| From: | Michael Paquier <michael(at)paquier(dot)xyz> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr>, Thomas Munro <thomas(dot)munro(at)gmail(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, PostgreSQL Developers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: refactoring - share str2*int64 functions |
| Date: | 2019-07-18 00:28:28 |
| Message-ID: | 20190718002828.GA1416@paquier.xyz |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Jul 17, 2019 at 11:14:28AM -0700, Andres Freund wrote:
> That'd be considerably slower, so I'm *strongly* against that. These
> conversion routines are *really* hot in a number of workloads,
> e.g. bulk-loading with COPY. Check e.g.
> https://www.postgresql.org/message-id/20171208214437.qgn6zdltyq5hmjpk%40alap3.anarazel.de
Thanks for the link. That makes sense! So stacking more function
calls could also be an issue. Even if using static inline for the
inner wrapper? That may sound like a stupid question but you have
likely more experience than me regarding that with profiling.
> I doubt it - it's not of that long-standing vintage (23a27b039d9,
> 2016-03-12), and if so they are very likely to use base 10. We shouldn't
> keep some barely tested function around, just for the hypothetical
> scenario that some extension uses it. Especially if that function is
> considerably slower than the potential replacement.
Okay, I won't fight hard on that either.
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Michael Paquier | 2019-07-18 00:45:14 | Re: Add parallelism and glibc dependent only options to reindexdb |
| Previous Message | David Rowley | 2019-07-17 23:57:44 | Re: Custom table AMs need to include heapam.h because of BulkInsertState |