This project is read-only.

PhyloDCcs.exe

A tool to submit PhyloD.exe jobs to a Windows Compute Cluster.

Usage:

PhyloDCcs -help
PhyloDCcs [many parameters]

Required Arguments:

Optional Arguments:

  • -keepTest, default AlwaysKeep
  • -skipRowIndexFile, a file that tells which output rows to skip because they have already been run, e.g. a file containing "100-200,300-400")
    • As described below, Tabulate.exe audits the output of PhyloD.exe and reports which rows were found. Use this parameter to have PhyloD.exe run again and produce the missing output rows.
  • -optimizer BrentThenGrid (default) OR BrentWithNoGrid OR Grid
    • You can append a tolerance to the Brent methods: e.g. BrentThenGrid.001.
    • The default Brent tolerance is .001.
  • -cluster,
    • The name of the cluster's headnode.
    • The default is to extract from "externalRemoteDirectory"
    • For example, Cluster1
  • -internalRemoteDirectory
    • The compute nodes' name for directory containing your input files.
    • The default is to be the same as the externalRemoteDirectory.
    • (Some clusters are configured so that the cluster's compute nodes have a different name for +the file
share containing user files.)

Output

In addition to the normal output from PhyloD.exe, PhyloDCcs also creates these directories under the externalRemoteDirectory. You may wish to periodically delete these files to free up disk space.
  • exes, where it copies a snapshot of the current PhyloD.exe and associated *.dlls
  • Stderr, the standard error output of each cluster tasks. This can be very useful in understanding why a task failed.
  • Stdout, the standard output of each cluster task. Also very usefuil in understanding why a task failed.

Configurating a Windows Computer Cluster for PhyloDCcs and HlaAssignCcs

Last edited Apr 20, 2007 at 12:10 AM by CarlK, version 8

Comments

No comments yet.