(1) update MOM6 to 20230206 main branch (2) remove cpld_esmfthread_P8 (3) hafs bug fix #1606
Conversation
remove cpld_esmfthreads_p8
jkbk2004
requested review from
BrianCurtis-NOAA,
DeniseWorthen and
DusanJovic-NOAA
jkbk2004
added
Baseline Updates
|
Do we need a new baseline? Can't we just add the one new debug test to the existing baseline? |
Sure, I can revert the BL_DATE change and do manually add hafs_regional_storm_following_1nest_atm_ocn_debug. |
|
@BrianCurtis-NOAA I am going to manually create hafs_regional_storm_following_1nest_atm_ocn_debug and copy over to develop-20230201 |
|
@jkbk2004 Can you explain how the baselines are identical? Debug should run optimizations different than non-debug which would cause different results. |
I think there is no rounding-off difference with -g -traceback or debug options at lower level optimization. |
|
I agree we should see some level of rounding-off difference by triggering different level of optimization. The difference not necessarily by debug option. I think this is another reason to double check our compile options. |
|
I still think we need to be sure about the compile, it makes no sense to me that the answers aren't different. For example, the atmf000.nc control_p8 and control_debug_p8 files are different. I think that is after a single startup timestep (ie, 000 is really after the first timestep of the atm). |
|
@jkbk2004 As I explained in the slack, the results from the new HAFS debug test are different from the HAFS control hafs_regional_storm_following_1nest_atm_ocn results. The baseline filenames are different, and they were put in the control baseline by mistake, but the RT will still pass because the code changes have no impact on other tests and previous baseline is not touched in the full RT. Hope this clarifies the confusion. |
|
I know what is happening now. The 001 files from the debug run are still present on hera in the non-debug baseline hafs_regional_storm_following_1nest_atm_ocn. We should remove the files that were created by the debug test from the non-debug baseline. |
The next PR will create new baselines, right? We can just wait for that, it won't hurt to keep the extra baselines in the non-debug directory. |
|
@BrianCurtis-NOAA Normally I might agree about the next baseline, but in this case I think the amount of confusion that has been generated makes it worthwhile to clean up. |
Dogwood is cleaned up. New debug baselines are created in the correct dir. Dogwood is set for this merge. Acorn can officially be skipped. The RT's will compile but not run, no logs are output. Seems a system issue, but it will need attention ASAP for the next set of PR's |
|
OK! |
|
the ecflow nodes on Dogwood went down before i could re-run one test that failed, but it passed earlier. The debug and non-debug tests passed, so WCOSS2 is good to go. |
|
cool! I am merging NOAA-EMC/fv3atm#622. @jiandewang can you merging mom6 pr? |
|
@jiandewang NOAA-EMC/fv3atm#622 was merged. Please, go ahead to update submodule pointers and revert the changes in gitmodules after you merge mom6 pr. |
MOM6 merged |
|
@jiandewang cool! yeah, we can merge once submodule pointers (fv3/mom6) are updated and gitmodules changes are reverted. |
|
@jiandewang The hash for MOM6 needs to be d0c4550 |
|
I reverted MOM6 and committed, but for FV3, this is what I did (after revert its submodule): but I got many files being removing information, am I doing correct ? (I haven't committed yet) |
|
@jiandewang I'm going to guess you checked out your own develop branch and not the one from FV3? you need to add https://github.com/NOAA-EMC/fv3atm to your remotes, update it, then i.e. in FV3: |
|
After: git checkout upstream develop (or maybe upstream/develop), please
also add
git submodule update --init --recursive
to get all the subcomponents.
…On Tue, Feb 14, 2023 at 4:02 PM Brian Curtis ***@***.***> wrote:
@jiandewang <https://github.com/jiandewang> I'm going to guess you
checked out your own develop branch and not the one from FV3? you need to
add https://github.com/NOAA-EMC/fv3atm to your remotes, update it, then git
checkout <NOAA-EMC/fv3atm remote name> develop
i.e. in FV3:
git remote add upstream https://github.com/NOAA-EMC/fv3atm
git remote update
git checkout upstream develop (or maybe upstream/develop)
cd ..
git add FV3
git push origin update-MOM6-20230206
—
Reply to this email directly, view it on GitHub
<#1606 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TKCJBJJLNGL6RIFNUTWXPXETANCNFSM6AAAAAAUW6RO3M>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
many thanks to Denise's help, FV3 and MOM6 are pointing to the right hash # now |
Description
MOM6 main branch is updated on 20230206, this is based on EMC ungridded dimension pstokes PR (mom-ocean/MOM6#1591), but being added through cherry-pick the ci failure fixing from dev/gfdl.
need to make corresponding updating in dev/emc and ufs-weather-model. No answer change is expected.
Top of commit queue on: TBD
Input data additions/changes
Anticipated changes to regression tests:
- No changes are expected to any regression test.
- Changes are expected to the following tests:
a new test is added, hafs_regional_storm_following_1nest_atm_ocn_debugSubcomponents involved:
Combined with PR's (If Applicable):
Commit Queue Checklist:
Linked PR's and Issues:
Testing Day Checklist:
Testing Log (for CM's):
HERA-gun-log.pdf