RE: Big Performance drop of Exceptions in UDFs between V11.2 and 13.4

-----Message d'origine-----
De : ldh(at)laurent-hasson(dot)com <ldh(at)laurent-hasson(dot)com>
Envoyé : mardi 31 août 2021 04:18
À : ldh(at)laurent-hasson(dot)com; Julien Rouhaud <rjuju123(at)gmail(dot)com>
Cc : Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>; Ranier Vilela <ranier(dot)vf(at)gmail(dot)com>; Andrew Dunstan <andrew(at)dunslane(dot)net>; Justin Pryzby <pryzby(at)telsasoft(dot)com>; pgsql-performance(at)postgresql(dot)org
Objet : RE: Big Performance drop of Exceptions in UDFs between V11.2 and 13.4
Importance : Haute

> I see on https://www.postgresql.org/download/ that there is a different
> installer from 2ndQuadrant. I am going to try that one and see what I
> come up with. Are there any other "standard" distros of Postgres that I
> could try out?
>
> I found out I could download Visual Studio community edition so I am
> trying this, but may not have the time to get through a build any time
> soon as per my unfamiliarity with the process. I'll follow Ranier's steps
> and see if that gets me somewhere.
>
> Thank you,
> Laurent.

Hello all,

I think I had a breakthrough. I tried to create a local build and wasn't able to. But I downloaded the 2nd Quadrant installer and the issue disappeared!!! I think this is proof that it's not my personal environment, nor something intrinsic in the codebase, but definitely something in the standard EDB installer.

create table sampletest (a varchar, b varchar); insert into sampletest (a, b) select substr(md5(random()::text), 0, 15), (100000000*random())::integer::varchar
from generate_series(1,100000);

CREATE OR REPLACE FUNCTION toFloat(str varchar, val real) RETURNS real AS $$ BEGIN
RETURN case when str is null then val else str::real end; EXCEPTION WHEN OTHERS THEN
RETURN val;
END;
$$ LANGUAGE plpgsql COST 1 IMMUTABLE;

explain (analyze,buffers,COSTS,TIMING) select MAX(toFloat(a, null)) as "a" from sampletest; --Aggregate (cost=2137.00..2137.01 rows=1 width=4) (actual time=2092.922..2092.923 rows=1 loops=1)
-- Buffers: shared hit=637
-- -> Seq Scan on sampletest (cost=0.00..1637.00 rows=100000 width=15) (actual time=0.028..23.925 rows=100000 loops=1)
-- Buffers: shared hit=637
--Planning Time: 0.168 ms
--Execution Time: 2092.957 ms

explain (analyze,buffers,COSTS,TIMING) select MAX(toFloat(b, null)) as "b" from sampletest; --Aggregate (cost=2137.00..2137.01 rows=1 width=4) (actual time=369.475..369.476 rows=1 loops=1)
-- Buffers: shared hit=637
-- -> Seq Scan on sampletest (cost=0.00..1637.00 rows=100000 width=8) (actual time=0.020..18.746 rows=100000 loops=1)
-- Buffers: shared hit=637
--Planning Time: 0.129 ms
--Execution Time: 369.507 ms

Thank you,
Laurent!

_________________________________________________________
Hi,

Something which has nothing with the thread but I think it must be said :-)
Why substring(x, 0, ...)?
msym=> select substr('abcde', 0, 3), substr('abcde', 1, 3);
substr | substr
--------+--------
ab | abc

Michel SALAIS