Re: vectorized CRC on ARM64 - Mailing list pgsql-hackers

From John Naylor
Subject Re: vectorized CRC on ARM64
Date
Msg-id CANWCAZZ48GuLYhJCcTy8TXysjrMVJL6n1n7NP94=iG+t80YKPw@mail.gmail.com
Whole thread
In response to Re: vectorized CRC on ARM64  (John Naylor <johncnaylorls@gmail.com>)
Responses Re: vectorized CRC on ARM64
List pgsql-hackers
I wrote:
> In the end, I want to add a length check so
> that inputs smaller than 80 bytes go straight to the scalar path.
> Above 80, after alignment adjustments in the preamble, that still
> guarantees at least one loop iteration in the vector path.

Attached is how that would look. The idea is that small inputs will
encounter fewer branches. It'd be tricky to prove a difference with a
benchmark, and I see this as just making the small-input path more
similar to PG 18, as a risk-avoidance maneuver.

-- 
John Naylor
Amazon Web Services

Attachment

pgsql-hackers by date:

Previous
From: wenhui qiu
Date:
Subject: Re: Clean up remove_rel_from_query() after self-join elimination commit
Next
From: Álvaro Herrera
Date:
Subject: Re: using index to speedup add not null constraints to a table