Horton

Re: Memory usage during sorting

Peter Geoghegan <peter@2ndquadrant.com>

View thread
From: Peter Geoghegan <peter@2ndquadrant.com>
To: Jeff Janes <jeff.janes@gmail.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2012-01-22T00:51:58Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Improve performance of our private version of qsort. Per recent testing,

  2. Further performance improvements in sorting: reduce number of comparisons

On 16 January 2012 00:59, Jeff Janes <jeff.janes@gmail.com> wrote:
> I think it would be better to pre-deduct the tape overhead amount we
> will need if we decide to switch to tape sort from the availMem before
> we even start reading (and then add it back if we do indeed make that
> switch).  That way we wouldn't over-run the memory in the first place.
>  However, that would cause apparent regressions in which sorts that
> previously fit into maintenance_work_mem no longer do.  Boosting
> maintenance_work_mem to a level that was actually being used
> previously would fix those regressions, but pointing out that the
> previous behavior was not optimal doesn't change the fact that people
> are used to it and perhaps tuned to it.  So the attached patch seems
> more backwards-friendly.

Hmm. Are people really setting maintenance_work_mem such that it is
exactly large enough to quicksort when building an index in one case
but not another? Is the difference large enough to warrant avoiding
pre-deduction?

-- 
Peter Geoghegan       http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training and Services