> 3. It takes quite a bit of time to do the generation.  The minimum
> generation time for the four documents are from 1 to 4 seconds (the 4
> seconds one is 5 pages with 13 tables with various cell colors and
> border styles).  In the 30 user test, the average time are from 20 to 40
> seconds, and maximum well over 2 minutes.

This sounds quite like the experiences by Jim Cotugno (See subject
"Multithreading"). Some time ago I did some concurrency tests and it was
interesting to see what happened. When multiple conversions occur
concurrently FOP seems to align the processing like this:

Let's say FOP does steps 1, 2 and 3 and we have conversions A, B, C and
D which run more or less at the same time. The following seems to
A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3

I happened to see that when I sent the output of every thread into the
same console. It might very vell be, that this has something to do with
the use of static variables. This also explains your experienced maximum
of over 2 minutes.

There's a lot we can do here (everybody can help):
- Identify the bad spots
- Streamline the generation process (as currently done by Mark)
- Write a simple performance benchmark to see if we're making any
  progress. Document the results.
- Rework the bad parts

The topic "Avalon" has come up a few times over the last months. It has
never got much attention. But it might help us to get rid of some of the
performance problems just by reworking parts of FOP for Avalon. This
could also result in a stable high-level interface for FOP which would
probably cheer up a few people as well (Cocooners, Servlet-integrators
etc.). What do you think?

I know we're not really in the optimization phase but looking at the
number of Foppers.....

Jeremias Märki


Postfach 3954 - Rhynauerstr. 15 - CH-6002 Luzern
Fon +41 (41) 317 2020 - Fax +41 (41) 317 2029
Internet http://www.outline.ch

To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to