Add noahmp and UFSATM hashes // Add Spack CI #2654 // PR Project Board Label Updates #2902

[dpsarmie]

Commit Queue Requirements:

• This PR addresses a relevant WM issue (if not, create an issue).
• All subcomponent pull requests (if any) have been reviewed by their code managers.
• Run the full Intel+GNU RT suite (compared to current baselines), preferably on Ursa (Derecho or Hercules are acceptable alternatives). Exceptions: documentation-only PRs, CI-only PRs, etc.
  • Commit log file w/full results from RT suite run (if applicable).
  • Commit 'test_changes.list' (unless file is empty).
• Fill out all sections of this template.

---

Description:

Add noahmp and UFSATM hashes

This PR will bring in a UFSATM hash that has a bug fix to FV3. The Atmos%iau_offset was not being set if IAU was being used. This caused the clearing of the accumulated precipitation to no occur as intended.

This PR also brings in a hash from noahmp that removes goto statements and removes an unnecessary write statement that was being triggered during land IAU runs.

Both submodules were tested independently with the RT suite. No baseline changes were found when running the noahmp submodule changes (expected). Three baseline changes were found when running the UFSATM changes (expected)

---

From #2654: Add Spack CI

This PR adds a CI workflow, Spack.yml, that builds the UFS Weather Model using a newly added in-repo copy of its Spack recipe (ci/package.py) with the intent that it be modified to include dependency changes and new CMake build options. Spack.yml uses the same custom action, NOAA-EMC/ci-test-spack-package, used in the NCEPLIBS Spack CI workflows. Currently Spack.yml uses OpenMPI but other MPI libraries could be straightforwardly used instead or in addition. Software dependencies are installed through Spack, except for those listed under spack-externals:, which is to a) reduce workflow run time and b) reduce the size of the dependency cache (see below).

A few things to note:

The Spack.yml workflow automatically creates a GitHub Actions cache with the compiled software dependencies. The whole bundle is about 500MB (~120MB without MAPL; GitHub total limit for a repo's caches is 10GB). With the build cache in place, it takes about 8-9 minutes for the workflow for building ATM to run, and about 24 minutes for S2SWA. Without the build cache, the total run time is ~1 hr or more.
Currently the ci-test-spack-package action used in this workflow has not been tested with Intel or any other non-GCC compilers. I expect that some modifications to ci-test-spack-package would be needed in order to support this, but it can certainly be done if/when needed.
In order to keep the Spack recipe in step with the code, I have added the ci-check-spack-recipe action to the workflow, which checks that all options listed in the root CMakeLists.txt are reflected in the recipe, ci/package.py. This means that when a new CMake option is added, ci/package.py should be modified to reflect it. This should usually be a low effort change, but I am happy to volunteer as needed to assist developers in updating the recipe.
I see no reason not to squash-and-merge this PR.

---

From #2902: PR Project Board Label Updates

This PR adds functionality to adjust the PR project board status automatically based on labels that are added or removed. Specifically, it handles labels for baseline updates and input data. I also included a small update to the PR template, which DeniseWorthen requested.

I tested the auto-labeling functionality using PRs from my fork (feature/add-labels branch) to the NOAA-EPIC/ufs-weather-model (feature/pr-management branch). For example, see: NOAA-EPIC#12 and the EPIC WM PR Project Board (view).

For security reasons, these GitHub actions won't run in a fork, so this functionality will only work once added to the WM repo. Since I have switched the code to point to the ufs-community project board, it will not work until merged. However, NOAA-EPIC#13 (where I am working on logic for the priority labels) and the EPIC WM PR Project Board (view) show how it would work for anyone interested in taking a look.

The details around security issues and testing are basically the same as those in PR #2908. Check out that PR description for more details.

A later PR will automate priority labels, but that logic is still in the works.

Commit Message:

* UFSWM - Fix Atmos%iau_offset bug, remove goto statements from noahmp, add spack ci, and add changes to PR proj board
  * UFSATM - Set Atmos%iau_offset to iau_offset in module_fcst_grid_comp
  * NOAHMP - Remove goto statements and remove leftover debug write statement

Priority:

• High: Needed for GFSv17

Git Tracking

UFSWM:

• Closes Atmos%iau_offset needs to be set to iau_offset in UFSATM #2912
• Closes Add Spack CI #2654
• Closes PR Project Board Label Updates #2902

Sub component Pull Requests:

• UFSATM: Fix initialization of Atmos%iau_offset NOAA-EMC/ufsatm#1015
• NOAHMP: Removal of goto statements NOAA-EMC/noahmp#16

UFSWM Blocking Dependencies:

• None

Documentation:

• Documentation update NOT required.
  • Explanation: No fundamental changes to the model have been made.

---

Changes

Regression Test Changes (Please commit test_changes.list):

• PR Updates/Changes Baselines.

Input data Changes:

• None.

Library Changes/Upgrades:

• No Updates

---

Testing Log:

• RDHPCS
  • Hera
  • Orion
  • Hercules
  • GaeaC6
  • Derecho
  • Ursa
• WCOSS2
  • Dogwood/Cactus
  • Acorn
• CI
• opnReqTest (complete task if unnecessary)

[jkbk2004]

@dpsarmie can you sync up branches?

[dpsarmie]

Ok, this should be ready to go for testing. @jkbk2004

[FernandoAndrade-NOAA]

We should be all set with testing, leaving a note in sub PRs.

[BrianCurtis-NOAA]

UFSATM: NOAA-EMC/ufsatm@67961c1

[dpsarmie]

Ok, should be good for final reviews.

[NickSzapiro-NOAA]

Noting that there are major variances in runtimes on Ursa not related to the code changes here, where some tests are over 3x faster and others 3x slower