From: Alfredo Ferrari (alfredo.ferrari@cern.ch)
Date: Tue Feb 14 2006 - 15:26:08 CET
Hi Sebastien
let's try in order
a) FC3 has a very well known issue with gdb+g77, the stack get spoiled
and no useful info comes out. It is bug which was there around
RedHat 5.x/6.x and suddenly reappeared with FC3. I attach an object
file (setsig.o), which overrides the signal trapping routine
of g77 and ususally makes core dumps again readable on FC3.
Of course you have to link FLUKA with setsig.o in (like whichever
else user routine) and run again
b) In the directory left after the crash (fluka_xxxx) there should
be a ranxxx corresponding to the next random number. It is the last
random successfully printed by FLUKA before the crash. If you start
the run from this random it should crash almost immediately (within
5 minutes of CPU time, often much sooner). If you succeed in doing so
you can
1) check with the debugger where it crashed (assuming you relinked
including setsig.o)
2) send us the input file and the random which triggers the crash
Ciao
Alfredo
On Tue, 14 Feb 2006, Sebastien WURTH wrote:
> Dear Fluka users and developpers,
>
> I'm trying to simulate a 50 MeV / 10 ľA electron beam on a U238-C2 target, so
> far so good.
> The resnuclei distribution seems to be quite as I expected.
> When I calculate the radioactive decay "offline" with usersuwev program, I
> get good results, still OK.
> But, when I try the same "online" this time, using RADDECAY, IRRPROFILE,
> DCYTIMES and DCYSCORE cards, I get a core dump before the end, but not at the
> very beginning. The run goes normally till it crashes.
> I would like to know what seems to be the problem.
> I figure it's the RADDECAY option, since everything goes OK without it, but I
> don't find where.
> I don't get any understandable answer when using gdb command applied to the
> core file.
> I don't think that my machine has a memory problem or something like that. I
> run on Redhat FC3.
> Is it a problem of time left for termination (defined in START card) ? I
> tried several termination times, with the same negative result.
>
> Any clue ?
> Thank you.
>
> Sebastien
>
> PS : Attach file is the used input file "target.inp"
>
-- +----------------------------------------------------------------------------+ | Alfredo Ferrari || Tel.: +41.22.767.6119 | | CERN-AB || Fax.: +41.22.767.7555 | | 1211 Geneva 23 || e-mail: Alfredo.Ferrari@cern.ch | | Switzerland || Alfredo.Ferrari@mi.infn.it | +----------------------------------------------------------------------------+
This archive was generated by hypermail 2.1.6 : Tue Feb 14 2006 - 15:55:24 CET