Bug fixes to drag_suite.F90 and GFS_surface_composites.F90 #140
Conversation
mdtoyNOAA
requested review from
SamuelTrahanNOAA,
tanyasmirnova,
christinaholtNOAA,
joeolson42,
hannahcbarnes,
haiqinli and
zhanglikate
as code owners
|
Is this expected to change the results of any regression tests? |
|
@SamuelTrahanNOAA Sam, yes, my bug fix will change the results. |
|
If it changes results, then I need you to generate baselines. Once you have the baselines, I will copy them to common areas, run the final regression tests, and merge. |
|
@mdtoyNOAA Mike, could you please provide the baseline? |
|
@SamuelTrahanNOAA The baseline for my change is on jet in /mnt/lfs4/BMC/wrfruc/smirnova/RRFS_code/RRFS_regtest/tests |
|
You also need to re-enable the disabled regression test in rt.conf and rt_gnu.conf: |
|
@tanyasmirnova That is not the location of your baseline. That is the location of your "tests" directory where you ran rt.sh. I need to know where the REGRESSION_TEST_INTEL directory resides. |
|
@SamuelTrahanNOAA <https://github.com/SamuelTrahanNOAA> , I updated the PR description to include the directory location of the baselines on Hera. These baselines included the bugfixes to both the drag_suite and Tanya’s fixes to GFS_surface_composites.F90.
… On Mar 21, 2022, at 10:44 AM, Samuel Trahan (NOAA contractor) ***@***.***> wrote:
@tanyasmirnova <https://github.com/tanyasmirnova> That is not the location of your baseline. That is the location of your "tests" directory where you ran rt.sh. I need to know where the REGRESSION_TEST_INTEL directory resides.
—
Reply to this email directly, view it on GitHub <#140 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARRVLICMI7A64XSOZGTXFMDVBCRP7ANCNFSM5RDPUR3Q>.
You are receiving this because you were mentioned.
|
|
@tanyasmirnova Also, you did not generate the baseline for the test we had disabled. #COMPILE | -DAPP=S2S -DDEBUG=ON -DCCPP_SUITES=FV3_GFS_v16_coupled_p8 | - wcoss_cray | fv3 | AFTER I copy your baselines, you can rerun just that one test rather than the whole set. Do not do it yet, or it'll delete all of the existing baselines you generated! |
|
@mdtoyNOAA Your baselines lack the cpld_debug_p8 directory, which is generated by the test that we had disabled due to Tanya's bug fix breaking it. Please rerun the baseline generation using only these two lines: It'll generate just that test's baseline. |
|
@SamuelTrahanNOAA Sam, the baseline is in/lfs1/BMC/wrfruc/smirnova/wrfruc/Tanya.Smirnova/RT_RUNDIRS/Tanya.Smirnova/FV3_RT/rt_282750 |
|
@tanyasmirnova <https://github.com/tanyasmirnova> Yes, I did create baselines, which included your bugfixes as well as mine. I think mine are the ones we need to work with. I’m running the “cpld_debug_p8” tests now.
… On Mar 21, 2022, at 10:59 AM, tanyasmirnova ***@***.***> wrote:
@SamuelTrahanNOAA <https://github.com/SamuelTrahanNOAA> Sam, the baseline is in/lfs1/BMC/wrfruc/smirnova/wrfruc/Tanya.Smirnova/RT_RUNDIRS/Tanya.Smirnova/FV3_RT/rt_282750
@mdtoyNOAA <https://github.com/mdtoyNOAA> Mike, did you create the baseline in you RT?
—
Reply to this email directly, view it on GitHub <#140 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARRVLIABHPF7KDD6KIYG2YTVBCTGHANCNFSM5RDPUR3Q>.
You are receiving this because you were mentioned.
|
|
@tanyasmirnova This is also not the baseline directory. It is the directory in which your workflow executed. There is a directory named "REGRESSION_TEST_INTEL" somewhere, with your baselines. However, if yours are identical to Mike's, then I don't need them (it would be redundant). |
|
@SamuelTrahanNOAA @mdtoyNOAA Yes, Mike's baseline should be used because he has combined both mine and his bug fixes. |
|
@SamuelTrahanNOAA I have added the successful test "cpld_debug_p8" to my baselines at: Hera: Jet: Do I need to copy and paste the run log for "cpld_debug_p8" into the "RegressionTests_..log" files that were generated on 2022_03_19 for the full tests (minus "cpld_debug_p8") and push these to my fork? |
|
I have copied the baselines and I am running regression tests now. |
|
@SamuelTrahanNOAA <https://github.com/SamuelTrahanNOAA> I’ll rerun the tests.
… On Mar 21, 2022, at 10:53 AM, Samuel Trahan (NOAA contractor) ***@***.***> wrote:
@mdtoyNOAA <https://github.com/mdtoyNOAA> Your baselines lack the cpld_debug_p8 directory, which is generated by the test that we had disabled due to Tanya's bug fix breaking it. Please rerun the baseline generation using only these two lines:
COMPILE | -DAPP=S2S -DDEBUG=ON -DCCPP_SUITES=FV3_GFS_v16_coupled_p8 | - wcoss_cray | fv3 |
RUN | cpld_debug_p8 | - wcoss_cray | fv3 |
It'll generate just that test's baseline.
—
Reply to this email directly, view it on GitHub <#140 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARRVLICZHLQQSLCGJJIBGBDVBCSSLANCNFSM5RDPUR3Q>.
You are receiving this because you were mentioned.
|
Fixed bug in drag_suite.F90 that prevented stochastic physics from functioning for small-scale GWD and turbulent orographic form drag. Also, deleted commented-out lines from previous commit.
Fixed bugs in GFS_surface_composites.F90 (bug fixes by @tanyasmirnova ).
Testing
How were these changes tested?
What compilers / HPCs was it tested with? Hera: Intel and GNU. Jet: Intel.
Are the changes covered by regression tests? Yes.
Have the ufs-weather-model regression test been run? On what platform?
Baseline Reg test data on Hera at:
/scratch1/BMC/wrfruc/mtoy/stmp4/Michael.Toy/FV3_RT/REGRESSION_TEST_INTEL.2022_03_19
and
/scratch1/BMC/wrfruc/mtoy/stmp4/Michael.Toy/FV3_RT/REGRESSION_TEST_GNU.2022_03_19
Baseline Reg test data on Jet at:
/lfs4/BMC/wrfruc/mtoy/RT_BASELINE/Michael.Toy/FV3_RT/REGRESSION_TEST_INTEL.2022_03_19