MIME-Version: 1.0
References: 
 <CAOzEurSW8cNr6TPKsjrstnPfhf4QyQqB4tnPXGGe8N4e_v7Jig@mail.gmail.com>
 <CAN55FZ247JdiT8Sd1SRiyOJxk3Ei=pDCL4kpdP=HqLRjOhKf1Q@mail.gmail.com>
 <CAN55FZ2AxiwSah7TiQoMB==r=JKT0bOtooCB7ov4xRrGkVmJ1A@mail.gmail.com>
 <CAOzEurR5nFt=-SijfU7y0BHVcrT6RG9ovvdVfKt_uBZfEQew9w@mail.gmail.com>
 <CAOzEurSqgA69er9SzhPnXwmsVpO7-piUOuOy3dXcHOi__nSQcg@mail.gmail.com>
 <CA+K2RumC79NwWxBdofHOYo8SCSs0YCJic05Du=xOszRmoPf9FA@mail.gmail.com>
 <CAN55FZ0houfWHn8_MEEefhprZvc33jr07GrBYo+Bp2yw=TVnKA@mail.gmail.com>
 <CA+K2Ru=jHuz_Wpgar4Sobtxeb33qxx=o59ToOhZ=vpmkMqErnA@mail.gmail.com>
In-Reply-To: 
 <CA+K2Ru=jHuz_Wpgar4Sobtxeb33qxx=o59ToOhZ=vpmkMqErnA@mail.gmail.com>
From: Nazir Bilal Yavuz <byavuz81@gmail.com>
Date: Tue, 19 Aug 2025 15:33:38 +0300
Message-ID: 
 <CAN55FZ1J+6eM=F5GreWEBMJcNV_gifYyYY1b6xpYzun=nWPhMQ@mail.gmail.com>
Subject: Re: Speed up COPY FROM text/CSV parsing using SIMD
To: KAZAR Ayoub <ma_kazar@esi.dz>
Cc: Shinya Kato <shinya11.kato@gmail.com>, pgsql-hackers@postgresql.org
Content-Type: multipart/mixed; boundary="000000000000796541063cb711d8"
Archived-At: 
 <https://www.postgresql.org/message-id/CAN55FZ1J%2B6eM%3DF5GreWEBMJcNV_gifYyYY1b6xpYzun%3DnWPhMQ%40mail.gmail.com>
Precedence: bulk

--000000000000796541063cb711d8
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi,

On Thu, 14 Aug 2025 at 18:00, KAZAR Ayoub <ma_kazar@esi.dz> wrote:
>> Thanks for running that benchmark! Would you mind sharing a reproducer
>> for the regression you observed?
>
> Of course, I attached the sql to generate the text and csv test files.
> If having a 1/3 of line length of special characters can be an exaggerati=
on, something lower might still reproduce some regressions of course for th=
e same idea.

Thank you so much!

I am able to reproduce the regression you mentioned but both
regressions are %20 on my end. I found that (by experimenting) SIMD
causes a regression if it advances less than 5 characters.

So, I implemented a small heuristic. It works like that:

- If advance < 5 -> insert a sleep penalty (n cycles).
- Each time advance < 5, n is doubled.
- Each time advance =E2=89=A5 5, n is halved.

I am sharing a POC patch to show heuristic, it can be applied on top
of v1-0001. Heuristic version has the same performance improvements
with the v1-0001 but the regression is %5 instead of %20 compared to
the master.

--
Regards,
Nazir Bilal Yavuz
Microsoft

--000000000000796541063cb711d8
Content-Type: text/plain; charset="UTF-8";
	name="COPY-SIMD-add-heuristic-to-avoid-regression-on-sm.txt"
Content-Disposition: attachment;
	filename="COPY-SIMD-add-heuristic-to-avoid-regression-on-sm.txt"
Content-Transfer-Encoding: base64
Content-ID: <f_meiiya400>
X-Attachment-Id: f_meiiya400
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--000000000000796541063cb711d8--