> Yeah, I suspect there's 10% or more noise in these numbers. But then
> one could read the results as saying we can't reliably measure any
> difference at all ...
>
> I'd feel more confident if the measurements were done using operators
> repeated enough times to yield multiple-second runtimes. I don't
> trust fractional-second time measurements on Unix boxen; too much chance
> of bogus results due to activity of other processes.
Any idea to do that? I tried to do a measurements using something like
"SELECT * FROM t1 WHERE very-long-string-column LIKE 'aaa'", but I'm
afraid the I/O time masks the difference...
--
Tatsuo Ishii
