Edit detail for Running Abaqus on the servers revision 1 of 11

1 2 3 4 5 6 7 8 9 10 11
Editor:
Time: 2007/08/15 13:08:59 GMT+0
Note:

changed:
-
This <a href="./Abaqus_Tutorial_01.pdf">Abaqus tutorial</a> is a good place to start regarding the use of Abaqus. Once you understand how to create an input file you can do the following to have the job run on the server.

Also, <a href="http://abaqusdocs.ecn.purdue.edu:2080">Abaqus documentation is online</a>.

We are going to follow very close what is done in another <a href ="RunningJobsOnTheServers">document</a> that generally describes how to run jobs on the servers.

**Abaqus is fully supported only on Linux and Windows. The Unix servers have a very old version of Abaqus**

1 **'ssh'** into a server.

2 You will need an Abaqus input file.

3 You will need to place your job in a shell wrapper. This is done like so:

  - Decide where you want to have the wrapper (or script). 
    For this example I will have it in my home directory (or N: drive on windows)
    
  - Create the script, I will name it **'run.sh'**.

  - You can use whatever editor you like (emacs, vi, pico, whatever).

  - Let's take a look at a sample wrapper::

    Listing for run.sh
    =================================================
    =1= #!/usr/bin/sh
    =2=
    =3= cd /home/roger/a/gizmo/research/abaqus_model
    =4=
    =5= /usr/opt/bin/abaqus job=input_file_name
    =================================================


  The numbers =X= is the line number, **NOT** part of the script.

  Line 1: The shebang line, just letting the OS know where to find the
          interpreter for the script.

  Line 3: I'm explicitly changing to the directory where the job I want
          to run is. I am giving the full path so that there is no doubt
          that I will be in the proper directory.

  Line 5: This is the what we really want to run. Once again I'm
          giving the path to Abaqus.

4 We need to make this script executable: 

  **'chmod +x run.sh'**

5 Now we can send this off to a server and forget about it while it
  chews on the CPU. To do this we use the **'-f'** flag for **'ssh'**. The syntax is:

  **'ssh -f server_name file_to_run'**

  For example:

  **'ssh -f yuri /home/roger/a/gizmo/run.sh'**

  You will be prompted for your password. Once you enter it the program
  runs.

  This <a href="https://engineering.purdue.edu/AAE/help/wiki/DepartmentServers">page</a>
  lists the department servers. The compute servers are the ones to use.


This Abaqus tutorial is a good place to start regarding the use of Abaqus. Once you understand how to create an input file you can do the following to have the job run on the server.

Also, Abaqus documentation is online.

We are going to follow very close what is done in another document that generally describes how to run jobs on the servers.

Abaqus is fully supported only on Linux and Windows. The Unix servers have a very old version of Abaqus

  1. ssh into a server.
  2. You will need an Abaqus input file.
  3. You will need to place your job in a shell wrapper. This is done like so:
    • Decide where you want to have the wrapper (or script). For this example I will have it in my home directory (or N: drive on windows)
    • Create the script, I will name it run.sh.
    • You can use whatever editor you like (emacs, vi, pico, whatever).
    • Let's take a look at a sample wrapper:
          Listing for run.sh
          =================================================
          =1= #!/usr/bin/sh
          =2=
          =3= cd /home/roger/a/gizmo/research/abaqus_model
          =4=
          =5= /usr/opt/bin/abaqus job=input_file_name
          =================================================
      

    The numbers =X= is the line number, NOT part of the script.

    Line 1: The shebang line, just letting the OS know where to find the interpreter for the script.

    Line 3: I'm explicitly changing to the directory where the job I want to run is. I am giving the full path so that there is no doubt that I will be in the proper directory.

    Line 5: This is the what we really want to run. Once again I'm giving the path to Abaqus.

  4. We need to make this script executable:

    chmod +x run.sh

  5. Now we can send this off to a server and forget about it while it chews on the CPU. To do this we use the -f flag for ssh. The syntax is:

    ssh -f server_name file_to_run

    For example:

    ssh -f yuri /home/roger/a/gizmo/run.sh

    You will be prompted for your password. Once you enter it the program runs.

    This page lists the department servers. The compute servers are the ones to use.