PDS Basics

A single PDS file corresponds to a single GridServer Service. The computation represented by the Service usually does the following:

Split up the input data into several pieces, one for each task.

Run the tasks in parallel on the Engines.

Collect and combine the results.

If the data is too large to be passed as command-line arguments to the program running on the Engine, then place it into files. You can locate these files in a shared directory, or copy them to the Engine and copy back the result files.

The PDS language contains constructs for carrying out various statements, such as running executables and copying files, at each point in the lifetime of a Service.

A few words about the lexical structure of PDS: whitespace is not significant, but the case is. All text from the “#” character to the end of the line is a comment and is ignored.

Did you find this helpful?

Yes No