[Octopus-users] opteron pgf90

[Giacomo Mulas]

On Thu, 24 Nov 2005, Francesco Sottile wrote:

> Dear all,
> has anyone of you tried to compile octopus on an opteron using the portland 
> compiler ?
> I have managed to compile but then it's immediately a segmentation fault,
> it doesn't matter the compilation flags (even none).
> The version I'm using is the devel 1.99 (but downloaded few months ago).
> In the manual of 1.4 there are some suggested flags for this compiler on 
> opteron, that I interpret as is this version (1.4) should work on this 
> platform. has anyone experienced these problems with the 1.99 version?

Hi Francesco. I did compile (and successfully run) Octopus 1.4 on an Opteron
cluster using the portland compiler version 5.2-4. However, I did have to
compile locally all libraries using the same compiler (otherwise it would
crash unexpectedly), and I had to specify at configure time that I wanted to
use the portland C compiler as well, since I had problems linking objects
compiled by pgf90 and objects compiled by gcc. The only libraries I did not
recompile were the lapack/blas libraries, since I used the optimised ones
provided by AMD for opterons (acml), which come in many flavours including
one for the nag compiler. I did not try with the CVS version of Octopus,
though. I hope this helps you...

Bye
Giacomo

-- 
_________________________________________________________________

Giacomo Mulas <gmulas at ca.astro.it>
_________________________________________________________________

OSSERVATORIO ASTRONOMICO DI CAGLIARI
Str. 54, Loc. Poggio dei Pini * 09012 Capoterra (CA)

Tel. (OAC): +39 070 71180 248     Fax : +39 070 71180 222
Tel. (UNICA): +39 070 675 4916
_________________________________________________________________

"When the storms are raging around you, stay right where you are"
                          (Freddy Mercury)
_________________________________________________________________

-- 
Il messaggio e' stato analizzato alla ricerca di virus o
contenuti pericolosi da MailScanner, ed e'
risultato non infetto.