Ansys Remote Solve Manager (RSM) is a versatile tool that enhances computational workflows by enabling users to submit simulation jobs to remote clusters or high-performance computing (HPC) systems. Whether managing large simulations or optimizing resource usage, RSM simplifies job scheduling, execution, and resource allocation.
This guide provides step-by-step instructions for configuring RSM on a Windows machine using the Ansys Remote Cluster (ARC) job scheduler. While this example focuses on ARC, RSM also supports third-party job schedulers like Platform LSF, PBS Professional, Slurm, Altair Grid Engine, and Windows HPC. For details on compatibility with other schedulers, refer to the Ansys support documentation.
These instructions specifically apply to RSM for Fluids/Structures tools. A slightly different setup is required for Electronics tools. For those steps, please refer to our dedicated blog here.
This tutorial demonstrates configuring a multi-node ARC cluster for RSM and setting up a custom job scheduling system. In this example:
[RSM_Install_Location]\ARC\bin\arcConfigConsole.exe.With the master node designated, the next step is to add the other two machines to the cluster as execution nodes. These nodes will handle the actual job processing. The master node can also participate as an execution node if the corresponding option was selected earlier. Once configured, each node will appear in the dialog box. By selecting a node from the list, you can adjust its settings, such as limiting the maximum number of cores it can use for job execution. This helps prevent overloading and ensures at least 1-2 CPUs remain available for the operating system's background tasks. Similarly, when additional execution nodes are added, they will be listed below the master node and can be customized to suit the cluster’s needs.
For this example, we will choose the following settings:
A job queue serves as a virtual pipeline where computational tasks (jobs) are organized, prioritized, and executed using the resources available within a cluster. Creating a custom queue allows you to optimize resource allocation for specific requirements—for example, reserving high-capacity nodes for memory-intensive simulations or capping the number of concurrent jobs on a machine to prevent overloading. Custom queues provide greater flexibility and control than default or local queues, making them a valuable tool for efficiently managing resources and ensuring critical tasks are handled according to predefined priorities. As an illustration, we will configure a custom queue accessible exclusively to HOST2, specifically for handling larger, memory-intensive jobs. For smaller clusters, the default and local queues are often sufficient. To set up a custom queue, follow these steps.
high_mem.Active Time Range: Check the "Set Active Time Range" box to ensure the queue remains active at all times.
Job Limit: Specify the maximum number of jobs that can run simultaneously in this queue. In this example, we’ll set the limit to 100 jobs.
Assign Execute Nodes: Assign specific nodes to this queue by navigating to the Access Control settings. For this example, select HOST2 as the designated execution node, as shown below.
To enable RSM client machines to submit jobs to the ARC submit host (MASTER), a configuration must be set up on the client side of the configuration to establish communication between the client and the submit host, define the file transfer method, and map RSM queues to the corresponding ARC cluster queues. To do this, we can follow the steps below.
high_mem).When submitting test jobs through the RSM Configuration tool, the status bar provides real-time updates on submission and execution progress. Upon completion, whether the job is successful or fails, a detailed diagnostics report is available. This report includes critical information for troubleshooting and ensures the configuration is correctly set up.
C:\SharedRSMConfig\ANSYS\v2##\RSM.ARC.rsmcc and queues.rsmq to the shared directory.%APPDATA%\ANSYS\v2##\RSM.Following these steps ensures a robust and efficient setup of Ansys RSM, enabling users to maximize simulation performance and streamline computational workflows.
Now that we have our RSM configuration setup, we can review how we can submit jobs to our new cluster in Mechanical or Fluent.
Mechanical Setup
In Ansys Mechanical, you can set up your job as usual on your local machine. However, when it's time to submit the solve, an adjustment is needed to direct Mechanical to use RSM. To configure this, open the Mechanical interface and navigate to File → Solve Process Settings. A window will appear similar to the one shown below. By default, the option "My Computer" will be selected. To enable RSM, switch this setting to My Computer, Background.
From here, you can configure the RSM setup. Set the Solution Execution Target to RSM, which specifies where the job will be submitted. Next, choose the appropriate queue for the job. While the default is "local," you can select a different queue if multiple options are available. Because the RSM Configuration Tool was set up in previous steps, the fields for HPC Configuration, HPC Type, and HPC Queue will automatically populate based on the saved configuration.
Finally, provide a job name (default is "Mechanical") and select the license to use for solving. Once everything is set, click OK, specify the number of CPUs to use for the solve, and proceed with solving the model!
Fluent Setup
Submitting jobs in Fluent requires a slightly different workflow. If you are using Workbench Fluent, you can use the process below. For standalone Fluent, the options can be specified from the launcher.
Access Solution Properties:
Configure RSM Submission Settings:
Select an RSM Queue:
Set Execution Mode:
Specify Parallel Processes (if applicable):
Submit the Job:
Manage the Update Process: