fv3 changes for s2s one step cold start #202
Conversation
junwang-noaa
changed the title
|
@junwang-noaa should I run tests on cheyenne. intel with both gnumake and cmake? |
|
I did not run gnumake in ufs-weather model, only cmake. I did not add
link_mpi option to compile configuration files (conf/*) on platforms other
than wcoss, hera and orion, you may want to check if it is working on
cheyenne.
…On Wed, Sep 16, 2020 at 10:00 AM Dom Heinzeller ***@***.***> wrote:
@junwang-noaa <https://github.com/junwang-noaa> should I run tests on
cheyenne. intel with both gnumake and cmake?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#202 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TMQA6LPESIZCRSDNSLSGDAHXANCNFSM4RNCTENQ>
.
|
In fact, you added it to |
|
Great, I did not test it.
…On Wed, Sep 16, 2020 at 10:28 AM Dom Heinzeller ***@***.***> wrote:
I did not run gnumake in ufs-weather model, only cmake. I did not add
link_mpi option to compile configuration files (conf/*) on platforms other
than wcoss, hera and orion, you may want to check if it is working on
cheyenne.
… <#m_-6133230094313618156_>
In fact, you added it to conf/configure.fv3.cheyenne.intel, that's why I
was asking. I'll test it!
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#202 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TK56RBGM3KKBGZ2H43SGDDQVANCNFSM4RNCTENQ>
.
|
I think we need to tiny changes to The rest is looking good, although the regression tests are still running. |
Update: all debug compile jobs fail on cheyenne, because even though cheyenne.intel is using the Intel compiler, it is not using Intel MPI, but SGI MPT. We need to add the same logic to the cmake and gnumake config that you added for wcoss. And probably need the same for gaea, too. I can make those changes and send you a PR to update your branch. This will not require rerunning the regression tests, since these are only machine-dependent configuration changes for machines other than tier-1 platforms. |
|
Sure.
…On Wed, Sep 16, 2020 at 11:22 AM Dom Heinzeller ***@***.***> wrote:
Great, I did not test it.
… <#m_4868682027481103853_>
I think we need to tiny changes to conf/configure.fv3.cheyenne.intel and
conf/configure.fv3.cheyenne.gnu, not directly related to this PR. There
are missing dependencies on POST_LIB with gnumake that my usual
regression testing using ecflow/cmake didn't catch.
The rest is looking good, although the regression tests are still running.
Update: all debug compile jobs fail on cheyenne, because even though
cheyenne.intel is using the Intel compiler, it is not using Intel MPI, but
SGI MPT. We need to add the same logic to the cmake and gnumake config that
you added for wcoss. And probably need the same for gaea, too. I can make
those changes and send you a PR to update your branch. This will not
require rerunning the regression tests, since these are only
machine-dependent configuration changes for machines other than tier-1
platforms.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#202 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TM4LBDTKBRWHQ2M7L3SGDJ3FANCNFSM4RNCTENQ>
.
|
…rary dependencies for cheyenne.{intel,gnu} gnumake config
…200916 Cheyenne and Gaea updates for PR "fv3 changes for s2s one step cold start"
climbfuji
left a comment
climbfuji
left a comment
There was a problem hiding this comment.
This looks good, just need to update the fv3atm submodule pointer and revert the .gitmodules changes.
…, updates to other GSL physics (ufs-community#202) This PR updates the GSL physics and does a few other things: *add --debug to ccpp_prebuild.py call when model is compiled in DEBUG mode, and pretty print the output in the build log (fixes ufs-community#187 and ufs-community#135) *add unified gravity wave drag (called unified ugwp) that combines the UGWP v1 with the GSL drag suite *add suite definition file for GFS v16beta for coupled model *move static array zs (depth of soil layer), dimensioned as 1:lsoil_lsm from Sfcprop to Model (since this is a model configuration and only needs to be stored once, not for each block) Co-authored-by: DomHeinzeller <58610420+DomHeinzeller@users.noreply.github.com> Co-authored-by: samuel.trahan <Samuel.Trahan@noaa.gov> Co-authored-by: hannah barnes <hannah.barnes@noaa.gov> Co-authored-by: tanyasmirnova <tanya.smirnova@noaa.gov> Co-authored-by: Joseph Olson <Joseph.B.Olson@noaa.gov> Co-authored-by: Michael Toy <michael.toy@noaa.gov>
3 participants
Description
This PR contains fv3 code changes to set up on step cold start.
Issue(s) addressed
Testing
RT will be conducted on wcoss, hera and orion.
Dependencies