Skip to content

GitLab

  • Menu
Projects Groups Snippets
    • Loading...
  • Help
    • Help
    • Support
    • Community forum
    • Submit feedback
    • Contribute to GitLab
  • Sign in
  • T Trilinos
  • Project information
    • Project information
    • Activity
    • Labels
    • Planning hierarchy
    • Members
  • Repository
    • Repository
    • Files
    • Commits
    • Branches
    • Tags
    • Contributors
    • Graph
    • Compare
  • Issues 936
    • Issues 936
    • List
    • Boards
    • Service Desk
    • Milestones
  • Merge requests 22
    • Merge requests 22
  • CI/CD
    • CI/CD
    • Pipelines
    • Jobs
    • Schedules
  • Deployments
    • Deployments
    • Environments
    • Releases
  • Monitor
    • Monitor
    • Incidents
  • Analytics
    • Analytics
    • Value stream
    • CI/CD
    • Repository
  • Wiki
    • Wiki
  • Snippets
    • Snippets
  • Activity
  • Graph
  • Create a new issue
  • Jobs
  • Commits
  • Issue Boards
Collapse sidebar
  • James Willenbring
  • Trilinos
  • Issues
  • #3844

Closed
Open
Created Nov 09, 2018 by James Willenbring@jmwilleMaintainer

ATDM_CONFIG_MPI_EXEC Causes Problems on Chama

Created by: jmgate

Hey @bartlettroscoe, I'm having trouble getting a build of EMPIRE up and running on chama. It'll build, but I'm seeing all of EMPIRE-PIC's tests fail (see EMPIRE_EM-Plasma-Trilinos-chama-test-all-of-EMPIRE-intel-opt-openmp). I think what's happening is on chama the ATDM_CONFIG_MPI_EXEC variable gets set to srun in cmake/std/atdm/common/toss3/environment.sh. When we test EMPIRE, we wind up submitting the ctest command via srun. Unfortunately each EMPIRE-PIC regression test winds up running

$Trilinos_MPI_EXEC -n $NUM_PROCS $EXE --i=$INPUT

where $TRILINOS_MPI_EXEC is $ATDM_CONFIG_MPI_EXEC, which is srun, so we're effectively invoking srun within something that's been submitted via srun. I don't know slurm inside and out, but that sounds like a problem.

The question, then, is what should we do about it? In the CMake file that's used for all of EMPIRE-PIC's regression tests, do we need to replace $Trilinos_MPI_EXEC with $(which mpiexec) or something to that effect? Or do we need to do something more clever with the $ATDM_CONFIG_MPI_EXEC variable in cmake/std/atdm/common/toss3/environment.sh such that it works for whatever you need it to do, but then it doesn't also mess up EMPIRE's testing on toss3 systems?

@bathmatt, did you set up our CMake stuff to use $Trilinos_MPI_EXEC? Do you know the rationale behind it?

Assignee
Assign to
Time tracking