schism-dev / schism

Semi-implicit Cross-scale Hydroscience Integrated System Model (SCHISM)
http://ccrm.vims.edu/schismweb/
Apache License 2.0
93 stars 88 forks source link

CMake fails with ParMetis 4 #15

Closed platipodium closed 4 years ago

platipodium commented 4 years ago

Something's still awkward with the build system, notably the inclusion of the new ParMetis4. Prior to this change, my build succeeded with

cmake ../schism/src -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local

This yields an error in ParMetis, which I tracked down to

cmake ../schism/src/ParMetis-4.0.3/metis -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local

Which gives the error


CMake Error at /Volumes/Kea/devel/schism/build/CMakeFiles/CMakeTmp/CMakeLists.txt:13 (add_executable):
  Cannot find source file:

    GKlib/conf/check_thread_storage.c

  Tried extensions .c .C .c++ .cc .cpp .cxx .cu .m .M .mm .h .hh .h++ .hm
  .hpp .hxx .in .txx

CMake Error at /Volumes/Kea/devel/schism/build/CMakeFiles/CMakeTmp/CMakeLists.txt:13 (add_executable):
  No SOURCES given to target: cmTC_05be2

CMake Error at GKlib/GKlibSystem.cmake:110 (try_compile):
  Failed to generate test project build system.
Call Stack (most recent call first):
  CMakeLists.txt:20 (include)

-- Configuring incomplete, errors occurred!
platipodium commented 4 years ago

For completeness: when trying to compile SCHISM (as before), I get a similar error

cmake ../schism/src -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local


### Adding enabled libraries to project
CMake Error at ParMetis-4.0.3/CMakeLists.txt:28 (include):
  include could not find load file:

    /metis/GKlib/GKlibSystem.cmake

CMake Error at ParMetis-4.0.3/CMakeLists.txt:38 (add_subdirectory):
  add_subdirectory given source "/metis/libmetis" which is not an existing
  directory.

All with today's version (11 May) 6f15b0d63b

platipodium commented 4 years ago

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

josephzhang8 commented 4 years ago

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

jreniel commented 4 years ago

These paths should probably be populated without having to invoke a separate file, since the path to the parmetis source code is fixed.

On 5/12/20 9:22 AM, Joseph Zhang wrote:

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627341559, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEXYQTWBG5LPFJRJMCWEHL3RRFERXANCNFSM4M6RGT4A.

josephzhang8 commented 4 years ago

Hi Jaime:

At the moment we are not pointing to parmetis web site; we just download and include it inside schism repo so the path is different at different users site.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 9:43 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

These paths should probably be populated without having to invoke a separate file, since the path to the parmetis source code is fixed.

On 5/12/20 9:22 AM, Joseph Zhang wrote:

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627341559, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEXYQTWBG5LPFJRJMCWEHL3RRFERXANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627353915, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZZRFPI7OSLJQPT3Y3DRRFG5FANCNFSM4M6RGT4A.

jreniel commented 4 years ago

The relative path is fixed. To me, the standard way of using CMake is by calling it directly, without calling external configuration files. These are useful on some systems, but is not the general standard way.

On 5/12/20 9:45 AM, Joseph Zhang wrote:

Hi Jaime:

At the moment we are not pointing to parmetis web site; we just download and include it inside schism repo so the path is different at different users site.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 9:43 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

These paths should probably be populated without having to invoke a separate file, since the path to the parmetis source code is fixed.

On 5/12/20 9:22 AM, Joseph Zhang wrote:

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on

GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923, or

unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627341559, or unsubscribe

https://github.com/notifications/unsubscribe-auth/AEXYQTWBG5LPFJRJMCWEHL3RRFERXANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627353915, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZZRFPI7OSLJQPT3Y3DRRFG5FANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627355572, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEXYQTRUYAUFTUGFMUSN2P3RRFHH5ANCNFSM4M6RGT4A.

platipodium commented 4 years ago

I suggest to make this conditionally dependent on system availability of ParMetis.  If there is a system ParMetis (which could even be pointed to by an env variable, CMake build system should choose that one, if not, it should chose the (fixed path) one shipped within SCHISM.

Cheers, Carsten

The relative path is fixed. To me, the standard way of using CMake is by calling it directly, without calling external configuration files. These are useful on some systems, but is not the general standard way.

On 5/12/20 9:45 AM, Joseph Zhang wrote:

Hi Jaime:

At the moment we are not pointing to parmetis web site; we just download and include it inside schism repo so the path is different at different users site.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 9:43 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

These paths should probably be populated without having to invoke a separate file, since the path to the parmetis source code is fixed.

On 5/12/20 9:22 AM, Joseph Zhang wrote:

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on

GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923,

or

unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub

https://github.com/schism-dev/schism/issues/15#issuecomment-627341559, or unsubscribe

https://github.com/notifications/unsubscribe-auth/AEXYQTWBG5LPFJRJMCWEHL3RRFERXANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on

GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627353915,

or

unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZZRFPI7OSLJQPT3Y3DRRFG5FANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub

https://github.com/schism-dev/schism/issues/15#issuecomment-627355572, or unsubscribe

https://github.com/notifications/unsubscribe-auth/AEXYQTRUYAUFTUGFMUSN2P3RRFHH5ANCNFSM4M6RGT4A.

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627364764, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAUNNOLI3KMQSJTWZ4PNLBDRRFJDRANCNFSM4M6RGT4A.

-- Dr. rer. nat. Carsten Lemmen Environmental Scientist

Helmholtz Zentrum Geesthacht Centre for Materials and Coastal Research GmbH Institute for Coastal Research - Systems Analysis and Modelling Ecosystem Modeling (KSE) Max-Planck-Str.1, 21502 Geesthacht, Germany

Tel.: +49 4152 87-2013 / Fax: -2020 / Secretary: -1502 Building 11, 2nd floor, room 213 https://w3w.co/builds.grubby.chewable

Email: carsten.lemmen@hzg.de Internet: http://www.hzg.de Messaging: skype://platipodium or aim://platipodium Zoom: https://zoom.us/my/platipodium

My email is as private as regular mail, please 'put it in an envelope' by GPG encryption with my public key id 9249e2e0.

Geschäftsführer: Prof. Dr. W. A. Kaysser, Dipl.-Ing. M. Ganß Vorsitzender des Aufsichtsrates: Ministerialrat W. Kraus Amtsgericht Lübeck - HRB 285 GE USt.IDNr.: DE 135 131 669

Helmholtz-Zentrum Geesthacht Zentrum für Material- und Küstenforschung GmbH Max-Planck-Straße 1 I 21502 Geesthacht I Deutschland/Germany

Geschäftsführung/Board of Management: Prof. Dr. Matthias Rehahn, Silke Simon Vorsitzender des Aufsichtsrates/Chairman of the Supervisory Board: Ministerialdirigent Dr. Herbert Zeisel Amtsgericht Lübeck HRB 285 GE (Register Court) Internet: http://www.hzg.de

josephzhang8 commented 4 years ago

For some reason the relative path did not work for me. For Carsten’s suggestion, I still think it’s better to recompile PM using source for max compatibility, even if it’s available in the system. The compilation takes minimal amount of time anyway. Thx.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 10:01 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

The relative path is fixed. To me, the standard way of using CMake is by calling it directly, without calling external configuration files. These are useful on some systems, but is not the general standard way.

On 5/12/20 9:45 AM, Joseph Zhang wrote:

Hi Jaime:

At the moment we are not pointing to parmetis web site; we just download and include it inside schism repo so the path is different at different users site.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 9:43 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

These paths should probably be populated without having to invoke a separate file, since the path to the parmetis source code is fixed.

On 5/12/20 9:22 AM, Joseph Zhang wrote:

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on

GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923, or

unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627341559, or unsubscribe

https://github.com/notifications/unsubscribe-auth/AEXYQTWBG5LPFJRJMCWEHL3RRFERXANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627353915, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZZRFPI7OSLJQPT3Y3DRRFG5FANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627355572, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEXYQTRUYAUFTUGFMUSN2P3RRFHH5ANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627364764, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ3SI3OSQTJYFU7NMQTRRFJDRANCNFSM4M6RGT4A.

jreniel commented 4 years ago

Any way you do it is fine with me. It's just a comment. The -D options should be used when you want to switch to another version, but CMake itself should pick up by default the one shipped with SCHISM without having to manually enter the paths.

-J.

On 5/12/20 10:07 AM, Joseph Zhang wrote:

For some reason the relative path did not work for me. For Carsten’s suggestion, I still think it’s better to recompile PM using source for max compatibility, even if it’s available in the system. The compilation takes minimal amount of time anyway. Thx.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 10:01 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

The relative path is fixed. To me, the standard way of using CMake is by calling it directly, without calling external configuration files. These are useful on some systems, but is not the general standard way.

On 5/12/20 9:45 AM, Joseph Zhang wrote:

Hi Jaime:

At the moment we are not pointing to parmetis web site; we just download and include it inside schism repo so the path is different at different users site.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Jaime R Calzada notifications@github.com Sent: Tuesday, May 12, 2020 9:43 AM To: schism-dev/schism schism@noreply.github.com Cc: Y. Joseph Zhang yjzhang@vims.edu; Comment comment@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

These paths should probably be populated without having to invoke a separate file, since the path to the parmetis source code is fixed.

On 5/12/20 9:22 AM, Joseph Zhang wrote:

Thx Carsten. I noticed u were not using the ‘standard’ cmake interface with SCHISM.local. and SCHISM.local.build. The full path to ParMETIS should be speicified in SCHISM.local.

-Joseph

Y. Joseph Zhang Web: schism.wiki Office: 804 684 7466

From: Carsten Lemmen notifications@github.com Sent: Tuesday, May 12, 2020 5:56 AM To: schism-dev/schism schism@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [schism-dev/schism] CMake fails with ParMetis 4 (#15)

Solved by adding the ParMetis CMake flags to the global build. So now you can build ParMetis and SCHISM in one step again (note my locally specific location of build and netcdf dirs and adjust accordingly ):

cmake -DCMAKE_Fortran_COMPILER=mpifort -DCMAKE_CXX_COMPILER=mpicc -DCMAKE_C_COMPILER=mpicc -DMPI_CXX_COMPILER=mpicc -DNetCDF_C_DIR=/opt/local -DNetCDF_FORTRAN_DIR=/opt/local -DGKLIB_PATH=../schism/src/ParMetis-4.0.3/metis/GKlib -DMETIS_PATH=../schism/src/ParMetis-4.0.3/metis ../schism/src && make pschism

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on

GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627238923,

or

unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ2D2HTWX4W7JA4JQH3RREMMXANCNFSM4M6RGT4A.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub

https://github.com/schism-dev/schism/issues/15#issuecomment-627341559, or unsubscribe

https://github.com/notifications/unsubscribe-auth/AEXYQTWBG5LPFJRJMCWEHL3RRFERXANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on

GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627353915, or

unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZZRFPI7OSLJQPT3Y3DRRFG5FANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627355572, or unsubscribe

https://github.com/notifications/unsubscribe-auth/AEXYQTRUYAUFTUGFMUSN2P3RRFHH5ANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/schism-dev/schism/issues/15#issuecomment-627364764, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AFBKNZ3SI3OSQTJYFU7NMQTRRFJDRANCNFSM4M6RGT4A.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/schism-dev/schism/issues/15#issuecomment-627368456, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEXYQTWCDFN3LDKLYAUKYUDRRFJ2XANCNFSM4M6RGT4A.