[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: 19990324: Alpha/Linux problem: Floating point exception
- Subject: Re: 19990324: Alpha/Linux problem: Floating point exception
- Date: Wed, 24 Mar 1999 12:30:05 -0700
Root,
> To: address@hidden
> From: root <address@hidden>
> Subject: porting trouble
> Organization: .
> Keywords: 199903241611.JAA18524
In the above message, you wrote:
> I am having trouble getting NetCDF running on my system. It is
> a alpha-based PC running RedHat Linux 5.2. The package compiles
> OK and then fails on the "make test". I have a commercial version
> of fortran called NDP fortran. It has a lot of the VMS extensions
> and these important compile options
> -fastlib (affects lib linking)
> -O (same as -O1, must supply an optimization switch)
> -64 (for 64 bit mode; faster because pointer arithmetic is the natural
> word size of the machine)
> -X954 (suppresses trailing _ in common names)
> -X955 (suppresses trailing _ in procedure names)
> -X956 ( suppreses trailing underscore in main procedure name)
>
> Anyway the following is the non-logfile info:
> root@/usr/local/netcdf-3.4/src]# man which
> [root@/usr/local/netcdf-3.4/src]# which f77
> /usr/bin/f77
> [root@/usr/local/netcdf-3.4/src]# which cc
> /usr/bin/cc
> [root@/usr/local/netcdf-3.4/src]# which gcc
> /usr/bin/gcc
> [root@/usr/local/netcdf-3.4/src]# which fort77
> /usr/bin/fort77
> [root@/usr/local/netcdf-3.4/src]# which c++
> /usr/bin/c++
> [root@/usr/local/netcdf-3.4/src]# uname -a
> Linux ARMPL3.TOR.EC.GC.CA 2.0.35 #8 Mon Mar 22 15:17:29 EST 1999 alpha unknown
> [root@/usr/local/netcdf-3.4/src]# cat VERSION
> 3.4
>
> For some reason the computer doesn't know anything about the processor
> (the chip is a 21164 alpha and the implemetation/motherboard is LX164)
> so it writes unknown for that in the uname command.
>
> Please let me know if this test failure is important and if I can make
> the fortran compiler use 64 bit mode.
I don't know anything about the NDP Fortran compiler, so I don't know if
you can make it use 64 bit mode.
>
> - mark
>
> PS the log files from configure, make, & make test follow!
>
> loading cache ./config.cache
The above line indicates that you didn't do a "make distclean" or remove
the file "config.cache" before executing the configure script. Thus,
some of the following results result from a previous execution of the
configure script.
...
> Making `test' in directory /usr/local/netcdf-3.4/src/nc_test
>
> make[2]: Entering directory `/usr/local/netcdf-3.4/src/nc_test'
> cc -o nc_test -O nc_test.o error.o test_get.o test_put.o test_read.o
> test_write.o util.o -L../libsrc -lnetcdf -lm
> ./nc_test -c
> make[2]: *** [test.nc] Floating point exception (core dumped)
> make[2]: *** Deleting file `test.nc'
> make[2]: Leaving directory `/usr/local/netcdf-3.4/src/nc_test'
> make[1]: *** [subdir_target] Error 1
> make[1]: Leaving directory `/usr/local/netcdf-3.4/src'
> make: *** [nc_test/test] Error 2
We've seen this problem before. Please see the following URL
<http://www.unidata.ucar.edu/glimpse/netcdf/2843>
--------
Steve Emmerson <http://www.unidata.ucar.edu>