Re: Improve the performance of Unicode Normalization Forms.

Alexander Borisov <lex.borisov@gmail.com>

From: Alexander Borisov <lex.borisov@gmail.com>

To: Jeff Davis <pgsql@j-davis.com>, Victor Yegorov <vyegorov@gmail.com>, Heikki Linnakangas <hlinnaka@iki.fi>

Cc: PostgreSQL Hackers <pgsql-hackers@postgresql.org>

Date: 2025-09-29T10:22:29Z

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 →

Use C11 char16_t and char32_t for Unicode code points.
- 3853a6956c3e 19 (unreleased) cited

Attachments

v5-0001-Moving-Perl-functions-Sparse-Array-to-a-common-mo.patch (text/plain) patch v5-0001
v5-0002-Improve-the-performance-of-Unicode-Normalization-.patch (text/plain) patch v5-0002
v5-0003-Refactoring-Unicode-Normalization-Forms-performan.patch (text/plain) patch v5-0003

20.09.2025 04:03, Jeff Davis wrote:
> On Thu, 2025-09-11 at 20:51 +0300, Alexander Borisov wrote:
>>
>>> Hey.
>>>
>>> I've looked into these patches.
>>
>> Hi Victor,
>>
>> Thank you for reviewing the patch and testing it!
> 
> Heikki, do you have thoughts on this thread?

Hey,

In patch v5 (attached), I changed the approach to class caching.
Now, for small texts (less than 512 characters), we don't allocate
memory from the heap; we use the stack.

And according to pgbench tests, we have a 2x speedup.
According to Jeff's tests, we have a 10x speedup.
According to tests from the thread
https://www.postgresql.org/message-id/CAFBsxsHUuMFCt6-pU+oG-F1==CmEp8wR+O+bRouXWu6i8kXuqA@mail.gmail.com,
we also have a 2x speedup.

--
Regards,
Alexander Borisov