First step of pygate

In this page,we will run a example consists of two steps;

1. File main.mac generates.

• Before a GATE process,we generally need to make a file of main.mac ,in which we can set the compomnents( world,system,phantom,soure and digitizer etc. ) for GATE simulation.
• The question is that the definition of compomnents is too heavy and complicated. A lot of time and ennergy wastes on this.
• Pygate offers a function of File Gnetators .Users can confirured the compomnents easily by set several arguments of neccessary compomnents you want in a file of make_mac.py (you can name it freely). A file of mac.yml may needed containing some default settings of the main.mac ,or you can set these in make_mac.py directly.

• The command for main.mac generating is:

  • $ pygate generate mac script -t make_mac.py -o main.mac -c mac.yml .
  • make_mac.py and mac.yml should be included in current work folder.
  • Then you will find the main.mac in the folder.

• The main work for this step is to code make_mac.py .Usres can modify on the template for first time.

:ref:make_mac.py

2. Submit a task to subsystem

• Usually there are hundreds of millions,even billions events occuring during a GATE simulation,which is the reason why the process take a long time.
• However,it a repetable work for a GATE program to generate an event.*Pygate* offers a method to speed up the process.*Pygate* divides the task into a lot of parts,
• then submits these parts to server.Each part of the original task will be distributed to no-wroking machine of the net by SLurm.Thus,we get a very high speed for GATE simulation.

• Users should know the following steps to archive it:

  1. Users should get the needed configured files by excuting this command: $ pygate init ext .You will get the files in folder:

     • main.mac ,you get it last in the fomer.
     • GateMaterials.db ,significant file for GATE configuration,can’t be lack.
     • Hits2CSV.C ,may needed if you want the data of csv format.
     • Materials.xml
     • Surface.xml ,set the surface rendering.Or you can seclect volume rendering.

  2. When you get the neccessary files in the work folder,you need to divide the task into parts.

     • $ pygate init subdir -n --INTGER -f --STR ,you can set the number of parts and the name of subdirectories as you want.The default option is “sub.[10]” and you will get 10 subdirectories of “sub.[x]”(x~[0-10]).
     • $ pygate init bcast ,broadcast the files to subdirectories maken last step.
     • $ pygate generate shell ,generate run.sh for SLurm to distribute the task and post.sh to merge the results of each parts.
     • $ pygate submit ,submit the task to subsystem. SLurm will do the disribution.The details information of disribution will print on the screen.You can easily know which machine each part run.

     • There are two procedures before getting results:

       • First,the machines absorbs the mission and complete it,then feedback the results to subdirectories. run.sh is for this step.
       • Then the results from subdirectories are merged into one file of optical.root ,containing all collected data of Hits. post.sh is for this.

You can refer the detail of commands in Here is the command list of pygate