ANSYS Electronics Job Submission using End-to-End Desktops

Introduction

End-to-end Desktops create a seamless HPC experience on Rescale. End-to-End Desktops is a batch compute cluster with desktop visualization capability on the head node of your cluster. It provides a user interface for users to interact with the job setup directly from the analysis software.

Features of End-to-End Desktops

  • Enables you to work with your batch job submission interactively. You can monitor jobs on the fly
  • Enables you to dictribute your batch job submission across the entire cluster using an automated process
  • You can pause the analysis at anytime, edit the inputs or paramters of the job and either restart or rerun the analysis
  • Job results can be viewed during and right after completion of the simulation without any delays
  • Files are automatically synced from the batch job to the Desktop session and vice-versa
  • There is no additional charges for the visualization

Key Notes

  • The Desktop session is CentOS Linux
  • The files are synced to the work directory of the desktop session
  • Users need to explicitly terminate the job to shutdown the cluster

ANSYS Electronics Tutorial

This tutorial shows how to submit an ANSYS HFSS job using End-to-End Desktops. The tutorial shows you how to set up an ANSYS Electronics job with a parameter sweep on a 2 node cluster.

Users that submit jobs to the workstation, are limited to the resources on the workstation. ANSYS provides an option to scale up and submit jobs to a larger, remote cluster. However, the workflow is complicated. End-to-End Desktops automates this process and hence provides the option to submit jobs to a large cluster with ease. ANSYS Optimization and Parameter sweeps are complicated and End-to-End Desktops provide a seamless methodology to run such jobs on the Rescale platform.

The steps are as follows:

  • Click on the + New Job icon on the top left hand corner of the dashboard
  • Name the job
  • Click on Upload from this computer and browse to the location where the job is saved. Select the file and click Open

e2e-aedt-file

  • Jump to the Software Settings page, and select ANSYS Interactive Workflow

e2e-attach-software

  • Check the box for Make this job interactive?
  • Select the software version as 19.2
  • Select the desired licensing option
    If you are using a Rescale License, you may select the Use On-Demand License by checking the box as shown below. Also, check the box to agree to the Software Licensing Agreement.
    Alternatively, users using their own licenses may select the Use Existing License and specify the relevant licensing information.

software settings

  • Next, proceed to the Hardware Settings page and select the hardware configuration that you would like to run your simulation on. In this example, we select the Nickel hardware as shown below

  • Set the Number of Cores to 32

  • Also, set a suitable wall time so that the cluster will be terminated after the simulation

  • Now, you can click on Submit on the top right of the screen

e2e-hardware-settings

Once the job is launched, you will notice that it looks very similar to a batch job with an additional button to Connect Interactive Session on the top right of the screen

  • Go to the Status Tab to view the progress of the job submission

  • Click on Connect Interactive Session on the top right of the screen. This will open an in-browser desktop Desktop session for End-to-End Desktops that is running Linux

e2e-job-status

  • Once the Desktop is launched, you will see the icon for ANSYS Electronics Desktop. Click on the icon to launch

e2e-launch-ansys-electronics

  • Once ANSYS Electronics is launched, click on Open. Browse to the shared folder as shown below and select the input file. Click on Open. Note that the input file can be .aedt or .aedtz

e2e-open-hfss-job

  • The Project Manager tree will be displayed on the top left. In this example, we will be solving a parametric sweep found under the Optimetrics entry in the project tree. Navigate to Optimetrics and then to Parametric Setup. Right click on it and click on Submit Job

e2e-submit-job

  • The will open the Submit Job dialog box. Click on Compute Resources tab on the top

e2e-submit-job-options

  • Click on the Import button and navigate to the cluster_config.areg file located on the Desktop. The cluster_config.areg is auto-generated on cluster start up and contains cluster IP adresses and number of cores on each node. Click on the file and click Open

e2e-cluster-file

The information from the config file into the dialog box. Please note that when the Use automatic settings checkbox is checked, the tasks are auto-distributed. When unchecked, you can specify the numbers of tasks to solve simulatenouesly. In the example here, we are auto-distributing the tasks, so the box is checked

  • Click on Submit Job

  • On the dialog box that opens, click on Yes to close the project

e2e-submit-aedt-job

  • A dialog box will be diplayed with the message that the job has been submitted. Check the box next to Begin monitoring this job now and click on OK

e2e-job-submitted

  • Next, the Monitor Job dialog box will be displayed. In the Messages window on the bottom left, you will see status messages regarding the jobs

  • The Progress tab indicates the Analysis progress e2e-status-message

  • The Profile tab on the tab should show additonal information like regarding the analysis

  • Once the job is completed solving, you can click on Close button on the Monitor Job dialog box

e2e-job-profile

  • To view the results, click on Open and browse to the shared directory. Click on the .aedt file and click on Open. Note that the input file can be .aedt or .aedtz

e2e-open-solved-job

  • On the Project Manager tree, right click on Parametric Setup and click on View/Analyze Results

e2e-view-aedt-results

e2e-results

  • The files are automatically saved to the work folder
  • Terminate the job to end the Desktop session. Any files saved to the work folder will be uploaded to the job files under Results on the platform