LLNL / scr

SCR caches checkpoint data in storage on the compute nodes of a Linux cluster to provide a fast, scalable checkpoint / restart capability for MPI codes.
http://computing.llnl.gov/projects/scalable-checkpoint-restart-for-mpi
Other
99 stars 36 forks source link

WIP: SCR_ADD_TEST: run python based tests #522

Closed ofaaland closed 1 year ago

ofaaland commented 2 years ago

Run tests using the python scripts in addition to the existing tests which use the shell scripts.

Do not run bash tests under flux; bash scripts for flux do not exist.

Tested under flux and slurm so far.

Results under flux, where only the python-based tests are run since bash scripts do not exist:

[faaland1@fluke6 branch:b-cmake-test-9 build] $grep Passed /tmp/t1
 1/14 Test  #1: serial_test_api_restart_py .................   Passed    3.30 sec
 2/14 Test  #2: parallel_test_api_restart_py ...............   Passed    3.57 sec
 3/14 Test  #3: serial_test_api_shared_file_restart_py .....   Passed    2.78 sec
 4/14 Test  #4: parallel_test_api_shared_file_restart_py ...   Passed    3.42 sec
 5/14 Test  #5: serial_test_config_restart_py ..............   Passed    2.49 sec
 6/14 Test  #6: parallel_test_config_restart_py ............   Passed    3.03 sec
 7/14 Test  #7: serial_test_api_multiple_restart_py ........   Passed    1.62 sec
 8/14 Test  #8: parallel_test_api_multiple_restart_py ......   Passed    1.63 sec
 9/14 Test  #9: serial_test_ckpt_restart_py ................   Passed    2.48 sec
10/14 Test #10: parallel_test_ckpt_restart_py ..............   Passed    3.04 sec
11/14 Test #11: serial_test_ckpt_F_restart_py ..............   Passed    2.95 sec
12/14 Test #12: parallel_test_ckpt_F_restart_py ............   Passed    3.70 sec
13/14 Test #13: serial_test_ckpt_F90_restart_py ............   Passed    2.94 sec
14/14 Test #14: parallel_test_ckpt_F90_restart_py ..........   Passed    3.42 sec
ofaaland commented 2 years ago

We should only consider merging this after my other PRs are landed.

ofaaland commented 1 year ago

Abandoning this PR, replacing with #526.

We decided to start by supporting the python scripts only under Flux, and finish testing/cleaning up python scripts for other RMs only after that is stable and working well.