Update intel compile to Intel2022 #571
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,24 +1,23 @@ | ||
| help([[ | ||
| ]]) | ||
|
|
||
| prepend_path("MODULEPATH", "/scratch1/NCEPDEV/nems/role.epic/miniconda3/modulefiles") | ||
|
There was a problem hiding this comment. Does this There was a problem hiding this comment. Others should also confirm but the There was a problem hiding this comment. The updated prepend_path and module versions all come from ufs-weather-model develop for Hera, Jet, and Orion. There was a problem hiding this comment. Let's work with g-w and others to update as many modulefiles as possible. Tagging @KateFriedman-NOAA , natalie-perlin, and @GeorgeVandenberghe-NOAA , and steven.bongiovanni@noaa.gov. Should anyone be removed from this list? Should anyone be added? There was a problem hiding this comment. Looks like all of the (fully) supported R&D platforms are covered in this PR (Hera, Orion, Jet). @DavidHuber-NOAA is providing updates for S4 (low support). WCOSS2 is not considered in this PR since it's only intel 2019 still and not EPIC-maintained hpc-stacks. Will update WCOSS2 to intel 2022 at a later date (likely when we move to spack-stack but not 100% on that). Updates in this PR should not break WCOSS2 functionality. There was a problem hiding this comment. OK. GSI There was a problem hiding this comment. Paths look correct. A couple versions are ahead of where workflow is (and not in UFS), but we can update to match. There was a problem hiding this comment. Is this resolved, then? There was a problem hiding this comment. If GSI paths are ahead of g-w, will There was a problem hiding this comment. This will need to be resolved in the g-w. I linked this discussion into the relevant issue NOAA-EMC/global-workflow#812. I believe this conversation can be resolved. |
||
| miniconda3_ver=os.getenv("miniconda3_ver") or "4.12.0" | ||
| load(pathJoin("miniconda3", miniconda3_ver)) | ||
|
|
||
| prepend_path("MODULEPATH", "/scratch1/NCEPDEV/nems/role.epic/hpc-stack/libs/intel-2022.1.2/modulefiles/stack") | ||
RussTreadon-NOAA marked this conversation as resolved.
Show resolved
Hide resolved
|
||
|
|
||
| local hpc_ver=os.getenv("hpc_ver") or "1.2.0" | ||
| local hpc_intel_ver=os.getenv("hpc_intel_ver") or "2022.1.2" | ||
| local hpc_impi_ver=os.getenv("hpc_impi_ver") or "2022.1.2" | ||
| local cmake_ver=os.getenv("cmake_ver") or "3.20.1" | ||
| local prod_util_ver=os.getenv("prod_util_ver") or "1.2.2" | ||
|
|
||
| load(pathJoin("hpc", hpc_ver)) | ||
| load(pathJoin("hpc-intel", hpc_intel_ver)) | ||
| load(pathJoin("hpc-impi", hpc_impi_ver)) | ||
| load(pathJoin("cmake", cmake_ver)) | ||
|
|
||
| load("gsi_common") | ||
|
|
||
| load(pathJoin("prod_util", prod_util_ver)) | ||
|
|
||
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -753,6 +753,8 @@ end subroutine contents_binary_diag_ | |
|
|
||
| subroutine contents_netcdf_diag_ | ||
| use sparsearr, only: sparr2, readarray, fullarray | ||
| use constants, only: r_missing | ||
| use screen_to_ncdiag | ||
|
There was a problem hiding this comment. GSI coding standards require
To conform to this standard we need
There was a problem hiding this comment.
|
||
| integer(i_kind),dimension(miter) :: obsdiag_iuse | ||
| integer(i_kind) :: obstype, obssubtype | ||
| type(sparr2) :: dhx_dx | ||
|
|
@@ -766,27 +768,33 @@ subroutine contents_netcdf_diag_ | |
| obssubtype = gps_allptr%rdiag(2) | ||
| call nc_diag_metadata("Observation_Type", obstype ) | ||
| call nc_diag_metadata("Observation_Subtype", obssubtype ) | ||
| call screen_to_single_nc_diag_metadata("Latitude",(gps_allptr%rdiag(3)) ) | ||
| call screen_to_single_nc_diag_metadata("Longitude",(gps_allptr%rdiag(4)) ) | ||
| call screen_to_single_nc_diag_metadata("Incremental_Bending_Angle",(gps_allptr%rdiag(5)) ) | ||
| call screen_to_single_nc_diag_metadata("Pressure",(gps_allptr%rdiag(6)) ) | ||
| call screen_to_single_nc_diag_metadata("Height",(gps_allptr%rdiag(7)) ) | ||
| call screen_to_single_nc_diag_metadata("Time",(gps_allptr%rdiag(8)) ) | ||
| call screen_to_single_nc_diag_metadata("Model_Elevation",(gps_allptr%rdiag(9)) ) | ||
| call screen_to_single_nc_diag_metadata("Setup_QC_Mark",(gps_allptr%rdiag(10)) ) | ||
| call screen_to_single_nc_diag_metadata("Prep_Use_Flag",(gps_allptr%rdiag(11)) ) | ||
| call screen_to_single_nc_diag_metadata("Analysis_Use_Flag",(gps_allptr%rdiag(12)) ) | ||
|
|
||
| call screen_to_single_nc_diag_metadata("Nonlinear_QC_Rel_Wgt",(gps_allptr%rdiag(13)) ) | ||
| call screen_to_single_nc_diag_metadata("Errinv_Input",(gps_allptr%rdiag(14)) ) | ||
| call screen_to_single_nc_diag_metadata("Errinv_Adjust",(gps_allptr%rdiag(15)) ) | ||
| call screen_to_single_nc_diag_metadata("Errinv_Final",(gps_allptr%rdiag(16)) ) | ||
| call screen_to_single_nc_diag_metadata("Observation",(gps_allptr%rdiag(17)) ) | ||
| !TODO Verify that 'adjusted' is supposed to be the same as unadjusted | ||
RussTreadon-NOAA marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| if(isnan(gps_allptr%rdiag(17)) .or. isnan(gps_allptr%rdiag(5))) then | ||
RussTreadon-NOAA marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| call nc_diag_metadata("Obs_Minus_Forecast_adjusted", sngl(real(r_missing))) | ||
| call nc_diag_metadata("Obs_Minus_Forecast_unadjusted", sngl(real(r_missing))) | ||
| else | ||
| call nc_diag_metadata("Obs_Minus_Forecast_adjusted", sngl(gps_allptr%rdiag(17))*sngl(gps_allptr%rdiag(5)) ) | ||
| call nc_diag_metadata("Obs_Minus_Forecast_unadjusted", sngl(gps_allptr%rdiag(17))*sngl(gps_allptr%rdiag(5)) ) | ||
| endif | ||
| call screen_to_single_nc_diag_metadata("GPS_Type",(gps_allptr%rdiag(20)) ) | ||
| call screen_to_single_nc_diag_metadata("Temperature_at_Obs_Location",(gps_allptr%rdiag(18)) ) | ||
| call screen_to_single_nc_diag_metadata("Specific_Humidity_at_Obs_Location",(gps_allptr%rdiag(21)) ) | ||
|
|
||
| if (save_jacobian) then | ||
| call readarray(dhx_dx, gps_allptr%rdiag(ioff+1:nreal)) | ||
|
|
||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why is is necessary to add
w3ncoto the list of common modules to build with intel2022?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@RussTreadon-NOAA
w3ncowas loaded bynemsioversions 2.5.2 and earlier, but is no longer, so it now needs to be loaded explicitly.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the explanation.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@DavidHuber-NOAA
If you are loading
nemsio, 2.5.4, it depends onw3emc. You no longer needw3nco.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
c.f. https://github.com/NOAA-EMC/NCEPLIBS-nemsio/blob/v2.5.4/cmake/PackageConfig.cmake.in
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@aerorahul I removed the w3nco load. Thanks! Note there is a placeholder for the new ncdiag library here which provides a local install on Hera presently. I will update this when the new ncdiag PR is merged and the new tag is installed everywhere.