| From: | Neil Conway <neilc(at)samurai(dot)com> |
|---|---|
| To: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
| Cc: | Qingqing Zhou <zhouqq(at)cs(dot)toronto(dot)edu>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Which qsort is used |
| Date: | 2005-12-12 17:15:46 |
| Message-ID: | 1134407746.9179.13.camel@localhost.localdomain |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, 2005-12-12 at 11:50 -0500, Bruce Momjian wrote:
> Are you willing to say that we should always prefer pgport over glibc's
> qsort()?
glibc's qsort is actually implemented via merge sort. I'm not sure why
the glibc folks chose to do that, but as a result, it's not surprising
that BSD qsort beats it for typical inputs. Whether we should go to the
trouble of second-guessing glibc is a separate question, though: it
would be good to see some performance figures for real-world queries.
BTW, Luke Lonergan recently posted some performance results for a fairly
efficient public domain implementation of qsort to the bizgres list:
http://lists.pgfoundry.org/pipermail/bizgres-general/2005-December/000294.html
-Neil
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Volkan YAZICI | 2005-12-12 17:31:05 | number of loaded/unloaded COPY rows |
| Previous Message | Andreas Pflug | 2005-12-12 17:11:44 | Re: pg_relation_size locking |