CCPP updates for PUMAS round3 by cacraigucar · Pull Request #75 · ESCOMP/PUMAS

cacraigucar · 2026-01-16T16:58:29Z

Round3 updates to get PUMAS into CCPP. These updates allow the CCPP version of PUMAS to compile in CAM-SIMA. These changes should be bit-for-bit in CAM (currently just tested the aux_cam izumi gnu and nag tests and they were bit-for-bit).

Major change was to move the conversion to pumas_r8 out of PUMAS and into the CAM-SIMA interstitial code. The reasoning to do this is that two copies were occurring - one to subset to the PUMAS dimension in the interstitial and then one to convert to pumas_r8 inside PUMAS. By moving it out, both subsetting and unit conversion can occur in one copy.

There is also a placeholder routine and metadata: micro_pumas_tempfix.F90. This placeholder routine defines the proc_rates_type DDT. This is basically a stub version of the PUMAS diagnostics routine which has not been completely implemented. When the PUMAS CCPP diagnostics routine is brought in, the micro_pumas_tempfix.F90 can be removed. Until then, it is needed to allow PUMAS to build in CCPP.

…dditional PUMAS settings check.

…penacc directive

nusbaume

Thanks for making the changes needed to get PUMAS working in CAM-SIMA @cacraigucar! I have some questions and change requests, but most are just repeated standard name updates (or things I messed up in the first CCPP attempt). Of course if you have any questions or concerns with any of my requests please let me know!

nusbaume · 2026-06-23T02:37:29Z

  !> \section arg_table_micro_pumas_ccpp_init Argument Table
  !! \htmlinclude micro_pumas_ccpp_init.html
-  subroutine micro_pumas_ccpp_init(gravit, rair, rh2o, cpair, tmelt, latvap, latice,     &
+  subroutine micro_pumas_ccpp_init(Gravit, rair, rh2o, cpair, tmelt, latvap, latice,     &


Why capitalize Gravit here?

I don't know - some random keystroke in vi. Reverted

nusbaume · 2026-06-23T02:42:24Z

I realize this was my doing, but can we change this from 512 to * so that it can be whatever length the host model/framework wants it to be?

Suggested change

character(len=*), intent(out) :: errmsg !PUMAS/CCPP error message (none)

integer, intent(out) :: errcode !CCPP error code (1)

nusbaume · 2026-06-23T02:44:08Z

+  !Set error code to non-zero value if PUMAS returns an error message:
+  if (trim(pumas_errstring) /= "") then
+    errcode = 1
+    errmsg = trim(pumas_errstring)
+    return
+  end if


I don't think this code block is needed given that it is basically a copy of the code block immediately below it (which doesn't need a return as it is already at the end of the subroutine).

nusbaume · 2026-06-23T02:48:38Z

Can we get rid of the un-needed save here? I again realize this was my doing. Thanks!

nusbaume · 2026-06-23T02:52:10Z

                                  errmsg, errcode)

    !External dependencies:
    use ccpp_kinds,        only: kind_phys


Do we need ccpp_kinds here anymore?

nusbaume · 2026-06-30T02:26:10Z

+  units = none
+  dimensions = ()
+  type = character
+  kind = len=512


Just noting that this should probably be kind = len=*:

Suggested change

kind = len=512

kind = len=*

changed everywhere

nusbaume · 2026-07-01T02:22:34Z

+
+   implicit none
+   private
+   save


Remove unnecessary save here?

nusbaume · 2026-07-01T02:27:21Z

+
+
+   ! CCPP error handling variables
+   character(len=512), intent(out) :: errmsg


Make this len=* here?

Suggested change

character(len=512), intent(out) :: errmsg

character(len=*), intent(out) :: errmsg

nusbaume · 2026-07-01T02:30:00Z

+  #Core PUMAS dependencies:
+  dependencies = micro_pumas_v1.F90,pumas_kinds.F90
+  dependencies = pumas_gamma_function.F90,micro_pumas_utils.F90
+  dependencies = micro_pumas_diags.F90
+  dependencies = pumas_stochastic_collect_tau.F90,tau_neural_net_quantile.F90
+  dependencies = module_neural_net.F90,ML_fixer_check.F90
+
+  #External dependencies:
+  #Lives in NCAR/ccpp-physics
+  dependencies = ../../../to_be_ccppized/wv_sat_methods.F90


I think this can be simplified to the following:

Suggested change

#Core PUMAS dependencies:

dependencies = micro_pumas_v1.F90,pumas_kinds.F90

dependencies = pumas_gamma_function.F90,micro_pumas_utils.F90

dependencies = micro_pumas_diags.F90

dependencies = pumas_stochastic_collect_tau.F90,tau_neural_net_quantile.F90

dependencies = module_neural_net.F90,ML_fixer_check.F90

#External dependencies:

#Lives in NCAR/ccpp-physics

dependencies = ../../../to_be_ccppized/wv_sat_methods.F90

#Core PUMAS dependencies:

dependencies = pumas_kinds.F90

dependencies = micro_pumas_diags.F90

Ha, ha - this is the quick stub I threw together based off the full diagnostics package. The plan is to remove it when the diagnostics are officially committed. I'll clean it up, even though it will be shortlived.

nusbaume · 2026-07-01T02:34:38Z

+
+module micro_pumas_tempfix
+
+   use micro_pumas_diags, only: proc_rates_type


Can we move this use statement down to the subroutine level?

cacraigucar

Once this is approved (and I've run my CAM and CAM-SIMA tests), I will include updating to this PUMAS tag in my misc ESCOMP/CAM PR I am putting together.

cacraigucar · 2026-07-01T17:39:12Z

+
+module micro_pumas_tempfix
+
+   use micro_pumas_diags, only: proc_rates_type


cacraigucar · 2026-07-01T17:39:17Z

+
+   implicit none
+   private
+   save


cacraigucar · 2026-07-01T17:39:25Z

+
+
+   ! CCPP error handling variables
+   character(len=512), intent(out) :: errmsg


cacraigucar · 2026-07-01T17:42:15Z

+  #Core PUMAS dependencies:
+  dependencies = micro_pumas_v1.F90,pumas_kinds.F90
+  dependencies = pumas_gamma_function.F90,micro_pumas_utils.F90
+  dependencies = micro_pumas_diags.F90
+  dependencies = pumas_stochastic_collect_tau.F90,tau_neural_net_quantile.F90
+  dependencies = module_neural_net.F90,ML_fixer_check.F90
+
+  #External dependencies:
+  #Lives in NCAR/ccpp-physics
+  dependencies = ../../../to_be_ccppized/wv_sat_methods.F90


Ha, ha - this is the quick stub I threw together based off the full diagnostics package. The plan is to remove it when the diagnostics are officially committed. I'll clean it up, even though it will be shortlived.

cacraigucar · 2026-07-01T17:44:41Z

  !> \section arg_table_micro_pumas_ccpp_init Argument Table
  !! \htmlinclude micro_pumas_ccpp_init.html
-  subroutine micro_pumas_ccpp_init(gravit, rair, rh2o, cpair, tmelt, latvap, latice,     &
+  subroutine micro_pumas_ccpp_init(Gravit, rair, rh2o, cpair, tmelt, latvap, latice,     &


I don't know - some random keystroke in vi. Reverted

cacraigucar · 2026-07-01T19:16:25Z

-[micro_reff_snow_out]
-  standard_name = microphysics_effective_radius_of_stratiform_snow_particle
+[pumas_reff_snow_out]
+  standard_name = pumas_effective_radius_of_stratiform_snow_particle


changed two places

cacraigucar · 2026-07-01T19:17:04Z

+  units = none
+  dimensions = ()
+  type = character
+  kind = len=512


changed everywhere

cacraigucar · 2026-07-01T19:21:02Z

-[micro_numsnow_tend_external_in]
-  standard_name = microphysics_tendency_of_mass_number_concentration_of_snow_wrt_moist_air_and_condensed_water_from_external_microphysics
+[pumas_numsnow_tend_external]
+  standard_name = tendency_of_pumas_mass_number_concentration_of_snow_wrt_moist_air_and_condensed_water_from_external


cacraigucar · 2026-07-01T19:23:02Z

cacraigucar · 2026-07-01T19:24:51Z

nusbaume and others added 12 commits September 29, 2025 14:42

Fix bad units in CCPP metadata file.

81d173b

Update logical variables to have clearer standard names, and add an a…

ab966b9

…dditional PUMAS settings check.

Updates to support accum_enhan

ac06a70

First commit of rearrange where pumas_r8 conversion done

1bf50f5

Next round of changes which compiled in CAM

7e3164b

temporary workaround until get diags completed

6501aaa

Fixes for CAM-SIMA prebuild processing

4ba6311

fix compiler error

101b267

add proc_rates allocate into PUMAS CCPP code

747fcc6

fixes for proc_rates DDT and add timestep_final to deallocate proc_rate

e3073ea

Update some incorrect ccpp standard names and add a missing $ to an o…

c04b38a

…penacc directive

Update standard names

c4aec4a

cacraigucar requested a review from nusbaume June 18, 2026 21:37

nusbaume requested changes Jul 1, 2026

View reviewed changes

Address reviewer comments

62ff68d

cacraigucar commented Jul 1, 2026

View reviewed changes


	character(len=*), intent(out) :: errmsg !PUMAS/CCPP error message (none)
	integer, intent(out) :: errcode !CCPP error code (1)



		! CCPP error handling variables
		character(len=512), intent(out) :: errmsg

	character(len=512), intent(out) :: errmsg
	character(len=*), intent(out) :: errmsg


		module micro_pumas_tempfix

		use micro_pumas_diags, only: proc_rates_type

Uh oh!

Conversation

cacraigucar commented Jan 16, 2026

Uh oh!

nusbaume left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cacraigucar left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants