Building Molpro for Alpha Linux

[Peter Knowles]

The idea was simply to run that one-line program and be convinced that
it delivers "8" rather than "4". If so, everything should be OK, since
then the interface between fortran and C inside Molpro should be
happy.
Peter
At Tue, 01 Oct 2002 12:10:22 +0200,
David Smith wrote:
> 
> Peter Knowles wrote:
> > 
> > I have the impression that you can just forget about the -i8 flag. The
> > key thing is that type 'long' in C is actually compiled as 8
> > bytes. You should perhaps check this:
> > int main(){ printf("%d\n",(int)sizeof(long)); }
> > 
> > Peter
> 
> Hi, 
> 
> I'm still trying to get things to work, but at this stage trying is the
> operative word. I'm not sure I quite followed the "long" integer stuff.
> Certainly some of the source contains this kind of definition but my c
> is not good enough to really understand what is going on.
> 
> Anyway, I tried with CC="gcc" along with 
> FC="fort -i8 -assume no2underscores"
> F90="fort -i8 -assume no2underscores"
> 
> and the compilation proceeded with no errors.  However, if I try make
> quicktest or something similar I get: 
> 
> "Received signal 11 Segmentation violation" every time I try to execute
> the program.
> 
> At first I thought it might be a problem with malloc as I had read some
> stuff on the list about that not so long ago.
> 
> I didn't have the very latest version so I downloaded the current one
> which has the linuxconf and linux_gmalloc patches installed.
> Unfortunately, it didn't help. Still the same error when trying to run.
> 
> I looked in linux.conf at the following logic:
> 
>   elif [ `uname -m | grep -c alpha` -eq 1 ]; then # If i am an alpha
>     if test "x$3" = "xunix-linux" ; then
>       if [ x"$compflev" != x"77" -a -r /usr/bin/fort ]; then
>         gmalloc=1
>         if [ x$2 = xunix-i8 ]; then
>         fortran="fort -i8 -assume no2underscores"
>         cc="cc -i8 -DI64"
>         gmalloc=0
>         else
>         fortran="fort -assume no2underscores"
>         cc="cc"
>         fi
> 
> and thought I would try with the         
> 
> fortran="fort -assume no2underscores"
> cc="cc"
> 
> combination directly in the CONFIG file. This gave me a complier error,
> please contact the vendor...
> 
> I thought I would also see what happened when trying -i4 as an option to
> configure but this doesn't work either as there is no license object
> available for alpha with i4.
> 
> Any suggestions would be greatly appreciated.
> 
> Thanks.
> 
> ---------------------------------------
> Dr. David Smith
> Department of Chemistry
> Ludwig Maximilians University
> Butenandt-Str. 5-13, D-81377 Munich
> Germany
> Tel.: +49 (0)89 2180 7740
> Fax.: +49 (0)89 2180 7738
> e-mail: David.Smith at cup.uni-muenchen.de
> ---------------------------------------

--
Prof. Peter J. Knowles              
Email P.J.Knowles at bham.ac.uk  Phone +44-121-414-7472  Fax +44-121-414-7471
School of Chemical Sciences, University of Birmingham, Birmingham, B15 2TT, UK
WWW http://www.tc.bham.ac.uk/~peterk/