"Greg Sabino Mullane" <greg@turnstep.com> writes:
> I can't imagine that it was causing a factor of 2 difference, but I've
> changed the code in DBD::Pg to make do() a lot smarter, for brain-dead
> benchmarks such as MySQL's that use do() instead of prepare/execute
> like they should. In short, DBD::Pg will short-circuit a lot of the usual
> work and call PQexec directly if it can get away with it, which should
> make the profile Tom posted look a lot better. It also avoids overhead of
> going through Pg.pm, and is all C/XS code now. Hopefully, this should
> make the benchmarking results a little more fair. :) This is now in
> cvs, and will be part of the upcoming version 1.44 of DBD::Pg.
Cool. Thanks for doing that. Do you have an idea when 1.44 will
be out?
regards, tom lane