My test was trying to simulate (in a poor way, surely) a high traffic in a
website. In this environment, I cannot do a single transaction. I could have
used a parameterized query for all the three DBMS (I don't remember why I
didn't do it ;)).

I was not telling that I do the best for every DBMS (too much for me), simply
I do the same thing in all them. I am sure somebody can think this is not a
good method, although I have not time to make me an expert DBA, I need my own
benchmark.

(note: My opinion is that Interbase is the best OpenSource software until
today. This message is sent from a Web Groupware System which I have written
the Interbase layer)

>
> >From the look of this code you are not wrapping all the insertions in a

single

> transaction. Neither are you using a prepared, parameterised query. Those

two

> changes alone would vastly improve the InterBase performance.
>
> I know you are trying to compare like with like, but one of the reasons that
> mySQL performs so well is because it doesn't really have transactions. For

all I

> know it may not have prepared, parameterized queries, either. It assumes

that

> only one user is inserting data and that user knows what he is doing.

InterBase

> starts from different assumptions.
>
> See if you can run some tests with the above changes and let us know.
>

If I have some time I will do the tests and of course I will publish the
results in this list.

> Paul
> --
>
> Paul Reeves
> Fleet River Software
>

Thanks for your suggestions, Paul.

(my english continues being very bad. sorry.)

--
Programmer
Integracion de Metodologias y Sistemas (Spain)
http://www.netims.com