[molpro-user] MPI Molpro 2009.1/2 on MacPro i7
Andy May
MayAJ1 at cardiff.ac.uk
Tue Aug 3 08:25:47 BST 2010
Łukasz,
The first problem (incompatibility with icc) we know about. On Linux if
one builds object files with GCC then other C compilers can easily link
against them. This is not the case on Mac, so one must always use GCC.
This should cause no real problems since there is very little C code in
Molpro at present, so any performance gain would be minimal.
The second problem is not related to Molpro, but rather Intel's
compilers, which are not compatible with the latest xcode. See this
previous post:
http://www.molpro.net/pipermail/molpro-user/2010-June/003784.html
Best wishes,
Andy
On 03/08/10 08:07, Łukasz Rajchel wrote:
> Dear Molpro users,
>
> I've compiled both Molpro 2009.1 and 2009.2 (patchlevel 1583) on
> MacPro i7. I've used the MacPorts' MPICH2 library and GA-4-3-2. I have
> used two sets of compilers:
>
> 1 gnu (gcc + gfortran),
> 2 Intel (icc + ifort).
>
> The gnu installation worked well and Molpro's running allright.
> However, I came across some problems using Intel compilers.
>
> First, I couldn't use both icc and ifort. More specifically, using
>
> FC=/opt/intel/Compiler/11.1/088/bin/intel64/ifort
> CC=/opt/intel/Compiler/11.1/088/bin/intel64/icc
>
> I got the following message at the config stage:
>
> ----------------------------------------------
> unable to link with parse object using:
>
> /opt/intel/Compiler/11.1/088/bin/intel64/ifort -i8 -r8 -pc64 -auto -
> Vaxlib -vec-report0 -cxxlib -O0 -lz
>
> Please check the compiler is working, flags are correct and all
> libraries are present
>
> failed to determine PARSE
> ----------------------------------------------
>
> Then I tried to compile Molpro with gcc/ifort combination, and it
> worked, with the following CONFIG:
>
> ----------------------------------------------
> # MOLPRO CONFIG generated at Mon Aug 2 20:31:28 EDT 2010, for host
> goose.seb.oakland.edu
>
> CONFIGURE_OPTIONS="-gcc" "-ifort" "-mpp" "-mppbase" "/usr/local/ga/
> ga-4-3-2" "-gforker" "-var" "BLASLIB=-L/Library/Frameworks/
> Intel_MKL.framework/Libraries/em64t /Library/Frameworks/
> Intel_MKL.framework/Libraries/em64t/libmkl_solver_ilp64.a -
> lmkl_intel_ilp64 -lmkl_intel_thread -lmkl_core -openmp - lpthread" "-
> var" "INSTBIN=exec" "-var" "INSTLIB=aux" "-var" "INSTHTML=manual"
>
> AR=ar
> ARCHNAME=Darwin/x86_64
> ARFLAGS=-rS
> AWK=awk
> BIBTEX=
> BLASLIB=-L/Library/Frameworks/Intel_MKL.framework/Libraries/em64t /
> Library/Frameworks/Intel_MKL.framework/Libraries/em64t/
> libmkl_solver_ilp64.a -lmkl_intel_ilp64 -lmkl_intel_thread -lmkl_core -
> openmp -lpthread
> BUILD=p
> CAT=cat
> CC=/usr/bin/gcc
> CCVERSION=4.2.1
> CC_FRONT=
> CDEBUG=-g
> CDEFINE=-D
> CFLAGS=-m64 -DINT64 -DZLIB
> CLEARSPEEDLIB=
> CMPPINCLUDE=/usr/local/ga/ga-4-3-2/include
> COPT=-O3
> COPT0=-O0
> COPT1=-O1
> COPT2=-O2
> COPT3=-O3
> CP=cp -p
> CPROFILE=-p
> CSCN=cscn
> CSFLAGS=-O3 -I. --dynamic
> CUDACC=
> CUDACCVERSION=
> CUDACDEBUG=-g
> CUDACDEFINE=-D
> CUDACFLAGS=
> CUDACOPT=
> CUDACOPT0=-O0
> CUDACOPT1=-O1
> CUDACOPT2=-O2
> CUDACOPT3=-O3
> CUDACPROFILE=-p
> CXX=$(CC)
> CXXFLAGS=$(CFLAGS)
> DOXYGEN=
> ECHO=/bin/echo
> EXPORT=export
> F90FLAGS=
> FC=/opt/intel/Compiler/current/bin/intel64/ifort
> FCVERSION=11.1
> FDEBUG=-g
> FDEFINE=-D
> FFLAGS=-i8 -r8 -pc64 -auto -Vaxlib -vec-report0 -cxxlib
> FOPT=-O3
> FOPT0=-O0
> FOPT1=-O1
> FOPT2=-O2
> FOPT3=-O3
> FPP=-fpp
> FPROFILE=-p
> FSTATIC=
> FTCFLAGS=molpro unix unix-i8 Darwin lapack mpp blas
> HDF5INCLDE=
> HDF5LIB=
> HOSTFILE_FORMAT=
> INSTBIN=exec
> INSTHTML=manual
> INSTLIB=aux
> INTEGER=8
> LAPACKLIB=
> LATEX2HTML=
> LAUNCHER=/opt/local/bin/mpiexec -np %n %x
> LD_ENV=/Library/Frameworks/Intel_MKL.framework/Libraries/em64t
> LD_ENVNAME=DYLD_LIBRARY_PATH
> LIBRARY_SUFFIX=a
> LIBS=-lz
> LIBS_FRONT=
> LINKOPT=-Wl,-framework,CoreFoundation -Wl,-framework,IOKit
> LINKOPT_FRONT=
> LN=ln -s
> MACROS=MOLPRO MOLPRO_ifort GA_TOOLS GA_TCGMSG_MPI _I8_ BLAS_INT=8
> LAPACK_INT=8 MOLPRO_FORCE_VECTOR MOLPRO_NEXTSCALAR
> MOLPRO_NO_RECURRENCE MOLPRO_NOVECTOR MOLPRO_SHORTLOOP _MOLCAS_MPP_
> MAKEDEPEND_OPTIONS=
> MAKEINDEX=
> MAPLE=
> MKDIR=mkdir -p
> MODULE_FLAG=-I
> MODULE_SUFFIX=mod
> MOLPRO_CACHE=12288
> MOLPRO_FLOPDGM=0386
> MOLPRO_FLOPDGV=5432
> MOLPRO_FLOPMXM=5247
> MOLPRO_FLOPMXV=2369
> MOLPRO_MINDGC=12
> MOLPRO_MINDGL=3
> MOLPRO_MINDGM=16
> MOLPRO_MINDGR=1
> MOLPRO_MINDGV=12
> MOLPRO_MXMBLK=64
> MOLPRO_MXMBLN=64
> MOLPRO_UNROLL=2
> MPILIB=-pipe -O2 -m64 -L/opt/local/lib -I/opt/local/include/mpich2 -L/
> opt/local/lib -L/opt/local/lib -lpmpich -lmpich -lopa -lpthread
> MPPLIB=-L/usr/local/ga/ga-4-3-2/lib/MACX64 -ltcgmsg-mpi -lglobal -lma -
> lpario -larmci
> OBJECT_SUFFIX=o
> OPT0=explicit_util.f copyc6.f
> OPT1=koopro4.F rmp2_f12_drv2.f90
> OPT2=integrals.f90 basis_integral_shells.f geminal.f
> OPT3=
> PAPER=a4paper
> PARSE=parse-Darwin-x86_64-i8.o
> PATCHER=patcher.exe
> PDFLATEX=
> PNAME=molprop_2009_2_Darwin_x86_64_i8
> PTSIZE=11
> RANLIB=ranlib -c
> RM=rm -rf
> SHELL=/bin/sh
> STRIP=echo WARNING, strip not used; true
> SUFFIXES=f F f90 c cpp
> TAR=tar -cf
> UNTAR=tar -xf
> VERSION=2009.2
> XSD=/opt/local/bin/xmllint --noout --schema
> XSLT=/opt/local/bin/xsltproc
> YACC=bison -b y
>
> .SUFFIXES:
> MAKEFLAGS+=-r
>
> ----------------------------------------------
>
> The compilation was succesful, however, the resulting Molpro
> executable doesn't work. When I try to run Molpro, I get the errors:
>
> ----------------------------------------------
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image PC Routine Line
> Source
> molpro.exe 000000010044997A Unknown Unknown
> Unknown
> molpro.exe 00000001004046DD Unknown Unknown
> Unknown
> molpro.exe 000000010048FEB2 Unknown Unknown
> Unknown
> molpro.exe 0000000100000FDC Unknown Unknown
> Unknown
> molpro.exe 0000000100000F74 Unknown Unknown
> Unknown
> ----------------------------------------------
>
> I tried to compile Molpro with MPICH2 (no GA) and sequential MKL
> (i.e., MPI only), the results were the same. I know the error is not
> particularly informative, but maybe someone has had a similar problem
> and has a workaround.
>
> Thank you very much for your kind answers,
>
> best regards,
>
> Łukasz Rajchel
> Quantum Chemistry Laboratory
> Warsaw University
> http://tiger.chem.uw.edu.pl/staff/lrajchel/index.html
>
>
>
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user
More information about the Molpro-user
mailing list