Re: [fluka-discuss]: Fluka-Matlab link

From: Vittorio Boccone <dr.vittorio.boccone_at_ieee.org>
Date: Tue, 26 Apr 2016 08:30:37 +0200

> On Apr 25, 2016, at 10:44 PM, me_at_marychin.org wrote:
>
> Hi Marlon, Vittorio and Masoomeh,
>
> Isn't variation in material density just an instance of parameter sweeping?
No because density variation means and expansion/shrink of bodies/regions as total matter should be conserved.
In my case this was a voxel structure recalculation. Eventually complex movable targets are also not just sweeping a number or two.


> MATLAB is not just any other numerical code. Its muscle is in interactive data
> examination (incl plotting) while having all the variables in memory. The
> overhead for batch work and data processing is high -- just do a tic-toc and see
> for yourself. Even for things as simple as reading in data from a file, it could
> be like touching one's nose via the opposite side of the head. Compare the
> tic-toc with Perl.
Don’t get me wrong I’m not a fan of Matlab, at all. But sometimes you can’t choose the tool as adapting, starting and analyzing LD-DYNA (my experience) simulation is way easier in Matlab.

>
> Technically there is no problem starting FLUKA from MATLAB. Starting is far
> easier than stopping! Issue the less-graceful ctrl-c and the more-graceful
> ctrl-z + bg? That isn't going to work, right? A shell script would be far
> neater.
Locking/killing mechanism should be implemented in the bash scripts as much as possible.

Having the main loop in bash has clearly advantages, although matlab can plot you live the status of simulation if you need for example to monitor the quality of you results quickly. Every tools has its own range of application. The scientist has to decide what’s better for him/her.


> :) mary
>
>> On 26 April 2016 at 03:07 Vittorio Boccone <dr.vittorio.boccone_at_ieee.org>
>> wrote:
>>
>>
>> Hi Mary and Masoomeh,
>> there are plenty of good reasons why one could lightly couple FLUKA with any
>> other numerical code, as for example to compute variation in the material
>> density in particular region. Ofter Finite-Elements codes for
>> thermo-mechanical studies have a natural Matlab interface.
>>
>> The important part is to define the correct scripts and locking mechanism to
>> handle the simulation loop.
>>
>> In our case we performed a soft-coupling FLUKA of with LSDYNA using Matlab in
>> Linux.
>>
>> Matlab can easily run bash scripts.
>> The procedure in not technically difficult but requires a bit of scripting
>> karma.
>>
>> V.
>>
>>> On Apr 25, 2016, at 3:02 PM, me_at_marychin.org wrote:
>>>
>>> Dear Masoomeh,
>>>
>>> **Using MATLAB to post-process FLUKA output**
>>> there is absolutely no problem. The trouble is probably not worth it, unless
>>> you
>>> have exhausted FLAIR's potential, in which case please follow through this
>>> thread:
>>> http://www.fluka.org/web_archive/earchive/new-fluka-discuss/7292.html
>>>
>>> **Running FLUKA via MATLAB**
>>> I really can't think of any reason for wanting this. Why would one ever wish
>>> to
>>> do that? Are you thinking of some complex parameter-sweeping studies? Even
>>> that
>>> can be accomplished far better using simple shell scripts.
>>>
>>> :) mary
>>>
>>>> On 24 April 2016 at 15:25 Masoomeh Yarmohammadi <myarmohammadi_at_ipm.ir>
>>>> wrote:
>>>>
>>>>
>>>> Dear All,
>>>>
>>>> Could you please let me how I can make a link between Fluka and Matlab then
>>>> run Fluka via Matlab.
>>>>
>>>> Many thanks in advance,
>>>>
>>>> Masoomeh Yarmohammadi Satri
>>>> Post-Doctoral Research Fellow, School of Particles and Accelerator,
>>>> Institute
>>>> for Research in Fundamental Sciences (IPM)
>>>> E-mail : myarmohammadi_at_ipm.ir, myarmoha_at_cern.ch
>>>>
>>>> __________________________________________________________________________
>>>> You can manage unsubscription from this mailing list at
>>>> https://www.fluka.org/fluka.php?id=acc_info
>>>>
>>>
>>> __________________________________________________________________________
>>> You can manage unsubscription from this mailing list at
>>> https://www.fluka.org/fluka.php?id=acc_info
>>>
>>

__________________________________________________________________________
You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id¬c_info
Received on Tue Apr 26 2016 - 10:08:52 CEST

This archive was generated by hypermail 2.3.0 : Tue Apr 26 2016 - 10:08:53 CEST