On 2013-09-24 19:56:32 +0200, Andres Freund wrote:
> On 2013-09-24 13:51:04 -0300, Alvaro Herrera wrote:
> > David Rowley escribió:
> >
> > > I do see a 15-18% slow down with the patched version, so perhaps I'll need
> > > to look to see if I can speed it up a bit, although I do feel this
> > > benchmark is not quite a normal workload.
> >
> > Ouch. That's certainly way too much. Is the compiler inlining
> > process_log_prefix_padding()? If not, does it do it if you add "inline"
> > to it? That might speed up things a bit. If that's not enough, maybe
> > you need some way to return to the original coding for the case where no
> > padding is set in front of each option.
>
> From a very short look without actually running it I'd guess the issue
> is all the $* things you're now passing to do appendStringInfo (which
> passes them off to vsnprintf).
> How does it look without that?
That's maybe misunderstandable, what I mean is to have an if (padding >
0) around the the changed appendStringInfo invocations and use the old
ones otherwise.
Greetings,
Andres Freund
-- Andres Freund http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services
