Non-optimization jobs

Moldflow analyses which consist of a linear analysis sequence, such as Cool + Fill+Pack + Warp, can be executed without using the Moldflow Job Manager.

It is recommended that you create a unique directory on a node's read-write scratch area, to protect against jobs interfering with each other, and to facilitate job cleanup after the analysis has finished.

Note: Non-optimization jobs cannot be spread across multiple nodes; they must run entirely on one node.

Transfer the study file to a node

Using a file transfer tool, copy the study file (*.sdy) to a directory on a node's read-write scratch area.

Tip: This directory will be referred to as $TARGET in the remaining steps. Substitute this with the actual directory that you use.

Execute a shell on the node

To perform the analysis on a node, start an interactive Bourne shell and run the following commands, or make a shell script to perform the commands in order.

1. Change directory to $TARGET
2. Source $INSTALLATION/mfclient.env
3. Execute the runstudy command on the study file. You may wish to redirect standard output into a file.

After the analysis completes, the runstudy command returns, and $TARGET will contain result files.

Transfer the result files from the node

Using a file transfer tool, copy the newly-created files back from $TARGET. Synergy can read these result files directly if they are in the same directory as the study.