FVCOM integration

[uturuncoglu]

I am creating new issue for FVCOM integration. So, we could keep communicating about integration issues that we might have.

[uturuncoglu]

@pvelissariou1 I am using your fork (main branch) for FVCOM at this point. https://github.com/pvelissariou1/FVCOM.git I wonder if this also builds the dependencies that you were mentioning.

[uturuncoglu]

@pvelissariou1 I think we will use develop brach for FVCOM tests.

https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop

At this point I am seeing two tests: (1) scituateharbor.fvc and (2) scituateharbor.atm2fvc2ww3. Is it possible to create configuration for ATMMESH+FVCOM (two component case). I think we could take with WW3 later.

[pvelissariou1]

@Ufuk Turuncoglu @.***> I'll talk to the FVCOM developers. They are not using Github very much, they like email communications.

@Jianhua Qi @.> , @Siqi Li @.> , @Changsheng Chen @.***> Jianhua, is it possible to have an ATMESH/FVCOM case, similar to the scituateharbor.fvc test case but passing the atmospheric data to FVCOM via ATMESH?

Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Mon, Jun 5, 2023 at 12:42 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I think we will use develop brach for FVCOM tests.

https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop

At this point I am seeing two tests: (1) scituateharbor.fvc https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop/scituateharbor.fvc and (2) scituateharbor.atm2fvc2ww3 https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop/scituateharbor.atm2fvc2ww3. Is it possible to create configuration for ATMMESH+FVCOM (two component case). I think we could take with WW3 later.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1577210542, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP2I7ULXM5MIBFEXDNTXJYK7VANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 Okay. Sure. We could communicate with them by mail if we need some help.

[pvelissariou1]

Hi Takis,

We don't have the test suite example of CoastalApp with component = "atmesh fvcom" because we applied 3-D baroclinic model of FVCOM to CoastalApp which needs other surface forcing other than wind only.

If necessary, we can create the barotropic case of FVCOM to CoastalApp to realize the data transfer from ATMESH to FVCOM. It is not difficult but takes time.

Please let us know your thoughts.

Thanks, Jianhua

---

From: Panagiotis Velissariou - NOAA Affiliate @.> Sent: Monday, June 5, 2023 1:57 PM To: oceanmodeling/ufs-coastal @.>; Ufuk Turuncoglu @.>; Jianhua Qi @.>; Siqi Li @.>; Changsheng Chen @.> Cc: oceanmodeling/ufs-coastal @.>; Mention @.> Subject: Re: [oceanmodeling/ufs-coastal] FVCOM integration (Issue #6)

[EXTERNAL SENDER] @Ufuk @.***> I'll talk to the FVCOM developers. They are not using Github very much, they like email communications.

@Jianhua @.> , @Siqi @.> , @Changsheng @.***> Jianhua, is it possible to have an ATMESH/FVCOM case, similar to the scituateharbor.fvc test case but passing the atmospheric data to FVCOM via ATMESH?

Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.**@.>

On Mon, Jun 5, 2023 at 12:42 PM Ufuk Turunçoğlu @.**@.>> wrote:

@pvelissariou1https://github.com/pvelissariou1 I think we will use develop brach for FVCOM tests.

https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop

At this point I am seeing two tests: (1) scituateharbor.fvchttps://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop/scituateharbor.fvc and (2) scituateharbor.atm2fvc2ww3https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop/scituateharbor.atm2fvc2ww3. Is it possible to create configuration for ATMMESH+FVCOM (two component case). I think we could take with WW3 later.

— Reply to this email directly, view it on GitHubhttps://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1577210542, or unsubscribehttps://github.com/notifications/unsubscribe-auth/APC7TP2I7ULXM5MIBFEXDNTXJYK7VANCNFSM6AAAAAAY3IRNIM. You are receiving this because you were mentioned.Message ID: @.***>

[pvelissariou1]

Hi Jianhua,

I completely forgot about that, thanks Jianhua. When you have the time please create a barotropic test case that can use ATMESH for atmospheric forcing. Ufuk is trying to set ufs-coastal to use ATMESH+MODEL type of configuration to test his CMEPS coupling interfaces.

Thank you Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Mon, Jun 5, 2023 at 1:22 PM Jianhua Qi @.***> wrote:

Hi Takis,

We don't have the test suite example of CoastalApp with component = "atmesh fvcom" because we applied 3-D baroclinic model of FVCOM to CoastalApp which needs other surface forcing other than wind only.

If necessary, we can create the barotropic case of FVCOM to CoastalApp to realize the data transfer from ATMESH to FVCOM. It is not difficult but takes time.

Please let us know your thoughts.

Thanks, Jianhua

---

From: Panagiotis Velissariou - NOAA Affiliate < @.> Sent: Monday, June 5, 2023 1:57 PM To: oceanmodeling/ufs-coastal < @.>; Ufuk Turuncoglu @.>; Jianhua Qi @.>; Siqi Li < @.>; Changsheng Chen @.> Cc: oceanmodeling/ufs-coastal @.>; Mention @.> Subject: Re: [oceanmodeling/ufs-coastal] FVCOM integration (Issue #6)

[EXTERNAL SENDER] @Ufuk Turuncoglu @.***> I'll talk to the FVCOM developers. They are not using Github very much, they like email communications.

@Jianhua Qi @.> , @Siqi Li @.> , @Changsheng Chen @.***> Jianhua, is it possible to have an ATMESH/FVCOM case, similar to the scituateharbor.fvc test case but passing the atmospheric data to FVCOM via ATMESH?

Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Mon, Jun 5, 2023 at 12:42 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I think we will use develop brach for FVCOM tests.

https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop

At this point I am seeing two tests: (1) scituateharbor.fvc https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop/scituateharbor.fvc and (2) scituateharbor.atm2fvc2ww3 https://github.com/noaa-ocs-modeling/CoastalApp-testsuite/tree/develop/scituateharbor.atm2fvc2ww3. Is it possible to create configuration for ATMMESH+FVCOM (two component case). I think we could take with WW3 later.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1577210542, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP2I7ULXM5MIBFEXDNTXJYK7VANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 It seems that you are not building metis for FVCOM,

https://github.com/pvelissariou1/FVCOM/blob/476e8208c3f2086260dcfb1397dcddf65f6f413a/src/libs/makefile#L20

but you are setting FLAG_411=-DMETIS_5. Is this because you already installed it to a common place. That will work for CoastalApp but not for ufs-coastal. So, I think I need to install Metis through the CMake interface but I am getting error like following,

/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/metis/struct.h(30): error: "KeyValueType" has already been declared in the current scope

I wonder if I need to pass any special flag to Metis build?

[pvelissariou1]

@Ufuk Turuncoglu @.**> Ufuk, Metis can be installed in some common place as you mentioned, built within CoastalApp (as part of ParMETIS) or can be built internally from the FVCOM source code (in src/libs). In addition to METIS, FVCOM uses other required libraries like PROJ4, Fortran-Proj, Julian that are also built from within FVCOM (or can be installed in a common place via fpr example spack). METIS can be installed using spack (as there are no licensing issues) and be part of UFS. Now to compile METIS (metis 5.1.0 or 5.2.1) I don't use any special flags, and I don't use the one contained in the src/libs folder. The one that comes with FVCOM is an older version of metis (less than 5) so in this case FLAG_411=-DMETIS_5 should be turned off. Now, the src/libs/metis does not compile (at least for me) with current versions of GCC without modifying the sources. I suggest using metis >= 5.1.0.* The compilation with Intel compilers seems to be successful, but the compiler throws a lot of warnings that in GCC are considered to be errors. The METIS libraries (5.1.0) coming with CoastalApp work fine. If you want a more recent version of METIS you may download it from:

5.2.1: https://github.com/KarypisLab/METIS/archive/refs/tags/v5.2.1.zip 5.1.0: I can send you the sources.

Please, let me know

Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Tue, Jun 6, 2023 at 10:29 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 It seems that you are not building metis for FVCOM,

https://github.com/pvelissariou1/FVCOM/blob/476e8208c3f2086260dcfb1397dcddf65f6f413a/src/libs/makefile#L20

but you are setting FLAG_411=-DMETIS_5. Is this because you already installed it to a common place. That will work for CoastalApp but not for ufs-coastal. So, I think I need to install Metis through the CMake interface but I am getting error like following,

/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/metis/struct.h(30): error: "KeyValueType" has already been declared in the current scope

I wonder if I need to pass any special flag to Metis build?

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1578987802, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP6KOIDVOGCL7NO3YPTXJ5EG5ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

I was using files under src/libs but if you want me to use another version. I could try with it. I think it would be nice to build Metis along with the FVCOM interface that I am creating. This will simplify the porting of the code.

[pvelissariou1]

Ufuk,

We can easily incorporate the new version of metis into FVCOM/src/libs (new versions use cmake for build). We need, at this point, to be careful to not break their internal "make" structure. In CoastalApp/scripts/functions_build I have the bash function compileFVCOMLibs that compiles FVCOM required libraries (found in src/libs) as needed (currently Julian, PROJ4, Fortran-Proj). I can add METIS (with an updated metis version in src/libs). Also in CoastalApp/scripts/functions_build I have the bash function compileMetis that compiles ParMETIS/METIS libraries.

I have already talked to the FVCOM developers that the libraries in the src/libs folder need to be updated and they are willing to update the codes. Let me update metis in the FVCOM/src/libs, update a few files in CoastalApp/FVCOM and I'll let you know (tomorrow probably). We are working on my FVCOM fork.

What do you think?

Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Tue, Jun 6, 2023 at 12:34 PM Ufuk Turunçoğlu @.***> wrote:

I was using files under src/libs but if you want me to use another version. I could try with it. I think it would be nice to build Metis along with the FVCOM interface that I am creating. This will simplify the porting of the code.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579186883, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP5PP4C7VOJKG5BQFCDXJ5SZRANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 Okay. I think updating the libs will solve the issue. I am not touching anything inside of the FVCOM at this point. I have made good progress in CMake interface (under UFS level) and it is almost ready if we solve the issue with Metis. Then, we could compile FVCOM as a part of ufs-coastal. Let me know when libraries are updated.

[pvelissariou1]

@Ufuk Turuncoglu @.***> How do you build the rest of the required libraries (Julian, PROJ, ...)? Do you use the src/libs/makefile to build them?

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Tue, Jun 6, 2023 at 12:59 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 Okay. I think updating the libs will solve the issue. I am not touching anything inside of the FVCOM at this point. I have made good progress in CMake interface (under UFS level) and it is almost ready if we solve the issue with Metis. Then, we could compile FVCOM as a part of ufs-coastal. Let me know when libraries are updated.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579221014, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP7J3TO4EKKPDOLYL23XJ5VWVANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 I am just extracting those tar files and CMake interface builds them for me. I am not using FVCOM build system and there is no change in there. Once I have a working version, I'll push all those changes to the ufs-coastal. So, you could check the implementation.

[pvelissariou1]

@Ufuk Turuncoglu @.***> Great, I do the same in CoastalApp. I'll try to have the libraries ready later today.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Tue, Jun 6, 2023 at 1:07 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I am just extracting those tar files and CMake interface builds them for me. I am not using FVCOM build system and there is no change in there. Once I have a working version, I'll push all those changes to the ufs-coastal. So, you could check the implementation.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579231287, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP5QQR33OZE3H72FHMLXJ5WUVANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 Thanks for your help.

[uturuncoglu]

@pvelissariou1 I have also some issue with fproj it complains in cfortran.h since I am not using make system and it has lots of definition with OS and architecture. Do you have any suggestion? If not maybe I could trigger the make build from CMake interface for fproj.

[pvelissariou1]

@Ufuk Turuncoglu @.***> Yes, the latest version of fproj does not have these issues (does not use cfortran.h) but at present does not work with FVCOM. The source code needs to be modified slightly to accept the newer version of fproj. I am talking with the developers to upgrade the FVCOM code. I'll keep you posted.

Please see how I compile FPROJ in CoastalApp/scripts/functions_build lines 1708-1748. You need to set CPPFLAGS=-DIFORT Worst case is to update cfortran.h to a most recent version. I prefer to update the FPROJ library though.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Tue, Jun 6, 2023 at 3:09 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I have also some issue with fproj it complains in cfortran.h since I am not using make system and it has lots of definition with OS and architecture. Do you have any suggestion? If not maybe I could trigger the make build from CMake interface for fproj.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579377959, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP6DKC3CYGO4F7H4W53XJ6E7PANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 Okay. I could able to compile fproj too. Thanks for your help.

[uturuncoglu]

@pvelissariou1 I am getting following error,

[ 17%] Building Fortran object FVCOM-interface/CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o
cd /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface && /glade/u/apps/ch/opt/ncarcompilers/0.5.0/intel/2022.1/mpi/mpif90 -DDOUBLE_PRECISION -DMETIS_5 -DMULTIPROCESSOR -DSINGLE_OUTPUT -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/julian -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/mod/modO/Linux.intel.64.mpt.default -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/src/include -I/glade/work/epicufsrt/contrib/hpc-stack/intel2022.1/intel-2022.1/mpt-2.25/netcdf/4.7.4/include -g -traceback -fno-alias -auto -safe-cray-ptr -ftz -assume byterecl -sox -O2 -debug minimal -fp-model source -module mod -free -c /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F -o CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o
/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3112): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [OCPCOMM4]
   USE OCPCOMM4
-------^
/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3308): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [OCPCOMM4]
   USE OCPCOMM4
-------^
/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3146): error #6404: This name does not have a type, and must have an explicit type.   [INPUTF]
   OPEN(INPUTF,FILE=TRIM(FNAME)) ; REWIND(INPUTF) 
--------^
compilation aborted for /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F (code 1)

Is OCPCOMM4 part of Metis library (only library that I did not install yet)? or it is just a configuration issue. Maybe I did not need to compile mod_utils.F. Any idea?

[pvelissariou1]

@Ufuk Turuncoglu @.***> Ufuk, I updated the metis library in FVCOM. I moved the libs folder in FVCOM/libs. I have a small README.txt file with basic instructions on metis compilation. Either clone my FVCOM fork or CoastalApp develop branch.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 7, 2023 at 12:36 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I am getting following error,

[ 17%] Building Fortran object FVCOM-interface/CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o cd /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface && /glade/u/apps/ch/opt/ncarcompilers/0.5.0/intel/2022.1/mpi/mpif90 -DDOUBLE_PRECISION -DMETIS_5 -DMULTIPROCESSOR -DSINGLE_OUTPUT -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/julian -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/mod/modO/Linux.intel.64.mpt.default -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/src/include -I/glade/work/epicufsrt/contrib/hpc-stack/intel2022.1/intel-2022.1/mpt-2.25/netcdf/4.7.4/include -g -traceback -fno-alias -auto -safe-cray-ptr -ftz -assume byterecl -sox -O2 -debug minimal -fp-model source -module mod -free -c /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F -o CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3112): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [OCPCOMM4] USE OCPCOMM4 -------^ /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3308): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [OCPCOMM4] USE OCPCOMM4 -------^ /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3146): error #6404: This name does not have a type, and must have an explicit type. [INPUTF] OPEN(INPUTF,FILE=TRIM(FNAME)) ; REWIND(INPUTF) --------^ compilation aborted for /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F (code 1)

Is OCPCOMM4 part of Metis library (only library that I did not install yet)? or it is just a configuration issue. Maybe I did not need to compile mod_utils.F. Any idea?

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579934396, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP7FPJLGD3GCH6XLXG3XKAHN3ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[pvelissariou1]

@Ufuk Turuncoglu @.***> OCPCOMM4 is a SWAN model module. Did you have any errors before reaching this point (is ocpcomm4.mod built?)

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 7, 2023 at 12:36 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I am getting following error,

[ 17%] Building Fortran object FVCOM-interface/CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o cd /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface && /glade/u/apps/ch/opt/ncarcompilers/0.5.0/intel/2022.1/mpi/mpif90 -DDOUBLE_PRECISION -DMETIS_5 -DMULTIPROCESSOR -DSINGLE_OUTPUT -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/julian -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/mod/modO/Linux.intel.64.mpt.default -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/src/include -I/glade/work/epicufsrt/contrib/hpc-stack/intel2022.1/intel-2022.1/mpt-2.25/netcdf/4.7.4/include -g -traceback -fno-alias -auto -safe-cray-ptr -ftz -assume byterecl -sox -O2 -debug minimal -fp-model source -module mod -free -c /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F -o CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3112): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [OCPCOMM4] USE OCPCOMM4 -------^ /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3308): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [OCPCOMM4] USE OCPCOMM4 -------^ /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3146): error #6404: This name does not have a type, and must have an explicit type. [INPUTF] OPEN(INPUTF,FILE=TRIM(FNAME)) ; REWIND(INPUTF) --------^ compilation aborted for /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F (code 1)

Is OCPCOMM4 part of Metis library (only library that I did not install yet)? or it is just a configuration issue. Maybe I did not need to compile mod_utils.F. Any idea?

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579934396, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP7FPJLGD3GCH6XLXG3XKAHN3ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[pvelissariou1]

@Ufuk Turuncoglu @.***> Also do you have the include paths to contain -I_FVCOMLOC/src? The ocpcomm4.mod module should be located there.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 7, 2023 at 9:49 AM Panagiotis Velissariou - NOAA Affiliate < @.***> wrote:

@Ufuk Turuncoglu @.***> OCPCOMM4 is a SWAN model module. Did you have any errors before reaching this point (is ocpcomm4.mod built?)

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 7, 2023 at 12:36 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I am getting following error,

[ 17%] Building Fortran object FVCOM-interface/CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o cd /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface && /glade/u/apps/ch/opt/ncarcompilers/0.5.0/intel/2022.1/mpi/mpif90 -DDOUBLE_PRECISION -DMETIS_5 -DMULTIPROCESSOR -DSINGLE_OUTPUT -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/julian -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/mod/modO/Linux.intel.64.mpt.default -I/glade/scratch/turuncu/COASTAL_APP/esmf-8.3.0b09/src/include -I/glade/work/epicufsrt/contrib/hpc-stack/intel2022.1/intel-2022.1/mpt-2.25/netcdf/4.7.4/include -g -traceback -fno-alias -auto -safe-cray-ptr -ftz -assume byterecl -sox -O2 -debug minimal -fp-model source -module mod -free -c /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F -o CMakeFiles/fvcom.dir/FVCOM/src/mod_utils.F.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3112): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [OCPCOMM4] USE OCPCOMM4 -------^ /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3308): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [OCPCOMM4] USE OCPCOMM4 -------^ /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F(3146): error #6404: This name does not have a type, and must have an explicit type. [INPUTF] OPEN(INPUTF,FILE=TRIM(FNAME)) ; REWIND(INPUTF) --------^ compilation aborted for /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/FVCOM-interface/FVCOM/src/mod_utils.F (code 1)

Is OCPCOMM4 part of Metis library (only library that I did not install yet)? or it is just a configuration issue. Maybe I did not need to compile mod_utils.F. Any idea?

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1579934396, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP7FPJLGD3GCH6XLXG3XKAHN3ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 It seems ocpcomm4 is defined in FVCOM-interface/FVCOM/src/swmod1.F but this seems to be SWAN code. Do we need to compile SWAN to compile FVCOM by default? Any idea?

[uturuncoglu]

@pvelissariou1 Thanks. I'll pull the changes from develop and try again.

[pvelissariou1]

@Ufuk Turuncoglu @.***> SWAN is coupled internally with FVCOM (hardcoded). I don't think there is a way to turn SWAN off at this point.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 7, 2023 at 10:51 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 It seems ocpcomm4 is defined in FVCOM-interface/FVCOM/src/swmod1.F but this seems to be SWAN code. Do we need to compile SWAN to compile FVCOM by default? Any idea?

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1581102754, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TPY6Z4PJKNDFB2JXHM3XKCPOTANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 Okay. It is good to know. I am compiling now.

[uturuncoglu]

@pvelissariou1 I could able to build Metis but I have minor question. I think there is no such application that builds all ADCIRC, SCHSIM and FVCOM together but you might want to have an executable that includes all the components. This will give you flexibility to run different applications such as CDEPS+ADCIRC or CDEPS+FVCOM just by changing nems.configure without additional build step. The only drawback is the since Metis is used also with other components, all the components needs to be complied with same version of Metis (maybe with same options). Otherwise, different versions of the same library will create namespace clash. The could be two solution for this: (1) we could use consistent Metis versions in all components (the source code of those components include also Metis directory or tar file like FVCOM) or (2) we could create a common lib directory in the UFS level and point all common libraries from there. Anyway, this is just an issue we might want to tackle in the future.

[pvelissariou1]

@Ufuk Turuncoglu @.***> Usually I build NEMS.x with all components included and switch models based on the application. Now, in CoastalApp the user has the option to build ParMETIS that includes the METIS libraries so that SCHISM, FVCOM and WW3 share exactly the same METIS/ParMETIS libraries. ADCIRC on the other hand uses METIS 4.0.2 that is built internally by ADCIRC.

To document how things work in CoastalApp and in the various modeling components Calls to metis functions for all models are as follows: ADCIRC uses: metis_partgraphkway,metis_wpartgraphkway, metis_estimatememory FVCOM uses: metis_partmeshdual, metis_meshtodual, metis_partgraphkway, metis_partgraphrecursive, inducerowpartfromcolumnpart SCHISM uses: parmetis_v3_partgeomkway, metis_setdefaultoptions, metis_setdefaultoptions, metis_partgraphkway WW3 uses: parmetis_v3_partgeomkway, metis_setdefaultoptions, metis_setdefaultoptions, metis_partgraphkway

I haven't noticed any conflicts so far. Now I tried to address the issue of all models to use the same versions of the libraries built with the same options by allowing the user of CoastalApp to build external libraries within CoastalApp (e.g., METIS/ParMETIS). All these libraries (static) are installed in the THIRDPARTY_INSTALL folder and can be accessed by all modeling components. So, this type of functionality is there. When I developed the ParMETIS cmake modules for SCHISM, and the top level CMakeLists.txt file I designed the use of the patitioning part to allow the use of (1) externally built ParMETIS/METIS, (2) internally built ParMETIS/METIS and (3) SCHISM's internal partitioning scheme.

So basically WW3, SCHISM, FVCOM use exactly the same ParMETIS/METIS libraries (4.0.3 and 5.0.1 respectively). As I mentioned before, ParMETIS cannot be included under NOAA/UFS configurations due to licensing issues (METIS does not have this problem). Furthermore, WW3 development is moving away from ParMETIS by incorporating the Scotch libraries for partitioning (https://www.labri.fr/perso/pelegrin/scotch/). This library will be officially supported under UFS/spack (this is my impression).

I don't believe that the ADCIRC developers will address the metis library issue any time soon. as the model works as expected. Also, the interface in more recent versions of metis has changed a bit.

Ufuk, in ufs-coastal we need to support all the particulars of each modeling component until we are in the position to standardize the component compilation to use common built libraries. Any thoughts on how to improve by you or anyone else will be very beneficial. We might need to open a discussion on this subject with all the developers.

Last but not least, is the licensing and public vs. private repos (e.g., ADCIRC) that needs to be addressed with the developers as well.

These are my thoughts, and sorry for the long reply.

Takis

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 7, 2023 at 5:16 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I could able to build Metis but I have minor question. I think there is no such application that builds all ADCIRC, SCHSIM and FVCOM together but you might want to have an executable that includes all the components. This will give you flexibility to run different applications such as CDEPS+ADCIRC or CDEPS+FVCOM just by changing nems.configure without additional build step. The only drawback is the since Metis is used also with other components, all the components needs to be complied with same version of Metis (maybe with same options). Otherwise, different versions of the same library will create namespace clash. The could be two solution for this: (1) we could use consistent Metis versions in all components (the source code of those components include also Metis directory or tar file like FVCOM) or (2) we could create a common lib directory in the UFS level and point all common libraries from there. Anyway, this is just an issue we might want to tackle in the future.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1581586098, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TPZJ5JE26M6SAO457YLXKD4TZANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 i have couple of minor fixes in the FVCOM related with the build such as misplaced CPP directives etc. I could create work from your FVCOM fork or I could directly push to your branch (but I need access for it). Let me know what you think?

[uturuncoglu]

@pvelissariou1 I have a question about configuring the model. In the CoastalApp are you using https://github.com/pvelissariou1/FVCOM/blob/main/src/make.inc file. Or you are providing those option in somewhere else? I am asking because -DWAVE_CURRENT_INTERACTION is turned off in that file and NUOPC cap is complaining about it. So, I think that needs to be activated when NUOPC cap is compiled. Any idea? Thanks.

BTW: We could look at the issue of using consistent installation of libraries like Metis etc. I'll probably come up a solution that could be specific to ufs-costal.

[uturuncoglu]

@pvelissariou1 Enabling -DWAVE_CURRENT_INTERACTION also requires petsc (tires to compile FVCOM/src/mod_petsc.F). So, probably I am trying to compile the wrong configuration.

[pvelissariou1]

@ufuk please look at NEMS/src/incmake/component_FVCOM.mk. This where I define the flags for the make.inc file in FVCOM.You need to supply the -DNUOPC flag for make.inc, otherwise within NUOPC the compilation won't be successful. On Thursday, June 8, 2023, Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 Enabling -DWAVE_CURRENT_INTERACTION also requires petsc (tires to compile FVCOM/src/mod_petsc.F). So, probably I am trying to compile the wrong configuration.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.< https://ci4.googleusercontent.com/proxy/vvp1o2dsJRZZfVZO-4W3T0hVbf-ZqqYugQnlvHpgsH_dcKwQ3-GKNeb5fkTzTwWn19cGQjY4PMOemcmF4QBSoRsO8BWbR7sJ6j6kojhEfKg-bMDWqe553qWL4_oVdxAK_Og0NWKZwGF0ZgKA3gu-FJPUGLLN4E4v_LhYw_jsr77PoAAy9I113bkzfYb71CHWUCa4qEpExoESkwTw5gwE2i1InJQm2fgBsuMuZiHP_1sH3HvDbQ=s0-d-e1-ft#https://github.com/notifications/beacon/APC7TP6VAXSV77TKEKIMUA3XKH6G5A5CNFSM6AAAAAAY3IRNIOWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTS6LLLSO.gif>Message ID: @.***>

-- Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

[uturuncoglu]

I tried to pass -DNUOPCbut then the build also replace use NUOPC to use 1 in the preprocessing I think. Anyway, I'll look at that again to see what happens. Maybe there is an issue in my CMake interface that leads to this strange behavior.

[pvelissariou1]

Ufuk, I think you need to pass a flag FLAGXXX=-DNUOPC Take a look at FVCOM/src/make.inc how it is used.

On Thursday, June 8, 2023, Ufuk Turunçoğlu @.***> wrote:

I tried to pass -DNUOPCbut then the build also replace use NUOPC to use 1 in the preprocessing I think. Anyway, I'll look at that again to see what happens. Maybe there is an issue in my CMake interface that leads to this strange behavior.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.< https://ci4.googleusercontent.com/proxy/Nvnx3OorEUIUa0dyy6nHrlUftfy5W8-vVrIeSQWEUosU1c97wjyeMSpY4Hr8y1ptpMIAc3q23gW85Gon7K7WSoUO0hwzIUBP2W_qSGpwqluVfQiHW37-1JEyl67WC3i7Vh4HXL8QJeHNdUATwcCLfN_ySccVMDd27ww9albVLDhg4vaTP3AY780iA-vcuANn3_tPPq3rvY7HGQbwaSw8etr2xa2XurpiNkKAhtuiyHK1Ch54kg=s0-d-e1-ft#https://github.com/notifications/beacon/APC7TPZODZXUQPFS2F3CSSDXKIBAHA5CNFSM6AAAAAAY3IRNIOWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTS6LNEVE.gif>Message ID: @.***>

-- Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

[uturuncoglu]

@pvelissariou1 I am plaining to replace NUOPC CPP flags with COUPLED since it is creating a issue in my side. I think COUPLED is better acronym for it. Anyway, let me know what you think?

[pvelissariou1]

Yes, please do that. It also needs to be replaced in a a few files in the FVCOM source. It does create issues for me (CoastaApp) as well. It doesn't allow me to build a standalone executable of FVCOM along with the NEMS executable. I have already mentioned this to FVCOM developers. The NUOPC cap uses duplicate code from FVCOM while, the FVCOM code exports array definitions to the cap. The cap needs significant adjustments. I mention these just for the sake of documenting. If I may suggest why don't use something like CPLNUOPC, to be more clear. Just let me know to update the CoastalApp as well.

On Thursday, June 8, 2023, Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 I am plaining to replace NUOPC CPP flags with COUPLED since it is creating a issue in my side. I think COUPLED is better acronym for it. Anyway, let me know what you think?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.< https://ci3.googleusercontent.com/proxy/QvecpRMfetG9CpsC1npNh_lrrVG2sKY1gV7EcyBsPSUJD5WTHLMrCGjA77fzIlPnewyUGqdnD0gdyfpsZ2o7osvpjV51VZvfCmiVh8aURlyygoVVpTBnB-8UiRGLSumPtIDpLkpMj37ALAdbmVHRQBjvta2KURIdXy2tbG4b2ACLyoCy7zkJt7kt0pVAYwV5B_ZCkau5EJW-DGk1Qm7rvG979yKdKSTNrjQ-p-_3LwqZ3Cpq0Q=s0-d-e1-ft#https://github.com/notifications/beacon/APC7TPZUYSFHM6XSNNYS2ATXKJAYRA5CNFSM6AAAAAAY3IRNIOWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTS6MCTPK.gif>Message ID: @.***>

-- Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

[uturuncoglu]

@pvelissariou1 Okay. I will changes those and create new fork from yours. So, you could merge it later and make required changes in the CoastalApp.

[uturuncoglu]

@pvelissariou1 @saeed-moghimi-noaa It took longer than I expected but now i could able to build FVCOM under ufs-coastal. I pushed all the developments related to FVCOM to the ufs-coastal. I also changed cpp coupling flag for FVCOM (from NUOPC to FVCOM_COUPLED). My FVCOM fork can be found in here https://github.com/uturuncoglu/FVCOM/tree/hotfix/ufs. This is forked from Takis's fork (main branch) and the ufs-coastal points my FVCOM fork. Once Takis merges these changes to his fork, I could again point Takis's fork in ufs-coastal. BTW, I have not tried to run the executable yet but I'll do it as a next step and then I am playing to port the FVCOM related tests from Coastal App test suite to UFS RT system (of course coupled configurations will use NUOPC connectors until we fix the issue in ADCIRC side and port that fix to FVCOM). Anyway, let me know if you have any suggestion. I'll be out tomorrow (6/15) and next Monday since it is holiday.

[pvelissariou1]

@Ufuk Turuncoglu @.> Ufuk, thank you for the update, I'll review and merge the changes into my fork. @Jianhua Qi @.> Jianhua, next week we need to review the changes to the FVCOM source code in my fork and merge them into FVCOM's repo.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Fri, Jun 16, 2023 at 12:43 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 @saeed-moghimi-noaa https://github.com/saeed-moghimi-noaa It took longer than I expected but now i could able to build FVCOM under ufs-coastal. I pushed all the developments related to FVCOM to the ufs-coastal. I also changed cpp coupling flag for FVCOM (from NUOPC to FVCOM_COUPLED). My FVCOM fork can be found in here https://github.com/uturuncoglu/FVCOM/tree/hotfix/ufs. This is forked from Takis's fork (main branch) and the ufs-coastal points my FVCOM fork. Once Takis merges these changes to his fork, I could again point Takis's fork in ufs-coastal. BTW, I have not tried to run the executable yet but I'll do it as a next step and then I am playing to port the FVCOM related tests from Coastal App test suite to UFS RT system (of course coupled configurations will use NUOPC connectors until we fix the issue in ADCIRC side and port that fix to FVCOM). Anyway, let me know if you have any suggestion. I'll be out tomorrow (6/15) and next Monday since it is holiday.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1594136468, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP635CZJTCXGQEXZG7LXLPW53ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[pvelissariou1]

@uturuncoglu Ufuk, the changes you implemented (NUOPC -> FVCOM_COUPLED) have been merged into my fork (https://github.com/pvelissariou1/FVCOM). I fixed the missing FVCOM_COUPLED from some files (fortran and make). The libs directory has been restructured (and a bug in julian has been fixed). All these changes from my fork will be propagated (after reviewing) upstream to the FVCOM repo. After the changes, compilation of all components was successful. You may delete your fork.

[uturuncoglu]

@pvelissariou1 That is great. Thanks. I'll point your fork again and test it.

[uturuncoglu]

@pvelissariou1 I am trying to point your fork for FVCOM but now I am getting error from fproj. Did you change the versions of the libraries under libs/ directory. Something seems changed but I don't know what at this point. I am trying to resolve the issue.

[pvelissariou1]

@Ufuk Turuncoglu @.***> I have compiled FVCOM and fproj without any issue. What error(s) are you getting?

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Wed, Jun 21, 2023 at 5:19 PM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I am trying to point your fork for FVCOM but now I am getting error from fproj. Did you change the versions of the libraries under libs/ directory. Something seems changed but I don't know what at this point. I am trying to resolve the issue.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1601753020, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP4PSPM54QYMXEM47ATXMNXN3ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 I am getting error like following,

make[3]: Entering directory '/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj_bld-prefix/src/fproj_bld-build'
if mpicc -DPACKAGE_NAME=\"libfproj4\" -DPACKAGE_TARNAME=\"libfproj4\" -DPACKAGE_VERSION=\"1.0\" -DPACKAGE_STRING=\"libfproj4\ 1.0\" -DPACKAGE_BUGREPORT=\"Magnus.Hagdorn@ed.ac.uk\" -DPACKAGE=\"libfproj4\" -DVERSION=\"1.0\" -DHAVE_LIBM=1 -DHAVE_LIBPROJ=1  -I. -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj   -DIFORT -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/include -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/include  -g -O2 -MT fort-proj.o -MD -MP -MF ".deps/fort-proj.Tpo" -c -o fort-proj.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj/fort-proj.c; \
then mv -f ".deps/fort-proj.Tpo" ".deps/fort-proj.Po"; else rm -f ".deps/fort-proj.Tpo"; exit 1; fi
mpif90  -g -c -o proj4.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj/proj4.f90
make[3]: *** No rule to make target 'proj4.mod', needed by 'all-am'.  Stop.
make[3]: *** Waiting for unfinished jobs....

I could also try with old version of FVCOM (my fork) to see what happens. That was working.

[pvelissariou1]

@Ufuk Turuncoglu @.***> Did you build the PROJ4 library first? You need to point to the location of the installed PROJ4 library when building fproj. See the function compileFVCOMLibs in the CoastalApp/scripts/functions_build file.

Panagiotis Velissariou, Ph.D., P.E. UCAR Scientist National Ocean and Atmospheric Administration National Ocean Service Office of Coast Survey CSDL/CMMB Physical Scientist - Project Lead cell: (205) 227-9141 email: @.***

On Thu, Jun 22, 2023 at 12:02 AM Ufuk Turunçoğlu @.***> wrote:

@pvelissariou1 https://github.com/pvelissariou1 I am getting error like following,

make[3]: Entering directory '/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj_bld-prefix/src/fproj_bld-build' if mpicc -DPACKAGE_NAME=\"libfproj4\" -DPACKAGE_TARNAME=\"libfproj4\" -DPACKAGE_VERSION=\"1.0\" -DPACKAGE_STRING=\"libfproj4\ 1.0\" @.\" -DPACKAGE=\"libfproj4\" -DVERSION=\"1.0\" -DHAVE_LIBM=1 -DHAVE_LIBPROJ=1 -I. -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj -DIFORT -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/include -I/glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/include -g -O2 -MT fort-proj.o -MD -MP -MF ".deps/fort-proj.Tpo" -c -o fort-proj.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj/fort-proj.c; \ then mv -f ".deps/fort-proj.Tpo" ".deps/fort-proj.Po"; else rm -f ".deps/fort-proj.Tpo"; exit 1; fi mpif90 -g -c -o proj4.o /glade/scratch/turuncu/COASTAL_APP/ufs-coastal/tests/build_fv3_coastal/FVCOM-interface/fproj/proj4.f90 make[3]: No rule to make target 'proj4.mod', needed by 'all-am'. Stop. make[3]: *** Waiting for unfinished jobs....

I could also try with old version of FVCOM (my fork) to see what happens. That was working.

— Reply to this email directly, view it on GitHub https://github.com/oceanmodeling/ufs-coastal/issues/6#issuecomment-1602012825, or unsubscribe https://github.com/notifications/unsubscribe-auth/APC7TP6JTIAZUTQ7ZJL326TXMPGU5ANCNFSM6AAAAAAY3IRNIM . You are receiving this because you were mentioned.Message ID: @.***>

[uturuncoglu]

@pvelissariou1 Yes. I am building project first. I think the issue is just happening wench I build with multiple process (make with -j). If I build it with single process then it builds without any issue. There must be some dependency that I am missing. I'll look at it tomorrow. Thanks.

[pvelissariou1]

@uturuncoglu , Ufuk I have ported the FVCOM 2D case (scituateharbor.atm2fvc) in CoastalApp-testsuite. You still need to download the common data from: wget https://tacc-nos-coastalapp-testsuit.s3.amazonaws.com/hsofs-data-v3.tgz to run this case.

[uturuncoglu]

@pvelissariou1 Thanks. I'll working on now and let you know if I have any issue.

[uturuncoglu]

@pvelissariou1 I am trying to run scituateharbor.atm2fvc at this point but I think sciww3_forcing.nc is missing. This is used by ATMMESH right? If so and if it is vdercotral file. Could you provide me the original dataset (on lat-lon grid)?

[uturuncoglu]

@pvelissariou1 BTW, why fvcom_name: sci parameter in nems.configure is not part of the OCN_attributes:: group?