I'm on Ubuntu 13.04 64-bit and I'm attempting to build a "Hello world" program in PETSc using CMake.
I have the following program solve1.c
(loosely based on ex1.c from the PETSc examples), which solves a simple system of linear equations:
/* minimal example of solving a linear equation system in petsc
(in serial) */
/* based on http://www.mcs.anl.gov/petsc/petsc-current/src/ksp/ksp/examples/tutorials/ex1.c.html */
#include <petscksp.h>
#define SIZE 3
int main(int argc,char **argv) {
Vec x,b; /* approx solution, rhs */
Mat A; /* linear system matrix */
KSP ksp; /* linear solver context */
PC pc; /* preconditioner context */
int size;
PetscInt col[3],i;
PetscScalar temp[SIZE];
PetscInitialize(&argc,&argv,PETSC_NULL,PETSC_NULL);
MPI_Comm_size(PETSC_COMM_WORLD,&size);
if(size!=1) PetscPrintf(PETSC_COMM_WORLD,"warning, %d threads\n",size);
/* create vectors */
VecCreate(PETSC_COMM_WORLD,&x);
VecSetSizes(x,PETSC_DECIDE,SIZE);
VecSetFromOptions(x);
VecDuplicate(x,&b);
temp[0]=14.5; temp[1]=54; temp[2]=0.423;
for(i=0;i<3;i++) VecSetValues(b,1,&i,&temp[i],INSERT_VALUES);
/* need to assemble after setting values! do necessary
message passing etc to propagate matrix to all ranks */
VecAssemblyBegin(b);
VecAssemblyEnd(b);
/* create matrix */
MatCreate(PETSC_COMM_WORLD,&A);
MatSetSizes(A,PETSC_DECIDE,PETSC_DECIDE,SIZE,SIZE);
MatSetFromOptions(A);
MatSetUp(A);
temp[0]=1; temp[1]=1; temp[2]=1;
col[0]=0; col[1]=1; col[2]=2;
for(i=0;i<3;i++) {
MatSetValues(A,1,&i,3,col,temp,INSERT_VALUES);
temp[i]=0;
}
/* need to assemble matrix for the same reasons as above */
MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY);
MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY);
/* linear solver context! */
KSPCreate(PETSC_COMM_WORLD,&ksp);
/* operator is A matrix, also set matrix for preconditioning here */
KSPSetOperators(ksp,A,A,DIFFERENT_NONZERO_PATTERN);
/* get pc context from ksp context */
KSPGetPC(ksp,&pc);
/* set preconditioner type */
PCSetType(pc,PCJACOBI);
KSPSetTolerances(ksp,1e-6,PETSC_DEFAULT,PETSC_DEFAULT,PETSC_DEFAULT);
KSPSetFromOptions(ksp);
/* solve! */
KSPSolve(ksp,b,x);
/* display everything */
MatView(A,PETSC_VIEWER_STDOUT_WORLD);
VecView(b,PETSC_VIEWER_STDOUT_WORLD);
VecView(x,PETSC_VIEWER_STDOUT_WORLD);
KSPView(ksp,PETSC_VIEWER_STDOUT_WORLD);
/* get rid of everything */
KSPDestroy(&ksp);
VecDestroy(&x);
VecDestroy(&b);
PetscFinalize();
return 0;
}
This is my attempt at creating CMakeLists.txt
(based on various PETSc tutorials and projects found on the web):
CMAKE_MINIMUM_REQUIRED(VERSION 2.8)
PROJECT(helloworld)
SET(CMAKE_MODULE_PATH "${CMAKE_SOURCE_DIR}/cmake-modules")
FIND_PACKAGE(PETSc REQUIRED)
INCLUDE_DIRECTORIES(${PETSC_INCLUDES})
ADD_DEFINITIONS(${PETSC_DEFINITIONS})
ADD_EXECUTABLE(solve1 solve1.c)
TARGET_LINK_LIBRARIES(solve1 ${PETSC_LIBRARIES})
In addition, I have a directory cmake-modules
where I have the files CorrectWindowsPaths.cmake
, FindPETSc.cmake
, FindPackageMultipass.cmake
and ResolveCompilerPaths.cmake
which I have gotten from https://github.com/jedbrown/cmake-modules (as recommended by the PETSc FAQ).
I'm building the project in this way (initially standing in the root directory of my project, which is /home/ruben/hello
):
mkdir build
cd build
cmake ..
make
Unfortunately, the make commands fails with a linker error. Here's the output from CMake (which looks good):
-- The C compiler identification is GNU 4.8.1
-- The CXX compiler identification is GNU 4.8.1
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- petsc_lib_dir /home/ruben/petsc-3.4.2/arch-linux2-c-debug/lib
-- Recognized PETSc install with single library for all packages
-- Performing Test MULTIPASS_TEST_1_petsc_works_minimal
-- Performing Test MULTIPASS_TEST_1_petsc_works_minimal - Success
-- Minimal PETSc includes and libraries work. This probably means we are building with shared libs.
-- Found PETSc: /home/ruben/petsc-3.4.2/arch-linux2-c-debug/include;/home/ruben/petsc-3.4.2/include
-- Configuring done
-- Generating done
-- Build files have been written to: /home/ruben/hello/build
Here's the output from make VERBOSE=1
(which is not good):
/usr/bin/cmake -H/home/ruben/hello -B/home/ruben/hello/build --check-build-system CMakeFiles/Makefile.cmake 0
/usr/bin/cmake -E cmake_progress_start /home/ruben/hello/build/CMakeFiles /home/ruben/hello/build/CMakeFiles/progress.marks
make -f CMakeFiles/Makefile2 all
make[1]: Entering directory `/home/ruben/hello/build'
make -f CMakeFiles/solve1.dir/build.make CMakeFiles/solve1.dir/depend
make[2]: Entering directory `/home/ruben/hello/build'
cd /home/ruben/hello/build && /usr/bin/cmake -E cmake_depends "Unix Makefiles" /home/ruben/hello /home/ruben/hello /home/ruben/hello/build /home/ruben/hello/build /home/ruben/hello/build/CMakeFiles/solve1.dir/DependInfo.cmake --color=
Dependee "/home/ruben/hello/build/CMakeFiles/solve1.dir/DependInfo.cmake" is newer than depender "/home/ruben/hello/build/CMakeFiles/solve1.dir/depend.internal".
Dependee "/home/ruben/hello/build/CMakeFiles/CMakeDirectoryInformation.cmake" is newer than depender "/home/ruben/hello/build/CMakeFiles/solve1.dir/depend.internal".
Scanning dependencies of target solve1
make[2]: Leaving directory `/home/ruben/hello/build'
make -f CMakeFiles/solve1.dir/build.make CMakeFiles/solve1.dir/build
make[2]: Entering directory `/home/ruben/hello/build'
/usr/bin/cmake -E cmake_progress_report /home/ruben/hello/build/CMakeFiles 1
[100%] Building C object CMakeFiles/solve1.dir/solve1.c.o
/usr/bin/cc -D__INSDIR__="" -I/home/ruben/petsc-3.4.2/arch-linux2-c-debug/include -I/home/ruben/petsc-3.4.2/include -o CMakeFiles/solve1.dir/solve1.c.o -c /home/ruben/hello/solve1.c
Linking C executable solve1
/usr/bin/cmake -E cmake_link_script CMakeFiles/solve1.dir/link.txt --verbose=1
/usr/bin/cc CMakeFiles/solve1.dir/solve1.c.o -o solve1 -rdynamic /home/ruben/petsc-3.4.2/arch-linux2-c-debug/lib/libpetsc.so -Wl,-rpath,/home/ruben/petsc-3.4.2/arch-linux2-c-debug/lib
/usr/bin/ld: CMakeFiles/solve1.dir/solve1.c.o: undefined reference to symbol 'MPI_Comm_size'
/usr/bin/ld: note: 'MPI_Comm_size' is defined in DSO /home/ruben/petsc-3.4.2/arch-linux2-c-debug/lib/libmpich.so.10 so try adding it to the linker command line
/home/ruben/petsc-3.4.2/arch-linux2-c-debug/lib/libmpich.so.10: could not read symbols: Invalid operation
collect2: error: ld returned 1 exit status
make[2]: *** [solve1] Error 1
make[2]: Leaving directory `/home/ruben/hello/build'
make[1]: *** [CMakeFiles/solve1.dir/all] Error 2
make[1]: Leaving directory `/home/ruben/hello/build'
make: *** [all] Error 2
The PETSc FAQ recommends looking at CMakeLists.txt
from the Dohp project. Unfortunately, I found this CMake file to be difficult to understand (Dohp is not a particularly small example and it has additional dependencies other than PETSc) and I have not been able to make use of this file to fix my linker error.
My question: What is wrong with my CMakeLists.txt
and how can I fix it? (Or, in the case that my CMake file is correct: What other mistake(s) have I done?)
Some additional information: I'm using PETSc 3.4.2 (installed in /home/ruben/petsc-3.4.2/
, installed with MPI which I want to use), CMake 2.8.10.1, gcc 4.8.1, all on Ubuntu 13.04 64-bit on an Intel i7-3930k. I am able to compile and run the above program using a regular makefile, and I'm able to compile and run all PETSc examples I've tried so far.