[CPMD-list] Itanium SGI

[Eliana K Asciutto]

Dear all,

I've compiled succesfully CPMD 3.9 in a SGI Itanium cluster.
The problem is that it is very slow (compared with an Intel Xeon 
cluster). Perhaps I'm missing some compilation's option  in order to 
speed it up.
Here is my Makefile:



#--------------- Default Configuration for SGI_LINUX_IA64_INTEL-MPI 
---------------
SRC  = .
DEST = .
BIN  = .
#QMMM_FLAGS = -D__QMECHCOUPL
#QMMM_LIBS  = -L. -lmm
FFLAGS = -fpp -D__alpha -DPOINTER8 -DLAPACK -DFFT_DEFAULT -DALPHALINUX \
        -DPARALLEL=parallel -DMP_LIBRARY=__MPI -DLINUX_IA64_INTEL \
        $(shell date +'-D__DATE__=\""%b %e %Y"\" -D__TIME__=\"%T\"')
LFLAGS = -L/opt/intel/mkl70/lib/64 -lmkl_itp -lmkl_lapack -lmkl_itp \
             -lPEPCF90 -lm -lpthread -lmpi $(QMMM_LIBS)
CFLAGS = -DPGSIZE="getpagesize()" -D__alpha -DPOINTER8 -DLAPACK 
-DFFT_DEFAULT -DALPHALINUX \
        -DPARALLEL=parallel -DMP_LIBRARY=__MPI -DLINUX_IA64_INTEL
CPP = /bin/cp --remove-destination
CPPFLAGS =
NOOPT_FLAG =
CC = gcc -O2 -c -Wall
FC = efc -O2 -w -w90 -w95 -cm -c
LD = efc -O2
AR = /usr/bin/ar -r

 
Any suggestions?
Thanks,

Eliana


-- 
Eliana K Asciutto
North Carolina State University
Deparment of Physics
127 Stinson Drive, 308 Cox Hall
Raleigh, NC, 27695-8202
Tel:(919)513-0317