Dear Lan,
I have had the same troubles some weeks ago. First of all, you should
upgrade to pgi 6.2 : it works better,
but be careful, you have also to include the file dfscf.f to the
following section in the arch.make file, relative to
the PGI compiler :
#
#
# Important (at least for V5.0-1 of the pgf90 compiler...)
# Compile atom.f and electrostatic.f without optimization.
#
atom.o:
$(FC) -c $(FFLAGS_DEBUG) $(INCFLAGS) atom.f
#
dfscf.o:
$(FC) -c $(FFLAGS_DEBUG) $(INCFLAGS) dfscf.f
#
electrostatic.o:
$(FC) -c $(FFLAGS_DEBUG) $(INCFLAGS) electrostatic.f
otherwise the forces on the atoms will not be evaluated properly.
Regards,
Phil.
lan haiping a écrit :
Dear All,
I tried to compile siesta-2.0 with pgi 6.0 compilor under x86-64
environment.
With arch.make file that tested for siesta-1.31, i can compile the
siesta-2.0 package without any complain.
But siesta-2.0 cannot finish any computation even for simple h2o
energy caclulation. All error information is
" p4_error: interrupt SIGSEGV: 11".
I do find some information about this problem : memory or stack
issues. And i also ulimite memory to unlimited. But no help, with some
error information. I also try to compile with low level optimization
triggers, but still cannot solve it .
Would you please give me some help or share me with arch.make file
under similar environment ?
my arch.make is below , and i can finish large system calculation
successfully with siesta-1.31p.
Regards,
Hai-Ping
arch.make :
*FC=pgf90
FC_ASIS=$(FC)
#
# -Mlre option gives incorrect numerical results
# default is -Mnolre, but we put it here explicitly, TO REMEMBER!
FFLAGS= -O3 -Munroll -Mnoframe -tp k8-64 -Mcache_align -Mnolre
FFLAGS_DEBUG= -g
RANLIB=echo
LDFLAGS=-R/opt/acml/pgi64/lib
COMP_LIBS=
#NETCDF_LIBS= -L$(HOME)/soft/netcdf/pgi64/lib -lnetcdf
#NETCDF_INTERFACE=libnetcdf_f90.a
#DEFS_CDF=-DCDF
#
MPI_INTERFACE=libmpi_f90.a
MPI_INCLUDE= /opt/mpich/include
MPI_LIBS= -L/opt/mpich/lib -lfmpich -lmpich*
**
*MATHLIBS= -L/opt/acml/pgi64/lib -lacml -I/opt/acml/pgi64/include
BLACS=/opt/acml/pgi64
BLACS_LIBS= -I$(BLACS)/include \
$(BLACS)/lib/libblacsF77init.a \
$(BLACS)/lib/libblacsCinit.a \
$(BLACS)/lib/libblacs.a
SCALAPACK_LIBS= /opt/acml/pgi64/lib/libscalapack.a
#
DEFS_MPI=-DMPI
#
LIBS= $(SCALAPACK_LIBS) $(BLACS_LIBS) $(MATHLIBS) \
$(MPI_LIBS) $(NETCDF_LIBS) \
-L/usr/lib/gcc/x86_64-redhat-linux/3.4.3
#
SYS=bsd
DEFS= $(DEFS_CDF) $(DEFS_MPI)
MATHLIBS= -L/opt/acml/pgi64/lib -lacml -I/opt/acml/pgi64/include
BLACS=/opt/acml/pgi64
BLACS_LIBS= -I$(BLACS)/include \
$(BLACS)/lib/libblacsF77init.a \
$(BLACS)/lib/libblacsCinit.a \
$(BLACS)/lib/libblacs.a
SCALAPACK_LIBS= /opt/acml/pgi64/lib/libscalapack.a
#
DEFS_MPI=-DMPI
#
LIBS= $(SCALAPACK_LIBS) $(BLACS_LIBS) $(MATHLIBS) \
$(MPI_LIBS) $(NETCDF_LIBS) \
-L/usr/lib/gcc/x86_64-redhat-linux/3.4.3
#
SYS=bsd
DEFS= $(DEFS_CDF) $(DEFS_MPI)
#
.F.o:
$(FC) -c $(FFLAGS) $(DEFS) $<
.f.o:
$(FC) -c $(FFLAGS) $<
.F90.o:
$(FC) -c $(FFLAGS) $(DEFS) $<
.f90.o:
$(FC) -c $(FFLAGS) $<
#
*
--
Hai-Ping Lan
Department of Electronics ,
Peking University , Bejing, 100871
