search

FreeFem / FreeFem-sources

FreeFEM source code
https://freefem.org/
Other
797 stars 193 forks source link

error during compiling WIN10 + SYS2 +MINGW64 #163

Closed spara7C5 closed 2 years ago

spara7C5 commented 3 years ago

cloning the branch "develop", running makefile -j2 i get:

ranlib ../../../libopenblasp-r0.3.6.a make[8]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty/blas/OpenBLAS/lapack-netlib/LAPACKE/utils' make[7]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty/blas/OpenBLAS/lapack-netlib/LAPACKE' make[6]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty/blas/OpenBLAS/lapack-netlib' make[5]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty/blas/OpenBLAS' touch openblas.done cd OpenBLAS && ../../../build/links2files /bin/sh: line 1: ../../../build/links2files: No such file or directory make[4]: [Makefile:859: links.done] Error 127 make[4]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty/blas' make[3]: [Makefile:573: all] Error 2 make[3]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty/blas' make[2]: [Makefile:593: all-recursive] Error 1 make[2]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources/3rdparty' make[1]: [Makefile:797: all-recursive] Error 1 make[1]: Leaving directory '/home/Brain ADM/Devel/FreeFem-sources' make: *** [Makefile:747: all] Error 2

prj- commented 3 years ago

I'd suggest you compile PETSc which will provide a BLAS implementation so that OpenBLAS won't be compiled. You can type make petsc-slepc in the 3rdparty/ff-petsc folder.

spara7C5 commented 3 years ago

the make petsc-slepc fails:

=============================================================================== Configuring PETSc to compile on your system

TESTING: check from config.libraries(config/BuildSystem/config/libraries.py:157) *** UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log for details):

--with-mpi-lib=[''] and --with-mpi-include=[] did not work

make: *** [Makefile:220: petsc-3.14.2/tag-conf-real] Error 1

spara7C5 commented 3 years ago

the config. log file:

----------------------

Running config.status.

----------------------

This file was extended by FreeFEM config.status 4.7-1, which was generated by GNU Autoconf 2.69. Invocation command line was

CONFIG_FILES = CONFIG_HEADERS = CONFIG_LINKS = CONFIG_COMMANDS = $ ../../config.status --file=Makefile.inc:Makefile-PETSc.inc

on DESKTOP-QJQ3LRV

config.status:1209: creating Makefile.inc

prj- commented 3 years ago

Do you have MS-MPI installed?

spara7C5 commented 3 years ago

Yes, i installed it via the installar msmpisetup.exe, are any configurations required? I still get the same error...

prj- commented 3 years ago

Could you please send FreeFem-sources/config.log?

spara7C5 commented 3 years ago

it is exactly equal to the one above

prj- commented 3 years ago

No, you copy/paste'd FreeFem-sources/3rdparty/ff-petsc/config.log...

spara7C5 commented 3 years ago

sorry for that. I attach FreeFem-sources/config.log config.log

prj- commented 3 years ago

Where did you install MS-MPI? There is not a single reference in your config.log. You should get something like this:

configure:12556: checking for MPIRUN
configure:12615: result: C:\Program Files\Microsoft MPI\Bin\mpiexec.exe
spara7C5 commented 3 years ago

i re-cloned the repo and retried the compile sequence. I still get error in make petsc-slepc. The config.log file even shows the MS-MPI: config.log but i still get:

=============================================================================== Configuring PETSc to compile on your system

TESTING: check from config.libraries(config/BuildSystem/config/libraries.py:157) *** UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log for details):

--with-mpi-lib=[''] and --with-mpi-include=[] did not work

make: *** [Makefile:220: petsc-3.14.2/tag-conf-real] Error 1

prj- commented 3 years ago

No, the config.log does not show a functioning MPI, see at the end: configure:25676: -- with MPI: no. Do you have any files in 3rdparty/include/msmpi? If not, could you try to go into 3rdparty and type make reinstall-msmpi and then reconfigure and make sure you have -- with MPI: yes? You also probably want to add the flag FCFLAGS=-fallow-invalid-boz on your configure line.

Note that this has nothing to do with PETSc, until MPI is not recognized by ./configure, there is no point in going further and trying to compile PETSc (and thus BLAS).