CAM icon indicating copy to clipboard operation
CAM copied to clipboard

Broken PUMAS GPU code and GPU regression test

Open sjsprecious opened this issue 1 year ago • 0 comments

What happened?

The PUMAS GPU code is broken for two reasons (https://github.com/ESCOMP/PUMAS/issues/67 and https://github.com/ESCOMP/PUMAS/issues/55) and thus the GPU regression test won't work either.

One pull request is issued to fix the two PUMAS problems (https://github.com/ESCOMP/PUMAS/pull/68).

Once that PR is merged and tagged, we need to bring it in to make the GPU regression test suite work on Derecho.

Note that we also need to bring in the ccs_config_cesm0.0.99 tag (https://github.com/ESMCI/ccs_config_cesm/tree/ccs_config_cesm0.0.99), which is necessary to make the PUMAS GPU code work and fix the EOSHIFT function bug from NVHPC compiler (#883).

What are the steps to reproduce the bug?

  • ./create_newcase --case /glade/derecho/scratch/$user/cam6/case/your_case_name --mach derecho --compiler nvhpc --compset F2000dev --res f19_f19_mg17 --walltime 01:00:00 --run-unsupported --queue main --ngpus-per-node 4 --gpu-type a100 --gpu-offload openacc
  • cd /glade/derecho/scratch/$user/cam6/case/your_case_name
  • ./case.setup
  • ./case.build
  • ./case.submit

What CAM tag were you using?

cam6_3_153

What machine were you running CAM on?

CISL machine (e.g. cheyenne)

What compiler were you using?

NVHPC

Path to a case directory, if applicable

No response

Will you be addressing this bug yourself?

No

Extra info

No response

sjsprecious avatar Apr 10 '24 20:04 sjsprecious