Several issues with how some of the hrrr 'quilt restart' tests are configured

[DusanJovic-NOAA]

There are several issues with how some of the hrrr 'quilt restart' tests are configured.

All quilt restart tests are supposed to exactly reproduce the restart outputs compared to the corresponding non-quilt restart tests, for example hrrr_control_qr must produce identical restart files as hrrr_control. In order to test this both of these tests have the same baseline directory (CNTL_DIR variable in test configuration file). Because of this only one of these two tests, and only one of these two tests, should be labeled as a test that generates new baselines, by having the string 'baseline' in the 4th column in rt.conf. This is not the case for these tests:

hrrr_control_qr (intel version)
hrrr_control_qr_dyn32_phy32 (intel version)
hrrr_control_qr (gnu version)
hrrr_control_qr_dyn32_phy32 (gnu version)

Another issue is that some 'qr' tests, I found one so far, have different baseline directory compared to the corresponding non-qr version. Tests hrrr_control_dyn32_phy32 and hrrr_control_qr_dyn32_phy32 should have the same baseline directory but they do not:

$ grep CNTL_DIR tests/hrrr_control_dyn32_phy32
export CNTL_DIR=hrrr_control_dyn32_phy32

$ grep CNTL_DIR tests/hrrr_control_qr_dyn32_phy32
export CNTL_DIR=hrrr_control_qr_dyn32_phy32

At some point we will switch to running all or most of the tests using quilting restart, but before we do that these issues must be fixed so that we are sure that quilting restart works correctly.

[zach1221]

Hi, Dusan-

This change has been added for hrrr_control and hrrr_control_qr in PR 1777. I think we just need to do the same now for hrrr_control_dyn32_phy32 and hrrr_control_qr_dyn32_phy32 . Then this issue can be closed. @jkbk2004

[zach1221]

This issue can be closed now. These changes are reflected in rt.conf. Baseslines listed for only the non qr cases.
gnu

intel