As an additional note to the problem I posted yesterday, I run MM5 on a 32-bit machine, using the same testing data, and got run through.
The info about the 32-bit machine:
Linux: 2.4.29
PGI: Linux86/6.1
CPU:AMD Opteron
The compiling flags are:
RUNTIME_SYSTEM = “linux”
FC = pgf90
FCFLAGS= -I$(LIBINCLUDE) -g -Mcray=pointer -tp p6 -pc 32 -Mlfs -Mnoframe -bytes
wapio
#FCFLAGS = -I$(LIBINCLUDE) -O2 -Mcray=pointer -tp p6 -pc 32 -Mlfs -Mnoframe -by
teswapio
##FCFLAGS = -I$(LIBINCLUDE) -O2 -Mcray=pointer -tp p6 -pc 32 -Mnoframe -byteswa
pio -mp
##-Mnosgimp
CPP = /lib/cpp
CFLAGS = -O
CPPFLAGS = -I$(LIBINCLUDE)
LDOPTIONS = -L/usr/pgi/linux86/6.1/liblf -g -Mcray=pointer -tp p6 -pc 32 -Mnofr
ame -byteswapio
#LDOPTIONS = -L/usr/pgi/linux86/6.1/liblf -O2 -Mcray=pointer -tp p6 -pc 32 -Mno
frame -byteswapio
##LDOPTIONS = -O2 -Mcray=pointer -tp p6 -pc 32 -Mnoframe -byteswapio -mp
LOCAL_LIBRARIES =
MAKE = make -i -r
It is wired that I could run MM5 on the 32-bit machine, but got “seg fault” on a 64-bit machine, with the same input file. I used “pgdbg” and found the seg fault occurred at line 622 of the param.F file, which is simply this:
DX=BHR(9,1)
Thank you for any help.
Jane