Hi all,
I hacked out a port of Cernlib for AMD64, and it should now be available in
the AMD64 archive (thanks to Andreas Jochens for building it). If anyone
who has an Opteron machine could do some testing on it for me, please let
me know the results. I'm particularly interested in whether the FORTRAN
interpreter built into PAW works correctly.
I would also welcome test reports from any of these other architectures
that I don't have access to myself (not being a DD), especially the 64-bit
ones:
ia64, alpha, hppa, mips, mipsel, arm, s390, powerpc (G5 only)
Here is a minimal test you can try even if not familiar with PAW or
FORTRAN. Note that in pawtest.f each line begins with 6 spaces except for
"10 CONTINUE" which begins with 3 spaces. (It's a recursive function to
generate the Nth triangular number, for those who are wondering.)
$ sudo apt-get install paw
$ cat << EOF > pawtest.f
SUBROUTINE PAWTEST(N)
SUM=0.
DO 10 I=1,N
SUM=SUM+I
10 CONTINUE
PRINT *,SUM
END
EOF
$ paw
[PAW banner output snipped]
Workstation type (?=HELP) <CR>=1 : [hit ENTER once]
[More PAW output snipped]
[A graphics window will open if you have a $DISPLAY;
it can be minimized and ignored.]
PAW > call pawtest.f(5)
15.00000
PAW > call pawtest.f77(5)
15.
PAW > quit
$
The important requirement is that both PAW commands above should output the
real number 15.0 (with some arbitrary number of zeros after the decimal
point). Please email results directly to me so as not to clutter the
mailing lists. Be sure to let me know your architecture and sub-arch.
On 64-bit sub-arches whose .debs are compiled for 32-bit (sparc64, s390x,
powerpc G5), you may first need to "touch /etc/disable_64_gcc" as root for
the test to succeed. Please also let me know how the existence and
non-existence of that file affect the test results.
thanks and regards,
--
Kevin B. McCarty <[EMAIL PROTECTED]> Physics Department
WWW: http://www.princeton.edu/~kmccarty/ Princeton University
GPG: public key ID 4F83C751 Princeton, NJ 08544
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
