Select page from this chapter: Multiple Tasks, Project Sharing

Project Sharing

Overview

This page describes briefly how you can share projects across multiple machines or users. Each BWTT "project" is physically represented by a folder containing a number of files, each of which represents a "job". Since exclusive locking is implemented on a per-job basis, it is perfectly possible for the project to be open in two places at once, with some jobs locked in one place and some in the other. This means that you can be preparing to compute or reviewing results from some jobs in your office, while in the back office another machine is busy computing some of the jobs. We illustrate by example.

WARNING: This functionality has not been tested to any great degree. Please make sure you keep backups of any of your hard work.

Example

* Start a project, and add 100 jobs to it.

* Parametrize 20 of them, then unlock them by either of these methods:

- close the project, and reopen it, to release all locked jobs.

- select those 20 jobs, and call "Release Write Lock" on them (from the "Selected" menu).

* Open the project on your back office machine - all listed jobs will remain unlocked. Select the first 20, and press the "Play" button to start them computing. These jobs will now be locked by this machine.

* Return to your office, and start parametrizing the next 20 jobs. You will not be able to open the first 20 for writing, though you can open them for reading, which allows you to keep track of their status.

* After parametrizing the next 20 jobs, release the locks on those ones, and return to the back office, opening the project on yet another computer and setting it to compute these jobs.

* Carry on parametrizing the next 20 jobs. You will probably find that some of the first 20 are starting to complete, now - as they complete, they will be unlocked, so you can start reviewing them whilst they are still being computed.

* Rinse, and repeat.

In this way, you can distribute the processing of a large project across multiple machines, rather than having to break it up into smaller projects yourself. Similarly, two users in different places can work on different jobs from the same project.

If processing fails on any of the machines, you may find you get some use out of the command "Project/Select/All ready to run", which is also available through CTRL+R (APPLE+R).