.egg-info/PKG-INFO
  writing dependency_links to pycuda.egg-info/dependency_links.txt
  writing requirements to pycuda.egg-info/requires.txt
  writing top-level names to pycuda.egg-info/top_level.txt
  reading manifest file 'pycuda.egg-info/SOURCES.txt'
  reading manifest template 'MANIFEST.in'
  warning: no files found matching 'doc/source/_static/*.css'
  warning: no files found matching 'doc/source/_templates/*.html'
  warning: no files found matching '*.cpp' under directory 'bpl-subset/bpl_subset/boost'
  warning: no files found matching '*.html' under directory 'bpl-subset/bpl_subset/boost'
  warning: no files found matching '*.inl' under directory 'bpl-subset/bpl_subset/boost'
  warning: no files found matching '*.txt' under directory 'bpl-subset/bpl_subset/boost'
  warning: no files found matching '*.h' under directory 'bpl-subset/bpl_subset/libs'
  warning: no files found matching '*.ipp' under directory 'bpl-subset/bpl_subset/libs'
  warning: no files found matching '*.pl' under directory 'bpl-subset/bpl_subset/libs'
  adding license file 'LICENSE'
  writing manifest file 'pycuda.egg-info/SOURCES.txt'
  /home/cryosparc_user/cryosparc_worker/deps/anaconda/envs/cryosparc_worker_env/lib/python3.8/site-packages/setuptools/command/build_py.py:202: SetuptoolsDeprecationWarning:     Installing 'pycuda.cuda' as data is deprecated, please list it in `packages`.
      !!


      ############################
      # Package would be ignored #
      ############################
      Python recognizes 'pycuda.cuda' as an importable package,
      but it is not listed in the `packages` configuration of setuptools.

      'pycuda.cuda' has been automatically added to the distribution only
      because it may contain data files, but this behavior is likely to change
      in future versions of setuptools (and therefore is considered deprecated).




   Please make sure that 'pycuda.cuda' is included as a package by using
      the `packages` configuration field or the proper discovery methods
      (for example by using `find_namespace_packages(...)`/`find_namespace:`
      instead of `find_packages(...)`/`find:`).

      You can read more about "package discovery" and "data files" on setuptools
      documentation page.
2

Hi @BrianCuttler. Was this the entire output?
To which version of the CUDA toolkit did --cudapath point?

3

Two links down it points to /usr/local/cuda-12.0
(/usr/local/cuda → /usr/alternateives/cuda-12.0 → /usr/local/cuda-12.0)

thanks,
Brian

4

CUDA toolkit version 11.8 is the latest version currently supported by CryoSPARC, but we became aware of performance issues when CryoSPARC was configured with CUDA toolkit version 11.8 (fix pending). I would therefore recommend CUDA toolkit version 11.7, which could be installed with
cryosparcw install-3dflex (see details and prerequisites).

5

The fix was released as part of CryoSPARC v4.2.

6

@BrianCuttler Were you able to complete CryoSPARC installation using CUDA-11.8?

7

You know the old saying, one step back, one step back.
System came from the vendor with the OS drive and the cryo-temp drive switched.
2tb vs 8tb, so with the system going back for reinstallation I held off on the cuda reinstall.

I can let you know how it all works out, but there will be a delay before I reattempt, but
very much appreciate the advice and guideance.

thanks - Brian

8

We had a new system delivered, so my problem is slightly different than before.
Unclear on how to ask a good question.

I’ve a new(ly delivered) Ubuntu 20.04 system with GPU (unclear why vendor didn’t install 22.04). Trying to install cuda 11.8 but believe I have a mismatch with Nvidia drivers.

I’m not clear on how to match the nvidia driver with a version of cuda that will function with Cryosparc.

I know there is a methodology, so that I can then follow the correct process for the 22.04 system that came in the door and the other existing GPU-enabled Ubuntu servers on site and the GPU-enabled Suse server that needs to be upgraded to Ubuntu.
There is table that matches GPU model with nvidia driver that will work with cuda 11.8?
Thanks in advance,

Brian

9

It doesn’t matter all that much so long as the version is supported.

My normal strategy (from a clean Ubuntu 20.04 or 22.04 install):

  • Download network install .deb for Ubuntu(version) from nVidia
  • Install with dpkg
  • sudo apt install cuda-11-8
  • Reboot
  • Check driver loaded with nvidia-smi

I’ve done this on half a dozen systems for 20.04 and 22.04 and it’s not failed so far.

10

that you, we can close this out.
I have a version of nvidia that supports Cuda 11.x and all seems good now.

Important thing that I hand’t known before, version of CUDA displayed in nvidia-smi
output is max CUDA version supported, not current cuda version, explains the different
in output from nvidia-smi and nvcc --version.

thanks for your support,
Brian

11