[petsc-users] Calling convention from windows VC++ and IFC
Satish Balay
balay at mcs.anl.gov
Thu Mar 3 11:01:08 CST 2011
You can try adding the following to
/cygdrive/d/dev/petsc-3.1-p7/cygwin-cxx-debug/include/petscconf.h
and rebuild the libraries [witout rerunning configure] and
see if the fortran examples work.
#ifndef PETSC_STDCALL
#define PETSC_STDCALL __stdcall
#endif
Ideally configure should set this if -Gz was specified to CC aswell -
but changing CFLAGS is tricky on windows, and there is no equivalent
of CXX_CXXFLAGS - so the alternative would be COPTFLAGS.
However you could try the default build [witout any -Gz] - and send
all the logs to petsc-maint - and we can try figuring out the issue.
Satish
On Thu, 3 Mar 2011, Alexander Grayver wrote:
> Really! Thanks Satish. Errors apparently are different when I use stdcall
> convention for compiling instead of default one. But the question is how to
> combine these interfaces now?
>
> On 03.03.2011 16:48, Satish Balay wrote:
> > Oh - I see that with --CXX_CXXFLAGS=-Gz - you get different errors - not
> > blaslapack errors.
> >
> > > > > > libpetsc.lib(fuser.o) : error LNK2019: unresolved external symbol
> > > > > > __imp_GetUserNameA referenced in function "int __cdecl
> > > > > > PetscGetUserName(char *
> > > > > > const,unsigned__int64)" (?PetscGetUserName@@YAHQEAD_K at Z)
> > This is due to compiling C interface with -Gz - so now fortran
> > interface is unable to link with the c interface.
> >
> > Satish
> >
> > On Thu, 3 Mar 2011, Satish Balay wrote:
> >
> > > Currently I don't have access to ifort to try.
> > >
> > > For one - its strange that c/c++ examples work - but not fortran - esp
> > > because all blas calls are from the same c/c++ PETSc code.
> > >
> > > One thing you can try:
> > >
> > > cd src/ksp/ksp/examples/tutorials
> > > make ex2
> > > make ex2f
> > >
> > > [if ex2 works and ex2f fails - edit makefile - and change target ex2f:
> > > from using FLINKER to CLINKER and see if it makes a difference]
> > >
> > > Satish
> > >
> > > On Thu, 3 Mar 2011, Alexander Grayver wrote:
> > >
> > > > Actually, I've added this flag after sending first message to mailing
> > > > list
> > > > (sorry for inconveniences) because I though it could help, unfortunately
> > > > it
> > > > didn't.
> > > >
> > > > The problem probably could be reproduced compiling petsc using
> > > > --with-clanguage=cxx and VC++ and then trying to run FORTRAN examples.
> > > >
> > > > On 03.03.2011 16:26, Satish Balay wrote:
> > > > > perhaps you can remove '--CXX_CXXFLAGS=-Gz' and retry.
> > > > >
> > > > > Satish
> > > > >
> > > > > On Thu, 3 Mar 2011, Alexander Grayver wrote:
> > > > >
> > > > > > I've sent them, thanks!
> > > > > >
> > > > > > I also must note that this is a problem for fortran examples only!
> > > > > > Because
> > > > > > when I've done "make test" after building the situation was:
> > > > > >
> > > > > > mt at agrayver /cygdrive/d/dev/petsc-3.1-p7
> > > > > > $ make PETSC_DIR=/cygdrive/d/dev/petsc-3.1-p7
> > > > > > PETSC_ARCH=cygwin-cxx-debug
> > > > > > test
> > > > > > Running test examples to verify correct installation
> > > > > > C/C++ example src/snes/examples/tutorials/ex19 run successfully with
> > > > > > 1 MPI
> > > > > > process
> > > > > > C/C++ example src/snes/examples/tutorials/ex19 run successfully with
> > > > > > 2 MPI
> > > > > > processes
> > > > > > --------------Error detected during compile or
> > > > > > link!-----------------------
> > > > > > See
> > > > > > http://www.mcs.anl.gov/petsc/petsc-2/documentation/troubleshooting.html
> > > > > > /cygdrive/d/dev/petsc-3.1-p7/bin/win32fe/win32fe ifort -c -MT -Z7
> > > > > > -fpp
> > > > > > -I/cygdrive/d/dev/petsc-3.1-p7/cygwin-cxx-debug/include
> > > > > > -I/cygdrive/d/dev/petsc-3.1-p7/include
> > > > > > -I/cygdrive/d/Dev/HPC_2008_SDK/Include
> > > > > > -I/cygdrive/d/dev/petsc-3.1-p7/cygwin-cxx-debug/include
> > > > > > -I/cygdrive/d/Dev/HPC_2008_SDK/Include -o ex5f.o ex5f.F
> > > > > >
> > > > > > /cygdrive/d/dev/petsc-3.1-p7/bin/win32fe/win32fe ifort -MT -Z7 -fpp
> > > > > > -o
> > > > > > ex5f
> > > > > > ex5f.o -L/cygdrive/d/dev/petsc-3.1-p7/cygwin-cxx-debug/lib
> > > > > > -L/cygdrive/d/dev/petsc-3.1-p7/cygwin-cxx-debug/lib -lpetsc
> > > > > > -L/cygdrive/d/dev/petsc-3.1-p7/cygwin-cxx-debug/lib -lflapack
> > > > > > -lfblas
> > > > > > /cygdrive/d/Dev/HPC_2008_SDK/Lib/amd64/msmpi.lib
> > > > > > /cygdrive/d/Dev/HPC_2008_SDK/Lib/amd64/msmpifec.lib Gdi32.lib
> > > > > > User32.lib
> > > > > > Advapi32.lib Kernel32.lib Ws2_32.lib
> > > > > >
> > > > > > libpetsc.lib(fuser.o) : error LNK2019: unresolved external symbol
> > > > > > __imp_GetUserNameA referenced in function "int __cdecl
> > > > > > PetscGetUserName(char *
> > > > > > const,unsigned__int64)" (?PetscGetUserName@@YAHQEAD_K at Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_ReleaseDC referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct_p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_BitBlt referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct _p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_ExtFloodFill referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct _p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_SelectObject referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct _p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_CreateCompatibleBitmap referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct _p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_CreateCompatibleDC referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct _p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_GetDC referenced in function "int __cdecl
> > > > > > PetscDrawSetDoubleBuffer_Win32(struct _p_PetscDraw *)"
> > > > > > (?PetscDrawSetDoubleBuffer_Win32@@YAHPEAU_p_PetscDraw@@@Z)
> > > > > > libpetsc.lib(win32draw.o) : error LNK2019: unresolved external
> > > > > > symbol
> > > > > > __imp_UpdateWindow referenced in function "int __cdecl
> > > > > > PetscDrawRectangle_Win32(struct _p_PetscDraw
> > > > > > *,double,double,double,double,int,int,int,int)"
> > > > > > (?PetscDrawRectangle_Win32@@YAHPEAU_p_PetscDraw@@NNNNHHHH at Z)
> > > > > >
> > > > > > On 03.03.2011 14:49, Satish Balay wrote:
> > > > > > > can you send the corresponding configure.log and make.log to
> > > > > > > petsc-maint at mcs.anl.gov?
> > > > > > >
> > > > > > > Satish
> > > > > > >
> > > > > > > On Thu, 3 Mar 2011, Alexander Grayver wrote:
> > > > > > >
> > > > > > > > Hello,
> > > > > > > >
> > > > > > > > I built PETSc-3.1 using VC++ 2008 and IFC 10.1 with the
> > > > > > > > following
> > > > > > > > configuration line:
> > > > > > > >
> > > > > > > > ./config/configure.py --with-cc='win32fe cl' --with-fc='win32fe
> > > > > > > > ifort'
> > > > > > > > --with-cxx='win32fe cl' --download-f-blas-lapack=1
> > > > > > > > --with-precision=double
> > > > > > > > --with-scalar-type=complex --with-clanguage=cxx
> > > > > > > > --with-mpi-include="/cygdrive/d/Dev/HPC_2008_SDK/Include"
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --with-mpi-lib=["/cygdrive/d/Dev/HPC_2008_SDK/Lib/amd64/msmpi.lib","/cygdrive/d/Dev/HPC_2008_SDK/Lib/amd64/msmpifec.lib"]
> > > > > > > > --useThreads=0 --with-shared=0
> > > > > > > >
> > > > > > > > Note there is C++ as main language.
> > > > > > > >
> > > > > > > > Now, when I'm trying to build examples I always get similar
> > > > > > > > errors:
> > > > > > > >
> > > > > > > > Error 2 error LNK2019: unresolved external symbol ZGESVD
> > > > > > > > referenced
> > > > > > > > in
> > > > > > > > function "int __cdecl
> > > > > > > > KSPComputeExtremeSingularValues_GMRES(struct
> > > > > > > > _p_KSP
> > > > > > > > *,double *,double *)"
> > > > > > > > (?KSPComputeExtremeSingularValues_GMRES@@YAHPEAU_p_KSP@@PEAN1 at Z)
> > > > > > > > libpetsc.lib
> > > > > > > > Error 3 error LNK2019: unresolved external symbol ZGEEV
> > > > > > > > referenced
> > > > > > > > in
> > > > > > > > function "int __cdecl KSPComputeEigenvalues_GMRES(struct _p_KSP
> > > > > > > > *,int,double
> > > > > > > > *,double *,int *)"
> > > > > > > > (?KSPComputeEigenvalues_GMRES@@YAHPEAU_p_KSP@@HPEAN1PEAH at Z)
> > > > > > > > libpetsc.lib
> > > > > > > > Error 4 error LNK2001: unresolved external symbol ZGEEV
> > > > > > > > libpetsc.lib
> > > > > > > > Error 5 error LNK2019: unresolved external symbol DZNRM2
> > > > > > > > referenced
> > > > > > > > in
> > > > > > > > function "int __cdecl VecNorm_MPI(struct _p_Vec *,enum
> > > > > > > > NormType,double
> > > > > > > > *)"
> > > > > > > > (?VecNorm_MPI@@YAHPEAU_p_Vec@@W4NormType@@PEAN at Z)
> > > > > > > > libpetsc.lib
> > > > > > > > Error 6 error LNK2001: unresolved external symbol DZNRM2
> > > > > > > > libpetsc.lib
> > > > > > > > Error 7 error LNK2019: unresolved external symbol DZASUM
> > > > > > > > referenced
> > > > > > > > in
> > > > > > > > function "int __cdecl VecNorm_Seq(struct _p_Vec *,enum
> > > > > > > > NormType,double
> > > > > > > > *)"
> > > > > > > > (?VecNorm_Seq@@YAHPEAU_p_Vec@@W4NormType@@PEAN at Z)
> > > > > > > > libpetsc.lib
> > > > > > > > ...
> > > > > > > >
> > > > > > > > It look like calling conventions are in disagreement. Any ideas?
> > > > > > > > May
> > > > > > > > be I
> > > > > > > > should specify conventions myself in configuration line? If so,
> > > > > > > > how to
> > > > > > > > do
> > > > > > > > that?
> > > > > > > >
> > > >
> > >
>
>
More information about the petsc-users
mailing list