|From:||Andres Freund <andres(at)anarazel(dot)de>|
|To:||Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>|
|Subject:||Re: Performance improvements for src/port/snprintf.c|
|Views:||Raw Message | Whole Thread | Download mbox | Resend email|
On 2018-08-17 14:32:59 -0400, Tom Lane wrote:
> I've been looking into the possible performance consequences of that,
> in particular comparing snprintf.c to the library versions on macOS,
> FreeBSD, OpenBSD, and NetBSD. While it held up well in simpler cases,
> I noted that it was significantly slower on long format strings, which
> I traced to two separate problems:
> Perhaps there's a way to improve that
> without writing our own floating-point conversion code, but I'm not
> seeing an easy way offhand. I don't think that's a showstopper though.
> This code is now faster than the native code for very many other cases,
> so on average it should cause no real performance problem.
I kinda wonder if we shouldn't replace the non pg_* functions in
snprintf.c with a more modern copy of a compatibly licensed libc. Looks
to me like our implementation has forked off some BSD a fair while ago.
There seems to be a few choices. Among others:
- freebsd libc:
(floating point stuff is elsewhere)
- musl libc:
- stb (as Alexander referenced earlier)
I've not benchmarked any of these. Just by looking at the code, the musl
one looks by far the most compact - looks like all the relevant code is
in the one file referenced.
|Next Message||Tom Lane||2018-09-26 23:45:07||Re: Performance improvements for src/port/snprintf.c|
|Previous Message||Andres Freund||2018-09-26 23:13:31||Re: Let's stop with the retail rebuilds of src/port/ files already|