I am an application engineer responsible for the HF/SI domain. I have some questions regarding pyAEDT script-based control of AEDT HPC settings on Linux systems that I would like to consult with you all about.
Background
My customer is using Q3D through 3D VNC on a Linux HPC cluster managed by PBS. The cluster consists of two compute nodes with approximately 128 CPU cores in total.
In a Windows environment, they are able to configure HPC settings through the Q3D GUI, including task distribution options such as Optimetrics Variations, Solution Types, and Frequencies. However, in their PBS/Linux environment, the HPC configuration appears as an Interactive Scheduler Job, and the configuration buttons are disabled, so they cannot modify the settings through the GUI.
My understanding of the customer's goal is that they are not trying to perform distributed solving of a single design. Instead, they would like to run approximately 50 independent Optimetrics variation tasks across the PBS cluster, with 2 CPU cores allocated to each task.
Question
After discussing this with the customer, they had some follow-up questions regarding the recommended AEDT Job Submission workflow.
Their current workflow is based on PyAEDT. They use PyAEDT to build the model, configure the HPC settings (number of tasks and cores), and then launch the simulation.
Since you suggested using AEDT Job Submission for scheduler environments, they would like to better understand how this workflow is intended to work.
Specifically, they are asking:
1.Is there any example or documentation showing how AEDT Job Submission can be used from a scripted workflow (for example, in conjunction with PyAEDT)?
2.Can they use PyAEDT to:
- Create or modify the project,
- Configure the simulation setup,
- Save the AEDT project,
and then submit the simulation through AEDT Job Submission?
3.Conceptually, what is the main difference between:
Using PyAEDT to configure HPC settings and launch the solve directly,
- Using AEDT Job Submission to submit the job to PBS?
4.When using AEDT Job Submission, is there a mechanism to retrieve job status, completion status, or simulation results programmatically after the job finishes?
The customer is trying to understand the recommended workflow for running large-scale Optimetrics studies in a PBS-managed cluster environment, particularly how PyAEDT and AEDT Job Submission are expected to work together.
I would greatly appreciate it if experts familiar with these issues could provide their insights. Thank you all very much for your support.
