Compile.sh #104
Compile.sh #104
Conversation
MinsukJi-NOAA
commented
Jun 3, 2020
MinsukJi-NOAA
commented
- Modify Moorthi's compile.sh (changing compile.sh and updating before_components.mk, configure.fv3.… #93) and before_components
- Modify rt.sh and related files to work with new compile.sh
…hera.intel, and configure.fv3.wcoss_dell_p3
…lefiles in dell and hera
* Change to fv3_coupled module in before_components * Add MOM6_MAKEOPT and CICE_MAKEOPT in compile.sh for debug * Change compile specifications in rt.conf * Change arguments to function call compile.sh in rt.sh and rt_utils.sh
|
I recommend dropping "_coupled" as with compile.sh, we don't need duplicate
files.
…On Thu, Jun 4, 2020 at 1:31 PM Jun Wang ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In conf/before_components.mk
<#104 (comment)>
:
> CHOSEN_MODULE=$(BUILD_TARGET)/fv3_coupled_debug
endif
#$(info CHOSEN_MODULE is $(CHOSEN_MODULE))
-CONFIGURE_NEMS_FILE=configure.fv3_coupled.$(BUILD_TARGET)
+ifneq (,$(findstring INTEL16=Y,$(FV3_MAKEOPT)))
I think we eliminated intel 16. So we may not need this statement, also we
don't suggest CCPP to run with intel 16.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#104 (review)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALLVRYRDTGH2KMBMDJD25CDRU7K65ANCNFSM4NRUWCMA>
.
--
Dr. Shrinivas Moorthi
Research Meteorologist
Modeling and Data Assimilation Branch
Environmental Modeling Center / National Centers for Environmental
Prediction
5830 University Research Court - (W/NP23), College Park MD 20740 USA
Tel: (301)683-3718
e-mail: [email protected]
Phone: (301) 683-3718 Fax: (301) 683-3718
|
|
Well, I need to have my own post since post doesn't take care of my needs. If this option is not in the official version, every time I merge, I have to take care of this. While I can do this this illustrates user unfriendly nature of this whole system. |
module files for fv3 and fv3_coupled are different. RT's do not pass with the fv3 module. I think the main difference between them is parallel netcdf and esmf. One thing we can do is to introduce a logic into fv3 for weather-model vs s2s-model. |
|
I hope at some point we use single module file and a single configue file per machine as ufs-weather_model for GFSv17 will be the same as ufs-s2s-model and compile.sh works for both. There is no need to make things more complicated than needed. |
|
I just created a ticket on updating netcdf/esmf library, so we will have
one module file for both weather and climate once that is done.
I'd suggest to remove cust_post as other people may want to add their
personalized libraries here, then it will be hard to maintain the official
code. This is an official repository, the general rule is to have no
personalized code/directory.
…On Thu, Jun 4, 2020 at 2:38 PM Denise Worthen ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In conf/configure.fv3.hera.intel
<#104 (comment)>
:
>
include $(ESMFMKFILE)
ESMF_INC = $(ESMF_F90COMPILEPATHS)
+ifeq ($(CUST_POST),Y)
Is cust_post going to remain or not?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#104 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TOGNOCBOW4V7M3BSXLRU7S2PANCNFSM4NRUWCMA>
.
|
|
Yes and CMEPS too. In my branch I already inserted CMEPS and was able to compile on Dell. |
CMEPS is already in. |
|
Late to the party, but I want to confirm that all "if INTEL16" logic is no longer needed. |
|
What is the difference between CICE_MAKEOPT and CICE6_MAKEOPT? |
We are using *_MAKEOPT just for a debug flag, but potentially other component-specific flags can also be appended. In that sense, can we leave them separate for now? When the transition to CICE6 is complete, we can consolidate variables then? |
|
OK, it does not have to be done now, but since CICE5 and CICE6 cannot be
used together, in compile.sh, one could just define CICE_MAKEOPT which
could be used in either CICE5 or CICE6, depending on which one is turned on.
…On Thu, Jun 4, 2020 at 3:47 PM Minsuk Ji ***@***.***> wrote:
What is the difference between CICE_MAKEOPT and CICE6_MAKEOPT?
Can't there be only one generic "CICE_MAKEOPT" that can be used with
either CICE5 or CICE6?
We are using *_MAKEOPT just for a debug flag, but potentially other
component-specific flags can also be appended. In that sense, can we leave
them separate for now? When the transition to CICE6 is complete, we can
consolidate variables then?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#104 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALLVRYSERLWOJ5E6GFGNJ5DRU723JANCNFSM4NRUWCMA>
.
--
Dr. Shrinivas Moorthi
Research Meteorologist
Modeling and Data Assimilation Branch
Environmental Modeling Center / National Centers for Environmental
Prediction
5830 University Research Court - (W/NP23), College Park MD 20740 USA
Tel: (301)683-3718
e-mail: [email protected]
Phone: (301) 683-3718 Fax: (301) 683-3718
|
|
This seems ready to merge but do you want me to test the CICE6 build first? There would be no regression tests, just comfirmation that CICE6 builds in both regular and debug mode. |
Were you able to successfully build CICE6 with current appBuilder method? |
|
Yes, I was able to build CICE6 at the time (05/11) by using a CICE6 app builder file and component_CICE6 which Tony created. |
Maybe you should give it a try then... |
|
Neither build worked. Could you take a look at the compile logs ? /scratch2/NCEPDEV/climate/Denise.Worthen/WORK/S2S_c6sh/tests This is a from a branch I created here: |
|
Can you try setting clean_before and clean_after (lines 21 and 22) to NO in compile.sh, and retry? |
|
I got another but different error for the non-bug compile. The debug compile is still going. |
|
Let me clone your repository and test it out. |
|
Let me check the branch I created. It might not be what I think it is. I'll give it another try. |
|
I got the same error when I tried to build with the app builder file: |
|
I'm going to need to spend some time tracking this down. I see Tony made a commit to the Consortium repo about 10 days ago and it looks like something must have broken. I think what we've done in compile.sh to make room for CICE6 is probably fine, it is in the build of CICE6 itself that we need to fix something. So it is probably safe to go ahead and merge if you want to get this closed. |
|
I agree. |
