[Xapian-discuss] Initial benchmark results quartz and flint

[Olly Betts]

On Wed, Jun 29, 2005 at 01:19:54PM +0200, Arjen van der Meijden wrote:
> It appears from these results that flint is significantly faster to 
> search in, both with phrase-queries and normal queries.

Well, that's good!

> I cannot explain from the hardware or benchmark setup why the compacted 
> quartz databases are so much slower with phrase.

This is indeed most puzzling (although ultimately flint is faster than
either which is what really matters).

Your test does seem to be carefully designed to eliminate various possible
hand-waving explanations I can think of.

> To be sure it are not one-time-only numbers, I'm running the benchmarks 
> twice more but since that'll take almost a day per run I sent these 
> numbers to the list already.

This would be useful - my only current thought is that something else was
happening on the machine at the time (the "update the locate database"
cronjob which usually runs once a day is a prime candidate for this sort
of thing - it does a pretty effective job of flushing the kernel's cache
of disk blocks and swapping out any process which hasn't is currently
inactive).

Cheers,
    Olly