Hello,I have compiled Siesta with the arch.make file below. I am trying to get the mpi version working. It compiles and runs fine when I type make into the Tests directory (serial). But when I copy the fe.fdf file from the test directory with the pseudo from the Examples/Vps directory and try to run it with two processors (mpiexec) I get a segmentation fault just after the first memory allocation in the first CG step. (I have attached the fe.out file and the err.pbs file) The same error occurs when I run the h2o job from the Examples directory except it occurs in the MD step instead of the CG step.
I would greatly appreciate any help. Thank you Abram Van Der Geest
err.pbs
Description: Binary data
fe.out
Description: Binary data
________________________________________________________ SUFFIXES: .SUFFIXES: .f .F .o .a .f90 .F90 #Arch SIESTA_ARCH=x86_64-ra-linux-gnu-Intel-openmpi #Preprocessor SUFFIX=.f90 CPP_ = ./preprocess <$*.F | /usr/bin/cpp -P -C -traditional >$*$(SUFFIX) CPP = $(CPP_) \ -DMPI \ -DMPI_BLOCK=1048576 \ -DNGZhalf \ -Duse_collective FPP=$(CPP) FPP_OUTPUT= #Fortan Compiler FC=mpif90 RANLIB=ranlib SYS=cpu_time SP_KIND=4 DP_KIND=8 KINDS=$(SP_KIND) $(DP_KIND) FFLAGS=-O3 -xT FPPFLAGS= -DFC_HAVE_FLUSH -DFC_HAVE_ABORT LDFLAGS= ARFLAGS_EXTRA= FCFLAGS_fixed_f= FCFLAGS_free_f90= FPPFLAGS_fixed_F= FPPFLAGS_free_F90= #LIBS MKL = /opt/intel/mkl/10.0.1.014 BLAS_LIBS= -L$(MKL)/lib/em64t -lmkl_intel_lp64 -lmkl_intel_thread \ -lmkl_core -liomp5 -lpthread LAPACK_LIBS=-L$(MKL)/lib/em64t -lmkl_intel_lp64 -lmkl_intel_thread \ -lmkl_lapack -lmkl_core -liomp5 -lpthread BLACS_LIBS=#included in the scalapack listSCALAPACK_LIBS=-L$(MKL)/lib/em64t -lmkl_scalapack_lp64 - lmkl_blacs_lp64 \
-lmkl_intel_lp64 -lmkl_intel_thread -lmkl_lapack -lmkl_core \ -liomp5 -lpthread MPI_LIBS=-L/usr/mpi/intel/openmpi-1.2.5/lib/ -lmpi COMP_LIBS= NETCDF_LIBS=LIBS=$(SCALAPACK_LIBS) $(BLACS_LIBS) $(LAPACK_LIBS) $(BLAS_LIBS) $ (NETCDF_LIBS) $(MPI_LIBS)
#MPI stuff hereMPI_INTERFACE= libmpi_f90.a#/usr/mpi/intel/openmpi-1.2.5/lib/ libmpi_f90.so
MPI_INCLUDE= /usr/mpi/intel/openmpi-1.2.5/include/ DEFS_MPI=-DMPI DEFS= $(DEFS_CDF) $(DEFS_MPI) #Dependency rules are created by autoconf according to whether #discrete preprocessing is necessary or not. .F.o:$(FC) -c $(FFLAGS) $(INCFLAGS) $(FPPFLAGS) $ (FPPFLAGS_fixed_F) $(DEFS) $<
.F90.o: $(FC) -c $(FFLAGS) $(INCFLAGS) $(FPPFLAGS) $(FPPFLAGS_free_F90) $< .f.o: $(FC) -c $(FFLAGS) $(INCFLAGS) $(FCFLAGS_fixed_f) $(DEFS) $< .f90.o: $(FC) -c $(FFLAGS) $(INCFLAGS) $(FCFLAGS_free_f90) $<