Re: benchmark results comparing versions 15.2 and 16

Hi Mark,

On Tue, May 30, 2023 at 1:03 PM MARK CALLAGHAN <mdcallag(at)gmail(dot)com> wrote:
> Do you want me to try PG 16 without ICU or PG 15 with ICU?
> I can do that, but it will take a few days before the server is available.

Sorry for the late reply. Most of the Postgres developers (myself
included) are attending pgCon right now.

It would be nice to ascertain just how much of a boost we're getting
from our use of ICU as far as sysbench goes. I'd appreciate having
that information. We discussed the choice of ICU as default collation
provider at yesterday's developer meeting:

https://wiki.postgresql.org/wiki/PgCon_2023_Developer_Meeting#High_level_thoughts_and_feedback_on_moving_toward_ICU_as_the_preferred_collation_provider
https://wiki.postgresql.org/wiki/StateOfICU

Just confirming my theory about abbreviated keys (without rerunning
the benchmark) should be simple - perhaps that would be a useful place
to start. You could just rerun the two individual queries of interest
from an interactive psql session. There are even low-level debug
messages available through the trace_sort GUC. From a psql session
you'd run something along the lines of "set client_min_messages=log;
set trace_sort=on; $QUERY". You'll see lots of LOG messages with
specific information about the use of abbreviated keys and the
progress of each sort.

Thanks
--
Peter Geoghegan