I am trying to do some loops of parameters in CAMB and run it with MPI on a INTEL-8 like supercomputer. In Makefile,I set
F90C = efc
FFLAGS = -openmp -ip -O2 -Vaxlib -W0 -WB -fpp2 and run with 256 CPUs.
However the result is not correct in that CAMB has repeated the computation for 256 times with the same parameter. On the other hand , I have tested my code on a COMPAQ machine with the Makefile:
F90C = f90
FFLAGS = -omp -O4 -arch host -math_library fast -tune host -fpe1
and found the result correct. Any suggestions on that ? Thanks in advance.
CAMB: run with MPI
1 post • Page 1 of 1