1.5.2. Specifying Fluent Launcher Settings Within Workbench

You can use Fluent in Workbench on either Windows or Linux machines, both interactively and in batch mode. You can also start Fluent on Linux from a Workbench session running on Windows and you can use the same Fluent-specific project (and related files) in Workbench using both Linux and Windows hardware interchangeably. The information is this section is the same for both Windows and Linux, except where noted.

When you start ANSYS Fluent from either type of Fluent-based system within Workbench, Fluent Launcher will appear by default. Most Fluent Launcher settings are available, except for the following options:

The Do not show this panel again option allows you to bypass Fluent Launcher for subsequent Fluent sessions. This option is only available when running Fluent in Workbench.


Important:  When using LSF to schedule a Fluent run from Workbench, by default the working directory is used for checkpointing. You can specify an alternate directory for checkpointing using the LSB_CHKPNT_DIR environment variable.



Important:  Note that, when using Fluent with Workbench on Linux, the Fluent Root Path option is disabled in the General tab of Fluent Launcher.


To start your Fluent simulation on a Linux cluster from a Workbench session running on Windows, use the Use Remote Linux Nodes option in Fluent Launcher. This option is available when the Parallel option is enabled under Processing Options. Once this option is enabled, the Remote tab is visible when you click the Show More Options button.


Important:  The Remote tab of Fluent Launcher can only be used for 64-bit Linux machines.


For more information about Fluent Launcher, see Starting ANSYS Fluent Using Fluent Launcher. For more information about using Fluent Launcher to access remote Linux clusters, see Setting Remote Options in Fluent Launcher.

1.5.2.1. Specifying Fluent Launcher Settings Using Cell Properties

You can view the properties of a selected cell in Workbench by selecting the Properties option under the View menu, or by right-clicking a cell and selecting Properties from the context menu. The properties of the selected cell are displayed in the Properties Window in Workbench.

Figure 1.7:  Properties for Fluent-Based Systems in Workbench

Properties for Fluent-Based Systems in Workbench

The Fluent-based system Setup and Solution cells have the following properties that you can set for Fluent Launcher:

Use Setup Launcher Settings

(for Solution cells only) allows you to specify that the current systems Solution cell should use the Fluent Launcher property settings from the current system’s Setup cell. By default, the Fluent-based system Setup and Solution cells share Fluent Launcher settings, however, if required, these settings can be set autonomously using this property.

Precision

allows you to choose either the single-precision or the double-precision solver

Show Launcher at Startup

allows you to show or hide Fluent Launcher when Fluent starts.

Display Mesh After Reading

allows you to show or hide the mesh after the mesh or case/data is read into Fluent.

Embed Graphics Windows

allows you to embed the graphics windows in the Fluent application window, or to have them free-standing.

Use Workbench Color Scheme

allows you to use either the classic black background color in the graphics window, or to use the Workbench color scheme.

Set up Compilation Environment for UDF

allows you to specify compiler settings for compiling user-defined functions (UDFs) with Fluent.

Use Job Scheduler

allows you to specify settings for running Fluent with the available job scheduler. On Linux, you can use either LSF, SGE, or PBS Pro, as the job scheduler. On Windows, you can use the Microsoft Job Scheduler, or, when Run Parallel Version and Use Remote Linux Nodes are selected, you can use the LSF, SGE, or PBS Pro job schedulers.

Run Parallel Version

allows you to choose to run the parallel version of Fluent or not.

UDF Compilation Script Path

(if Set up Compilation Environment for UDF is selected) allows you to specify the path to the UDF compilation script.

Initialization Method

allows you to initialize your Fluent simulation. Available options are:

  • Program Controlled (default) allows Fluent to use any existing results (solution data) to initialize the solution.

    • For a new calculation, if no results are available, then the system initializes the solution using the settings specified in Fluent.

    • For existing projects, if the system is not connected to another upstream system through the Solution cell, then the system uses results available from the previous calculation. If the results are not compatible with the mesh, then the system initializes the solution using the settings specified in Fluent.

    • For existing projects, if the system is connected to another upstream system through the Solution cell, then the Initialization Method is not available for the downstream Solution cell. In this case, the downstream Solution cell always uses results coming from the upstream Solution cell to initialize its solution for the initial design point (DP0) as well as any other design points. If the solution data file from the upstream Solution cell is not available or not compatible with the mesh, then Workbench uses the upstream data interpolation (*.ip) file for solution initialization. If the *.ip file is also not available, or the results are not compatible with the mesh, then the system initializes the solution using the settings specified in Fluent.

  • Solver Controlled allows Fluent to always initialize the solution using settings specified in Fluent, enforcing the initialization method available in Fluent for new and existing projects. This option is not available for Solution cells connected with upstream Solution cells.

  • Use Solution Data from File allows you to import a solution data file (*.dat) or data interpolation file (*.ip) that Fluent will use to initialize the solution. If selected, you are prompted to provide an Initial Data File. For new and existing calculations, the Solution cell uses a registered file as the initial data in Workbench.

For more information, see Using the Update Command.

Use Remote Linux Nodes

(available only when Run Parallel Version is selected) allows you to run your Fluent simulation on 64-bit Linux machines.

If Run Parallel Version is selected, the following additional properties are available:

Number of Processors

allows you to set the number of processors you want to use for the parallel calculations (for example, 2, 4, and so on).

Interconnect

allows you to set the interconnects you want to use for the parallel calculations (for example, ethernet, myrinet, infiniband, and so on).

MPI Type

allows you to set the MPI type you want to use for the parallel calculations (for example, ibmmpi, intel, and so on).

Use Shared Memory

allows you to specify if shared memory is to be used or not.

Machine Specification

(if Use Shared Memory is not selected) allows you to specify a list of machine names, or a file that contains machine names.

Machine List

(if Use Shared Memory is not selected and Machine List is selected as the Machine Specification) allows you to specify a list of machine names to run the parallel job.

Machine Filename

(if Use Shared Memory is not selected and File Containing Machine List is selected as the Machine Specification) allows you to specify the name of the file that contains a list of machines to run the parallel job.

If Use Remote Linux Nodes is selected, the following additional properties are available (for 64-bit Linux machines only):

Remote Fluent Root Path

allows you to specify the remote Fluent Linux installation root path.

Use Specified Remote Working Directory

allows you to specify a directory other than temp directory as the working directory for the remote Linux nodes. When this property is selected, you can specify the directory in the Remote Working Directory property that appears.

Remote Spawn Command

allows you to specify the command to connect to the remote node (the default is RSH).

Use Remote Cluster Head Node

allows you to specify the remote node that Fluent will connect to for spawning (for example, via rsh or ssh). When this property is selected, you can specify the remote node in the Remote Host Name property that appears.


Important:  When the PBS Pro option is selected for the Job Scheduler, the host specified in the Use Remote Cluster Head Node field should be the PBS Pro submission host.


If Use Job Scheduler is selected, the following additional properties are available:

Computer Cluster Head Node Name

allows you to specify the name of the compute cluster head node.

Job Template

(available only when running Windows HPC 2008 Server Scheduler) allows you to create a custom submission policy to define the job parameters for an application. The cluster administrator can use job templates to manage job submission and optimize cluster usage.

Node Group

(available only when running Windows HPC 2008 Server Scheduler) allows you to specify a collection of nodes. Cluster administrators can create groups and assign nodes to one or more groups.

Processor Unit

(available only when running Windows HPC 2008 Server Scheduler) allows you to choose the following:

  • Core refers to a single, named host in the cluster.

  • Socket refers to a set of processors with a dedicated memory bus. This is also known as a non-uniform memory access (NUMA) node.

  • Node refers to an individual CPU on a node. For example, a dual-core processor is considered two cores.

For more information about running Fluent jobs using the Windows HPC 2008 Server Scheduler, see the Frequently Asked Questions section of the Customer Portal.

Start When Resources Are Available

allows you to start the job when resources are available.

Create Job Submission XML

allows you to create the job submission XML file.

Job Submission XML File

(if Create Job Submission XML is selected) allows you to specify the name of the job submission XML file.

If Use Job Scheduler and Use Remote Linux Nodes are selected, the following additional properties are available:

Job Scheduler

allow you to specify the job scheduler for the remote Linux nodes. Available options are LSF (the default), SGE, or PBS Pro.

If LSF is selected as the job scheduler, the following additional properties are available:

Use LSF Queue

allows you to use an LSF queue.

LSF Queue

(when Use LSF Queue is selected) allows you to specify the name of the LSF queue.

Use Checkpointing

allows you to use checkpointing with LSF.

Enable Automatic Checkpointing

(when Use Checkpointing is selected) allows you to automatically checkpoint within a specific period of time.

Checkpointing Period

(when Use Checkpointing and Enable Automatic Checkpointing are selected) allows you to specify a time period for automatic checkpointing.


Important:  When using LSF to schedule a Fluent run from Workbench, by default the working directory is used for checkpointing. You can specify an alternate directory for checkpointing using the LSB_CHKPNT_DIR environment variable.


If SGE is selected as the job scheduler, the following additional properties are available:

SGE Qmaster

allows you to set the machine in the SGE job submission host list.

SGE Queue

allows you to set the queue where you want to submit your Fluent jobs. Leave this field blank if you want to use the default queue.

SGE PE

allows you to set the parallel environment where you want to submit your Fluent jobs. Leave this field blank if you want to use the default parallel environment.

Use SGE Settings

allows you to use SGE settings.

SGE Settings

(when Use SGE Settings is selected) allows you to specify SGE settings.


Important:  When running Fluent in Workbench via SGE, there may be instances when not enough time is allotted for Fluent to start. By default, Workbench performs two steps: starting the Fluent process and establishing initial contact, with a default wait time of one minute; completing the Fluent start up process and establishing a final connection (for example, to spawn parallel nodes, to checkout licenses, and so on), with a default wait time of five minutes. You can change the waiting time increments for each step by defining the Fluent_WB_MAX_STARTUP_WAIT environment variable before starting Workbench. Note that you cannot define this environment variable within Fluent Launcher. The value of this environment variable is equivalent to the wait time in minutes for each start up step. So, the overall wait time is double the value specified for the environment variable. Fractions are allowed as values for the environment variable, and any value less than or equal to zero will let Workbench wait indefinitely for Fluent to start. Note that while Workbench is waiting for Fluent to start, the Workbench interface is locked, and if, for some reason Fluent fails to start, you will have to manually kill the Workbench session.


If PBS Pro is selected as the job scheduler, the following additional property is available (on Linux only):

PBS Submission Host

(optional) allows you to specify the name of the PBS Pro submission host if the machine you are using to run Workbench cannot submit a job to PBS Pro.

These properties are the same as the settings within Fluent Launcher. For more information about Fluent Launcher, see the Fluent Getting Started Guide.

Properties for Setup and Solution cells related to problem setup and solution processes are discussed in Specifying Other Setup and Solution Cell Settings.

1.5.2.2. Copying Fluent Launcher Property Settings

Typically, Fluent Launcher settings are specified in the Fluent Launcher dialog box when Fluent is launched from the Setup cell.

By default, the Solution cell uses the same Fluent Launcher property settings as the Setup cell. If you want the Solution cell’s Fluent Launcher settings to be different than those specified for the Setup cell, you can disable the Use Setup Launcher Settings property setting (see Specifying Fluent Launcher Settings Within Workbench) and then set the Solution cell’s Fluent Launcher property settings without impacting the settings for the Setup cell.

You can copy the Setup cell’s Fluent Launcher settings to the Solution cell by selecting the Copy Launcher settings to Solution cell command from the Setup cell’s context menu (available only when the Use Setup Launcher Settings property setting is disabled). Note that the values that are written are retained, even if you later disable this setting. Likewise, you can also copy the Solution cell’s Fluent Launcher settings to the Setup cell by selecting the Copy Launcher settings to Setup cell command from the Solution cell’s context menu (available only when the Use Setup Launcher Settings property setting is disabled).


Release 18.2 - © ANSYS, Inc. All rights reserved.