Subject: CCM3 on Deck Alpha
From: Maisa Rojas (maisa@dgf.uchile.cl)
Date: Thu Nov 08 2001 - 13:15:45 MST
Dear Chien and anyone else on the ccm3 user list:
thanks to your quick answer to my previous mail about running CCM3 on a
DEC Alpha station. I followed your suggestions and downloded gnumake.
CCM3 seems to compile well now, but while running I get a segmentation
fault error early in the run. I have the impresion that it has nothing to
do with segmentation fault, but is rather a compilation problem, although
the error arrises when trying to equal two arrays defiened as real*16 in
the subroutine that calculates the sine of the latitudes (gauaw.F and
bsslzr.F) (I have stacksize and memoryuse set unlimit).
I am compiling the model with the following options:
#------------------------------------------------------------------------
# DIGITAL OSF1
#------------------------------------------------------------------------
CPP := /lib/cpp
FC := f90
CFLAGS := $(cpp_path) -64 -cpp
FFLAGS := $(cpp_path) -64 -r8 -i4 -O3 -c -cpp
SPEC_FFLAGS := $(cpp_path) -64 -r8 -i4 -O2 -c -cpp
LDFLAGS := -L$(LIB_NETCDF) -lnetcdf
I took out everything I believed had to do with MPI, as our Dec/Alpha
station doesn't have MPI.
(details of the machine are:
OSF1 V4.0 878 alpha
and the fortran compiler is:
DIGITAL Fortran 90 V5.1-594
to anyone with some insight on this i'm would be very greatfull to hear
from you!
thanks,
Maisa
-------------------------------------------------
Maisa Rojas Corradi
Departamento de Geofisica
Universidad de Chile
56-2-678 4302
This archive was generated by hypermail 2b27 : Tue Jan 22 2002 - 11:16:21 MST