Re: Restructured Shared Buffer Hash Table

From: Dhruv Aron <dhruv(dot)aron(at)gmail(dot)com>
To: Alexandre Felipe <o(dot)alexandre(dot)felipe(at)gmail(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Restructured Shared Buffer Hash Table
Date: 2026-10-13 04:38:11
Message-ID: CAAStW3LzZeO48st2Dyr03Vkej2DMJS5WbEKpLZTNWYK-tovUaQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Oct 12, 2026 at 3:08 PM Alexandre Felipe <
o(dot)alexandre(dot)felipe(at)gmail(dot)com> wrote:

> 1.5 * num_bucket implicitly casts to floating point, and later back to
> integer, that was not meant to be permanent.
> A safer option would be (num_buckets | (num_buckets >> 1)), no wrap-around
> if some day
> this gets increased to UINT_MAX. Or simply num_buckets, without
> multiplication at all.
>

Apologies, I am experiencing trouble finding the line where we perform "1.5
* num_bucket"; perhaps it has been too long since I last worked on the
patch.
Could you point me towards where it is in the
restructured_shared_buffer_table.patch file?
Otherwise, yes, the bitwise arithmetic in this case should also be more
performant.

Best,
Dhruv

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message vignesh C 2026-10-13 04:48:38 Re: Test publication.sql racing with subscription.sql might fail due to cache lookup error
Previous Message Tom Lane 2026-10-13 04:34:35 Re: [PG19] Stale state after the join removal restart gives wrong results