Date: Thu, 23 Apr 2015 22:19:00 +0200
Dear Yuni,
thank you for your very detailed e-mail.
You don't need flair to merge your data, you can script it using the
embedded tool or you can modify it easily in order to take advantage of
your cluster.
If the nodes are the same you don't need to recompile on each machine.
Using OpenPBS the submission script might look like: (credits to Vasilis
Vlachoudis)
=====
#!/bin/bash
MY_JOBFILE=${MY_NAME}.job
MY_NAME=`echo ${MY_NAME} | cut -c 1-15`
cat > ${MY_JOBFILE} << EOF
#!/bin/bash
#PBS -q normal
#PBS -N ${!#}
export FLUKA=$FLUKA
cd $PWD
$*
EOF
echo "Submitting job ${MY_JOBFILE}"
chmod +x ${MY_JOBFILE}
qsub ${MY_JOBFILE}
=====
Best, V.
On Thu, Apr 23, 2015 at 5:07 AM, 汪金龙 <wangjinlong_at_nssc.ac.cn> wrote:
> Dear Vittorio
>
> Thank you very much. you are right, submission script do almost the copy
> work. But it seems not that simple when useroutine files *.f have to be
> compiled on every node and link to fluka run. The attached simple script
> are what I used before, but always get errors during a long time simulation
> when large pramary was set (up to 1E8). Thus I just simplified my GEO model
> and no magenet, no lattice etc.
>
> And now in the following picture is what I'm trying to do. Do you have
> any suggestion for this tast?
>
> Thank you and best wishes
> Yuni
>
>
> -----原始邮件-----
> *发件人:* "Vittorio Boccone" <dr.vittorio.boccone_at_ieee.org>
> *发送时间:* 2015年4月23日 星期四
> *收件人:* "汪金龙" <wangjinlong_at_nssc.ac.cn>
> *抄送:* "fluka-discuss_at_fluka.org" <fluka-discuss_at_fluka.org>
> *主题:* Re: Re: [fluka-discuss]: How to merge data with weight by flair?
>
>
> Well you can start parallel simulation with a simple script (many cluster
> or batch system provide users with examples.)
> In the flair preferences you can customize the script used to submit the
> fluka jobs (i.e. to a cluster).
>
> Typically the submission script takes care of copying the artifact of the
> simulation to an aread accessible to the user once the process is
> terminated. At this point you would need to have a second job processing
> the data.
>
> Best,
> Vittorio
>
> On Wed, Apr 22, 2015 at 4:45 AM, 汪金龙 <wangjinlong_at_nssc.ac.cn> wrote:
>
>> Dear Vittorio：
>> Thank you very much, it is good to use *usbsuw.f *on local computer,
>> however, in order to decrease simulation time I use parallel with several
>> cluster nodes, unfortunatly, I still not solve another problem that start a
>> parallel fluka simulation with useroutine.
>> How about SimpleGEO, is there any plugin can deal **.bnn data?
>>
>> Best regards
>> Yuni
>>
>>
>>
>> -----原始邮件-----
>> *发件人:* "Vittorio Boccone" <dr.vittorio.boccone_at_ieee.org>
>> *发送时间:* 2015年4月22日 星期三
>> *收件人:* "汪金龙" <wangjinlong_at_nssc.ac.cn>
>> *抄送:* "fluka-discuss_at_fluka.org" <fluka-discuss_at_fluka.org>
>> *主题:* Re: [fluka-discuss]: How to merge data with weight by flair?
>>
>>
>> Dear Yuni,
>> I'm not aware of any tool distributed with FLUKA/flair which allow you
>> to do what request.
>> I would personally do it via an external script (once the simulations of
>> energy are merged).
>>
>> You have the sources of the tools *usbsuw.f* and the specification about
>> how the *USRBIN* are read/written is available both in the source code
>> and in the manual.
>>
>> Best,
>> Vittorio
>>
>> On Tue, Apr 21, 2015 at 5:43 AM, 汪金龙 <wangjinlong_at_nssc.ac.cn> wrote:
>>
>>> Dear Vasilis and fluka experts:
>>>
>>> I use USRBIN to score dose-eq, the beam has several energies for
>>> exmaple E1 E2, and I get some **fort.24 files for each energy after several
>>> simulaitons. My question is how to merge these fort files into one
>>> **24.bnn file with weight according to different energy?
>>>
>>> Thanks
>>> yuni
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>>
>>
>>
>
>
>
>
>
__________________________________________________________________________
You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id=acc_info
(image/png attachment: ______1.png)