On Sat, Jul 25, 2020 at 12:41 PM Peter Geoghegan <pg@bowt.ie> wrote:
> I have added a new open item for this separate
> LookupTupleHashEntryHash()/lookup_hash_entry() pipeline-stall issue.
Attached is a rebased version of Andres' now-bitrot 2020-06-12 patch
("aggspeed.diff").
I find that Andres original "SELECT cat, count(*) FROM
fewgroups_many_rows GROUP BY 1;" test case is noticeably improved by
the patch. Without the patch, v13 takes ~11.46 seconds. With the
patch, it takes only ~10.64 seconds.
Didn't test it against v12 yet, but I have no reason to doubt Andres'
explanation. I gather that if we can get this patch committed, we can
close the relevant LookupTupleHashEntryHash() open item.
Can you take this off my hands, Jeff?
Thanks
--
Peter Geoghegan