| From: | Michael Paquier <michael(dot)paquier(at)gmail(dot)com> |
|---|---|
| To: | Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> |
| Cc: | Jeff Davis <pgsql(at)j-davis(dot)com>, PostgreSQL mailing lists <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: 9.5: Better memory accounting, towards memory-bounded HashAgg |
| Date: | 2015-01-16 12:26:40 |
| Message-ID: | CAB7nPqQAj3paSjZO00L4wwWx+rB37nBGhXuQQUoP46-tDck0RA@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Jan 8, 2015 at 4:07 AM, Tomas Vondra
<tomas(dot)vondra(at)2ndquadrant(dot)com> wrote:
> which is IMHO obviously wrong, because that accounts only for the
> hashtable itself.
> In those cases the patch actually does no memory accounting and as
> hashcontext has no child contexts, there's no accounting overhead.
> [blah]
> So the performance drops 2x. With more groups, the performance impact is
> even worse.
Hmm. It seems that this patch is not there yet, so marking it as
returned with feedback, especially knowing that the patch is incorrect
by using hascontext as mentioned by Tomas.
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Michael Paquier | 2015-01-16 12:43:43 | Re: Turning recovery.conf into GUCs |
| Previous Message | Manuel Kniep | 2015-01-16 12:21:28 | Re: segmentation fault in execTuples.c#ExecStoreVirtualTuple |