Make stochastic physics independent of fv3atm and its submodules

[climbfuji]

Description

This PR and associated PRs make the stochastic_physics submodule independent of fv3atm and its submodules. For details, see NOAA-PSL/stochastic_physics#27.

Issue(s) addressed

This is required for updating/improving the cmake build system. I did not try to update the existing CMakeLists.txt, because I know that others are working on this. I only moved the existing CMakeLists.txt from FV3/stochastic_physics to stochastic_physics after making the code changes in NOAA-PSL/stochastic_physics#27.

Testing

Regression testing was done on hera.intel in REPRO and PROD mode for the two tests fv3_ccpp_stochy and fv3_ccpp_ca after applying the bugfix described in NOAA-PSL/stochastic_physics#24 to both the top of develop and the code in the PR. The results are bit-for-bit identical.

Dependencies

NOAA-PSL/stochastic_physics#27
NOAA-EMC/fv3atm#145
#171

[climbfuji]

Regression testing against existing baseline on hera.intel and hera.gnu: one test fails (fv3_ccpp_ca) as expected because of the bugfix for NOAA-PSL/stochastic_physics#24 for hera.intel; for hera.gnu, this test is not run and all tests pas.

rt_hera_intel.log
rt_hera_gnu.log

Note: if the bugfix for NOAA-PSL/stochastic_physics#24 is applied to the current head of develop, the results are bit for bit identical to this PR for fv3_ccpp_ca.

[climbfuji]

Creating new baselines on hera.gnu and hera.intel: all tests pass

rt_hera_intel_create.log
rt_hera_gnu_create.log

[climbfuji]

Manual testing on macosx.gnu and hera.intel: code compiles with compile.sh as well as with compile_cmake.sh. Code runs and gives same results for fv3_ccpp_stochy compared to head of develop on macsox.gnu.

[climbfuji]

Regression tests passed on orion.intel, hera.intel, hera.gnu; logs updated in the PR.

[pjpegion]

Changes look good to me.

[pjpegion]

The stochastic_physics repo has been updated. I'm cannot approve anything here, only leave comments.

[aerorahul]

The stochastic_physics repo has been updated. I'm cannot approve anything here, only leave comments.

Are you sure @pjpegion? Does github let you do a review? If so, it should let you approve/comment/request changes.

[climbfuji]

Regression tests passed on wcoss_cray and wcoss_dell_p3; logs updated in the PR.

[pjpegion]

I can only leave comments.

[junwang-noaa]

Phil, I am not sure if you were added to ufs-weather-model developer group, would you try again?

…

On Thu, Aug 6, 2020 at 12:09 PM Phil Pegion ***@***.***> wrote: I can only leave comments. — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <#171 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AI7D6TMVU6K7T5GXDM2TEOTR7LIS7ANCNFSM4PDVMJEA> .

[pjpegion]

Now I can approve.