The problem is that BIRT needs to be started once, and only once, in any given process. Once it has been started up, then it is plain sailing to load reports and render them to HTML. The BIRT runtime should be kept around until the process shuts down.

Two problems came up with this approach - if each PHP process makes its own BIRT process any settings/current data will be wiped out and the BIRT instance would go away whenever the PHP script finishes. He found something that solves both of these problems - the Zero Global Context. Its a "storage area" that can contain just about anything. He used it to store the BIRT runtime until its told to finish and die off.

Each PHP request is a separate thread. There are some workarounds like using pcntl_fork, starting multiple commandline php processes using the exec command or even using ajax. Another possibility is using the Curl library. Besides the basic functions described above Curl offers the "multi" functions for retrieving content from several url's at the same time.

Examples are included showing how to make these "multi" requests via PHP's cURL support (calling URLs with a "seconds" value on the end).

He does, however, branch off into something much more interesting that seems to be somewhat ignored by developers - the use of ticks on their code:

This is something named "ticks" - I wonder how many of the PHP developers heard about it and of those how many actually used it. Could it be used for offloading long-running I/O-bound tasks or grouping them together (e.g. so we could wait for DB and HTTP in parallel and not sequentially)? Would there be any use at all for such functionality and if so - how it's supposed to work? I.e. how would you know it's done and how you would collect and use the results?

It's suggested in the comments that it could be used for any kind of application that might need the pseudo-multithreading it offers (including something like scripts needing multiple TCP connections).

All threads can now both hear and report on conditions. A problem with the original send/receive mechanism made threads hang the controlling thread while long operations were occurring.

After updating some of their HVAC class to make it a little easier to understand/use, they ran some benchmarks on running processes two different ways - multithreaded and sequential.

The package is available for download if you'd like to work with it/run the tests yourself.