Re: FPE crash, PMUPEV: rather large balance ?!

Anna,
Thanks for the suggestion about precision on input. I tried it but it
does not fix the problem on my mchines. May be a difference in
machine architecture or compiler or something. The job failed again
on FPE after
4.76 hours.

The failed output is at

http://www.slac.stanford.edu/~arnold/fluka_32606/

   So we try again.

Thanks much for your good observations, and we'll get this fixed some way.

     Ray

On Fri, 26 Sep 2008, Anna Ferrari wrote:

>
>
> Hi Ray,
> your FPE crash has probably nothing to do with the warnings you get from
> PMUPEV: these warnings - which, as you wrote, report about particular
> kinematic situations - are not the symptom of an incoming fatal error of
> the type you got.
>
>
> I have run your input file changing this line:
>
> *..+....1....+....2....+....3....+....4....+....5....+....6....+....7....+....8
> PHOTONUC 1111. 1.0D-4 HYDROGEN @LASTMAT 1.0MUMUPAIR
>
> in:
>
> *..+....1....+....2....+....3....+....4....+....5....+....6....+....7....+....8
> PHOTONUC 1111.1.00000D-4 HYDROGEN @LASTMAT 1.0MUMUPAIR
>
> that is: I used for the double precision number corresponding to what(2)
> all the available 10 digits, to avoid uncorrect readings and then possible
> FPE crashes coming from here.
>
> This works: I reached 20 hours of CPU time without getting
> errors...
>
> cheers,
>
> Anna
>
>
>
>
>
>
>
> On Mon, 22 Sep 2008, Raymond G. Arnold wrote:
>
>> Dear FLUKA Dr.
>>
>> I'm having a mysterious problem that seeems to generate bad kinematics
and lots
>> of PMUPEV error messages (example below) from Peanut. I'm running 500 GeV
>> electrons in a simulation of a dump and shielding for development of
possible
>> Linear Collider Beam Dumps. Until recently I inched forward
sucessfully adding
>> features and physics process and all was working fine. Lots of errors
came out
>> but they didn't stop the program so I continued. I'm running the standard
>> executable $FLUPRO/flukahp with no user routines.
>>
>> A few days ago I noticed that some of the errors were complaining about
>> problems transporting Kaons, so I put in a higher energy tranport
cutoff on
>> Kaons and the message volume decreased, but did not go away.
>>
>> I'm studying energy depositon, neutron fluence, and (in a separate .inp)
>> activation decay radiation dose. The FLUKA model includes a large bug free
>> geometry, and lots of physics process (LOW-NEUT, PHOTONUC with and without
>> MUMUPAIR,...) with careful tuning of the EMFCUT to preserve the physics
but
>> increase efficiency, and some biasing to increase statistics in the low
flux
>> regions. The .inp was created with flair, run many times for short (<
1 hour
>> runs) and then slightly modified to change number of primaries and
tweek the
>> RANDOMIZe card for long batch running.
>>
>> Now I'm trying to run long jobs for statistics, and have hit a brick wall.
>> After 4 hours or more CPU time, FLUKA exits with a floating point
exception,
>> from somewhere, I don't know where. Or even if the FPE is realated to the
>> PMUPEV errors.
>>
>> I guessed PMUPEV might be associaed with photo-nuclear interactions by
muons.
>> So I commented out the MUPHOTON card, but no efect.
>>
>> I'm stuck. Can anybody tell me how to fix this, or what to chop out so
I can
>> get past the fatal FPE.
>>
>> The latest input and output from a failed run can be found at
>>
>> http://www.slac.stanford.edu/~arnold/fluka_13873/
>>
>>
>> Thanks for any hints...
>>
>> Ray Arnold
>>
>> P.S. - Hello to my friend Alberto, wherever you are. We miss you
here at
>> SLAC.
>>
>> ==============================================================================
>>
>>
>> NEXT SEEDS:319A96CE 27 0 0 0 0 181CD 3039 0
>> 0
>> PMUPEV: rather large balance ?!
>> PXRCL,PYRCL,PZRCL 0. 0. 0.0627817928
>> ERCL,EKRCL 0.940347661 0.00207535099
>> PMUPEV: rather large balance ?!
>> PXRCL,PYRCL,PZRCL 0. 0. 0.140975004
>> ERCL,EKRCL 28.8444688 0.000241636737
>> PMUPEV: rather large balance ?!
>> PXRCL,PYRCL,PZRCL 0. 0. 0.126833818
>> ERCL,EKRCL 22.3361239 0.000320194522
>> *** Peanut:Ptres,Ibres,Icres 0. 16 7
>> 1915 8085 8085 8.7497104E+00
>> 1.0000000E+30 23109966
>> NEXT SEEDS:34A07743 27 0 0 0 0 181CD
3039 0
>> 0
>> PMUPEV: rather large balance ?!
>> PXRCL,PYRCL,PZRCL 0. 0. 0.0939789178
>> ERCL,EKRCL 0.942943942 0.00467163214
>> Eventv: ekin+am < pla,ij,igreyt 6.88954317 6.8996321 14 2
>> 1916 8084 8084 8.7491050E+00
>> 1.0000000E+30 23120448
>> NEXT SEEDS:35B78595 27 0 0 0 0 181CD
3039 0
>> 0
>> PMUPEV: rather large balance ?!
>> PXRCL,PYRCL,PZRCL 0. 0. 0.0564107626
>> ERCL,EKRCL 0.939966536 0.0016942256
>> Eventv: ekin+am < pla,ij,igreyt 6.90732917 6.91668049 14 5
>> 1917 8083 8083 8.7483434E+00
>> 1.0000000E+30 23130154
>> NEXT SEEDS:36C556D3 27 0 0 0 0 181CD
3039 0
>> 0
>> =============================================================================
>>
>> Raymond G. Arnold arnold_at_slac.stanford.edu
>> M.S. 42 SLAC voice (650) 926-2755
>> P.O. Box 4349 fax (650) 926-2407
>> Stanford, CA 94309
>>
>
>
>

Raymond G. Arnold arnold_at_slac.stanford.edu
M.S. 42 SLAC voice (650) 926-2755
P.O. Box 4349 fax (650) 926-2407
Stanford, CA 94309
Received on Sat Sep 27 2008 - 23:46:09 CEST