Re: Different events with identical kinematics

Hi Joseph

Do I understand correctly that you run -N0 -M5 and get two out of them
with identical output even though each of the five runs starts with a
different random number seed (the one from the previous cycle)?

Do you start always from the same entry in your source particle file? If
so (and especially in case you do not go through the source sample
several times during the same cycle) it would be advisable to start
randomly at an entry in that file. However, this is only a general comment
and may not explain your observation.

If you would like us to check your problem please send us input, source
files and the Version.tag of your FLUKA installation.

Best regards
Stefan

On Sat, 14 Jun 2008, Joseph Comfort wrote:

> I am doing a beamline simulation, running particles (e.g., neutrons)
> through a set of collimators. At various points along the way, I produce
> a file of particle kinematics (X,Y,Z, momentum components, age, weight,
> and Fluka case number). I am mainly interested in what gets through to
> the end. (The particle source is actually a set of similar Fluka output
> files from earlier stages of the simulation, with some selection criteria
> imposed.)
>
> During some post-processing, I noticed that many groups of events had
> identical output kinematic data, differening only in the Fluka case
> number. More specifically, I did 5 reclycles of the input data. For one
> set of final position/momentum data, I had 2 identical outputs with
> different case numbers in one of the cycles; the entire output including
> the case number was duplicated into a separate cycle; some other cycles
> had the same output data, with still different case numbers.
>
> In looking at the history of such events, I could see that they had gone
> through the collimator with identical kinematic values at each of the
> inspection points. But the events prior to that were quite different.
>
> I would not expect such similarities if the random number sequence is
> anywhere near random. I start the 5 cycles with the default random
> number, and the values between cycles are properly written. It seems very
> peculiar to me. One would normally notice such duplication. I found it
> because my post-processing imposed some very tight constraints on the
> events (i.e., did the particles scatter from something upstream).
>
> The issue is important because I count the number of particles that reach
> the end of the line to determine production rates. I'm not sure how to
> interpret the duplications.
>
> Thank you
> Joe Comfort
>
>

-- 
___________________________________
Stefan Roesler
CERN, SC/RP
CH-1211 Geneva 23
Switzerland
Phone:  +41-22-7679891
Fax:    +41-22-7669639
E-mail: Stefan.Roesler@cern.ch