At 04:46 PM 8/14/00 -0700, you wrote:

>Well, the results compared to Interbase were based on a single benchmark
>(AS3AP), since they couldn't get the TPC benchmark to work with the
>Intersolv ODBC driver.

How ironic. I was just talking about the ODBC (or the lack thereof) driver
in the IBDI listserve.

>Also, they don't mention whether the Interbase version was classic or
>Superserver; I'd expect Classic to fall apart after a large number of users,
>based on the lack of a shared cache. The benchmark PC Week did a while ago
>showed the same thing (using 5.6 for Linux).

What I found interesting was that the text seemed to indicate that both NT
and Linux servers were used. Setting up postgres on NT is a royal pain in
the behind and I would expect it to run slower in the NT then native NT
apps. The article mentioned some graphs which were missing and "commertial"
databases I wonder which databases they were testing.

It is possible they did not have expertise in tuning IB or mysql and maybe
that was the problem. I was under the impression that superserver did not
run so well on linux maybe that's why they went with classic. It's
interesting reading but for me it raises more questions then answers.
----------------------------------------------
Tim Uckun
Mobile Intelligence Unit.
----------------------------------------------
"There are some who call me TIM?"
----------------------------------------------