Subject RE: [Firebird-general] [ANNOUNCE] paper covering Firebird
Author Leyne, Sean

> I have just published a performance benchmark paper which includes
> Firebird in its tests.
> I hope you will find some of these results interesting, as I did.
> I am planning on making a new series of tests, including an optimized
> setup.
> Please bear in mind that the scope of the benchmark is limited (not
> trying to do what TPC does).

Unfortunately, I find, the graphics of the results are almost impossible
to read -- the lines on the graphs are not wide enough to differentiate
color so you can't follow the results. The legends are also impossible
to read.

While I have absolutely no experience with Java, I have a couple of
questions regarding the testing:

- When where the tests performed?
(Given that FB v2.0 has been released since November 2006, I would have
thought that it could have been used for the tests.)

- Did I read the results correctly; you were testing up to 1000+
simultaneous operations through separate threads?

- Did each thread establish its own database connection?