Remove Noah-WRFv4, bug fix in Thompson MP inner loop, update HWRF regression tests, set ECF_TRIES to 2, update Cheyenne job submission scripts; contains "Reduce memory required by MERRA2 option; fix diag_tables for P7 tests; adds C384 P7 tests to Cheyenne.intel" (#866), contains "Update CMEPS" (#775) #831
Conversation
|
I had a compile failure for the I added the library in the ufs_wcoss_cray In the compile_001/err I see as the first occurence of But then: |
|
Denise, thanks for catching the wrong esmf lib on cray. Let's check with
nceplibs team if esmf820bs14 is installed on cray
…On Fri, Oct 29, 2021 at 7:59 AM Denise Worthen ***@***.***> wrote:
I had a compile failure for the hafs_regional_atm and
hafs_regional_atm_ocn test on cray because this PR requires esmf8.2bs14.
The update to 8.2bs14 was done in PR #776
<#776> but it
looks like the module was not updated on cray. This PR adds a new NUOPC
call in CMEPS which now fails using the older 811 esmf.
I added the library in the ufs_wcoss_cray
diff --git a/modulefiles/ufs_wcoss_cray b/modulefiles/ufs_wcoss_cray
index 01ca6e76..9586ca24 100644
--- a/modulefiles/ufs_wcoss_cray
+++ b/modulefiles/ufs_wcoss_cray
@@ -63,7 +63,7 @@ module load gni-headers
module load udreg
module load ugni
-module load esmf/811
+module load esmf/820bs14
module load fms/2021.03
module swap pmi pmi/5.0.11
In the compile_001/err I see as the first occurence of ESMF:
+ export ESMFMKFILE=/gpfs/hps/usrx/local/nceplibs/NCEPLIBS/cmake/install/NCEPLIBS-v1.3.0/esmf/esmf-820bs14/lib/esmf.mk
But then:
CMake Error at /gpfs/hps3/emc/nems/noscrub/emc.nemspara/soft/cmake/cmake-3.20.1-linux-x86_64/share/cmake-3.20/Modules/FindPackageHandleStandardArgs.cmake:230 (message):
Could NOT find ESMF (missing: ESMF_LIBRARY_LOCATION) (found version
"8.2.0")
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#831 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TMQUSR5CZXYW4L6KDDUJKEDJANCNFSM5EWRCOAQ>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
Yes, thanks Denise, also for running the wcoss tests in general. After cray is done, we are ready to merge. That was a quick one then, despite creating new baselines. |
|
Just started the cray verify step. |
|
CMEPS has been merged; hash 5beead0 |
….com/climbfuji/ufs-weather-model into remove-noah-wrfv4-bugfix-mp-thompson
CMEPS submodule pointer updated, waiting for fv3atm to be merged |
|
fv3atm has is correct (4214e6d), waiting to resolve a .gitmodules change question. Please all review carefully once more, many unrelated changes. |
|
Double checked again. All my changes are present as expected. Thanks for pulling all these in together. |
|
Should be good now! |
PR Checklist
Ths PR is up-to-date with the top of all sub-component repositories except for those sub-components which are the subject of this PR. Please consult the ufs-weather-model wiki if you are unsure how to do this.
This PR has been tested using a branch which is up-to-date with the top of all sub-component repositories except for those sub-components which are the subject of this PR
An Issue describing the work contained in this PR has been created either in the subcomponent(s) or in the ufs-weather-model. The Issue should be created in the repository that is most relevant to the changes in contained in the PR. The Issue and the dependent sub-component PR
are specified below.
If new or updated input data is required by this PR, it is clearly stated in the text of the PR.
Description
This PR removes unused/broken HAFS-HWRF regression tests that used Ferrier-Aligo MP, and updates the old HAFS-HWRF regression tests to use Noah MP instead of Noah-WRFv4 as instructed by @yangfanglin. It is understood that these old HAFS-HWRF regression tests will eventually be replaced with new HAFS regression tests.
The submodule pointers for fv3atm and ccpp-physics are updated to remove Noah-WRFv4 from fv3atm/ccpp-physics and to fix a bug in Thompson MP when the inner loop logic is used.
The changes in this PR and associated PRs require new baselines for the following four tests:
All other regression tests are b4b reproducible against the current baselines (tested on Hera with Intel and GNU). Note in particular that the Thompson MP results do not change, because the inner loop logic is not used in the regression tests.
The PR also contains an update from @DusanJovic-NOAA that uses the automatic rerun capability for ecflow:
ECF_TRIESis set to 2 in this PR, which will attempt to rerun a failed test once before reporting a failure. This is to avoid unnecessary reruns of the entire regression test suite if single tests timed out.Further, this PR as of 2021/10/28 contains the following changes:
Reduce memory required by MERRA2 option; fix diag_tables for P7 tests; adds C384 P7 tests to Cheyenne.intel #866 (Reduce memory required by MERRA2 option; fix diag_tables for P7 tests; adds C384 P7 tests to Cheyenne.intel) from @DeniseWorthen. These updates require new baselines for several of the P7 tests, the details are described in Reduce memory required by MERRA2 option; fix diag_tables for P7 tests; adds C384 P7 tests to Cheyenne.intel #866: Bit-for-bit comparisons fail only because of the changes to the forecast files associated with the updated
diag_table. All common fields in the forecast files are B4B and all restart files are B4B. This has been verified before including Update CMEPS #775 (see next bullet).Update CMEPS #775 (Update CMEPS) from @DeniseWorthen. These updates require new baselines for all coupled tests.
Update Cheyenne job submission scripts due to a recent change on Cheyenne w.r.t. how the user environment gets propagated to the jobs: replace
#!/bin/shwith#/bin/bashto fix random compile errors (`module command not found), as per recommendation by the CISL helpdesk. This change does not affect the results on Cheyenne.No new input data required.
Issue(s) addressed
Testing
Tested on Hera/Intel on 2021/10/25 against official baseline, only the four tests mentioned above failed with b4b differences.
How were these changes tested? What compilers / HPCs was it tested with? Are the changes covered by regression tests? (If not, why? Do new tests need to be added?) Have regression tests and unit tests (utests) been run? On which platforms and with which compilers? (Note that unit tests can only be run on tier-1 platforms)
Dependencies