EGEE Budapest Contributions

Demo (medical physics)

These instructions have been provided from /afs/

Do: setenv DIANE_USER_WORKSPACE /afs/ before sourcing 1.6.0/DIANE/etc/environment.csh. Use Ganga version 4.4.0.

Diane and Ganga are setup as follows:

source /afs/

setenv PATH /afs/$PATH


(For more details see the sections below!!!)

1. Setup

 setenv ORBgiopMaxMsgSize 35000000
 setenv ORBdumpConfiguration 1
 cd /afs/
 eval ` -e csh`
 cd /afs/
 eval ` -e csh`

2. Simulation start

 cd /afs/
 env GANGA_CONFIG_PATH=$PWD/gangarc-submit ./diane.startjob2 -j macros/macroIridium50mio20k.job -w30@lcg --ganga -ORBendPoint giop:tcp::4069 --wms=$PWD/celists/cern_ce_list

3. Plot generation

 cd /afs/
 sh <DIANE output directory>
(e.g. "sh 3" where "3" is the subdirectory of above directory)

4. Plot presentation (once the plot histo10.jpg was created in the DIANE output directory) kview histo10.jpg


I) Setup of DIANE environment for the G4Brachy application:

0) Let your diane.workspace-symbolic link point to /afs/

1) Please type:

setenv ORBgiopMaxMsgSize 35000000
setenv ORBdumpConfiguration 1

2) Please type:

cd /afs/
eval ` -e csh`

In case of success you will see: Setting up /afs/ completed setup for /afs/

3) Please type:

cd /afs/
eval ` -e csh`

In case of success you will see: setting up /afs/ completed setup for/afs/

II) Simulation run on the Grid:

Please start the simulation runs from the following directory: /afs/

by typing:

env GANGA_CONFIG_PATH=$PWD/gangarc-submit ./diane.startjob2 -j macros/macroIridium50mio20k.job -w30@lcg --ganga -ORBendPoint giop:tcp::4069 --wms=$PWD/celists/cern_ce_list

NOTE: This example (with this macro and number of workers) was working well for updating the merged histograms with PAW! This macro runs for about 45 min. For longer runs take macros/macroIridium500mio20k.job (ten times as long), but don't change the number of workers. However, you need to change the z-Range of the plots (for instructions see below, i.e. end of section III)

Site selection:

Site lists are specified in the directory "celists" of above submission directory. You can create your own site list used for the simulations. I experienced the best response from,, Please note: Not all sites specified in the lists in the directory "celists" are in good conditions for the Geant4 simulations. But above ones are the safest option.

Job file:

The job files can be found in the directory "macros" of above submission directory. E.g. the file macroIridium20mio20k.job contains 20 million events grouped into tasks of 20000 (This task size I think is optimal for update of the plots, so please keep it or slightly increase it: Otherwise PAW will not be able to merge if too many ouputs arrive due to smaller tasks). For the above script the simulations runs about 15 - 20 min. You can e.g. use the job file macroIridium500mio20k.job, which should run a few hours. If you want to adapt the event number by your own through changing the lines: 'eventNumber' : 500000, 'eventsPerWorker' : 20000, in the macro files (Please create new ones instead of editing existing ones).

Number of workers:

I would recommend not to use more than 30 workers, otherwise PAW is not capable of merging the output files (for the given task size of 20000 events).

III) Simulation output and graphics generation/view:

The output is stored in the subdirectories of /afs/

The explicit name of the output directory is given by the run number of DIANE (e.g. directory /5).

The outputs shipped backed from the different worker nodes are named taskout*.aida, where * is the task number.

The scripts responsible for the graphics generation are 1) 2) cumulativehistoplot.kumac (is called by script 1) ) These files are found in /afs/ (i.e in the job output base directory). The scripts must stay in this directory to work well (since paw does not except the absolute path of the input script due to a too long path name, thus I had to specify a relative path).

Once DIANE is running and has created the output directory, you can invoke the graphics creation file by: sh (where for the current output directory must be specified). e.g. sh 185

The scripts performs the following: 1) It periodically merges the output files into mergeddata.aida and removes the output files (Once started the script stays in an endless loop). 2) It creates an eps plot and converts it into a jpg (filename histo10.jpg, stored in output directory) which always overwrites the previous jpg file (and thus updates the histogram plot).

NOTE: In the startup phase of the simulation, no outputs or plots are available (it will take some time until plots are generated).

NOTE: Since I cannot estimate the fixed height of histogram plot (i.e. z-axis) for a long term simulation (with lots of outputs), you might want to change the following line in the script cumulativehistoplot.kumac: Just modify the last argument: H/SET/MAX [2] 1500 (e.g. here 1500 specifies the fixed maximum z in the plots). My experience is that the height is 1700 for 20 mio events.

To view the plots you can use kview (Just type "kview histo10.jpg" once an output file exists). Unfortunately kview does sometimes not automatically update the file.


In case you lose the scripts for plot generation, I they are attached here:


check () {

  if [ $1 -ne 1 ] ; then
    echo "Wrong number of arguments specified"
    echo "Usage: $0 <DIANE-outp-dir> "
    exit 1


createplot () {

  pushd $1
  if [ $? != 0 ] ; then
     echo  "Error: Cannot change to directory $1"
     exit 1

  ls taskout* > /dev/null 2>&1
  if [ $? = 0 ] ; then
     /afs/ -b ../cumulativehistoplot.kumac > /dev/null

  if [ -e  $ ] ; then
     ps2epsi $
     convert -quality 200x200 -density 90 $FILE.epsi $FILE.jpg


periodic () {

  while true
  do sleep 2
     createplot $1


check $#
periodic $1
exit 0


* * * * * * * * * * * * * * * * * * * * *
*   PAW script
*   Purpose: Adding all common histograms
*            of all .aida files located
*            in the current directory
* * * * * * * * * * * * * * * * * * * * *

SHELL mv aux.aida mergeddata.aida
EXEC CREATEPS mergeddata.aida 10

SHELL 'rm -f aux.kumac aux2.kumac listofaidafiles'
SHELL 'ls  *.aida > listofaidafiles'
SHELL 'cat listofaidafiles | sed -e ''s/[^ ]*/H\/FILE 0 &; HRIN 0 ! 99999; CLOSE 0/'' > aux.kumac'
SHELL 'sed ''1,1s/! 99999//1'' aux.kumac > aux2.kumac'
SHELL 'cat listofaidafiles | grep taskout | sed -e ''s/[^ ]*/rm &/'' >'
EXEC aux2.kumac
MESSAGE NOTE: Removing old output files
SHELL '/bin/bash'
SHELL 'rm -f aux.kumac aux2.kumac listofaidafiles'

MESSAGE NOTE: Writing histograms to file [1]
HISTOGRAM/FILE 0 aux.aida ! N

H/FILE 0 [1]
HRIN [2]
H/SET/MAX [2] 15000
FOR/FILE 44 histo[2].ps
METAFILE 44 -111
H/2D/L [2](1:40,1:40) ! ! 2
MESSAGE Plot to be found in file histo[2].ps


This list is a cut and paste from Massimo's mail. Start updating the item by adding the relevant documents to this page. Finshed items probably should be changed t0 bold.

  • PR part
    • Logos
    • Tutorial
    • H5N1 is another demo
    • multigrid (OSG and NDGF)
    • # of certificates in EGEE (total, 4 experiments)
    • median (Ian)

  • Ganga
    • poster for Manchester

  • QCD
    • Andrew stuff
    • #files vs time (log files)

  • ITU poster (conference)

  • DEMO
    • Tony?
    • Garfield
    • Tutorial?

  • Network requirements
    • ssh to cern
    • ports? inbound + 4001-4005

-- AndrewMaier - 19 Sep 2007

Topic attachments
I Attachment History Action Size Date Who Comment
PowerPointppt Ganga_CHEP07_-_final.ppt r1 manage 6312.5 K 2007-09-21 - 17:25 AndrewMaier Final version of CHEP talk in PPT format. (Includes the simplfied tables for the plots)
PowerPointppt LatticeQCD.ppt r1 manage 562.5 K 2007-09-21 - 17:28 AndrewMaier The (almost) empty poster. Need to be completed....
Microsoft Excel Spreadsheetxls cpuhisto.xls r1 manage 35.0 K 2007-09-21 - 17:26 AndrewMaier table with all the CPUs used. Good for creating more plots
PowerPointppt gangaPR-poster-EGEE2007-budapest.ppt r1 manage 16263.5 K 2007-09-27 - 17:36 JakubMoscicki updated version of PR
PDFpdf gangaPR.pdf r1 manage 17680.6 K 2007-09-27 - 15:57 MassimoLamanna PR poster
PowerPointppt gangaPR.ppt r1 manage 16235.5 K 2007-09-27 - 16:13 MassimoLamanna PR poster (PPT format)
PDFpdf gangaSU3.pdf r1 manage 1138.1 K 2007-09-27 - 16:11 MassimoLamanna Ganga SU3
PowerPointppt gangaSU3.ppt r1 manage 1817.5 K 2007-09-27 - 16:54 MassimoLamanna SU3 poster (PPT format)
PowerPointppt manchester.ppt r1 manage 3134.5 K 2007-09-19 - 11:24 AndrewMaier The Poster from the Manchester Users forum with updated Ganga Logos
Microsoft Excel Spreadsheetxls tld.xls r1 manage 46.5 K 2007-09-21 - 17:27 AndrewMaier table with all the top level domains used. Good for extra plots
Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2007-09-30 - PatriciaMendezLorenzo
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    ArdaGrid All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback