Re: 16-bit page checksums for 9.2

Nicolas Barbier <nicolas.barbier@gmail.com>

From: Nicolas Barbier <nicolas.barbier@gmail.com>

To: Simon Riggs <simon@2ndquadrant.com>

Cc: Andres Freund <andres@anarazel.de>, pgsql-hackers@postgresql.org, Kevin Grittner <Kevin.Grittner@wicourts.gov>, aidan@highrise.ca, stark@mit.edu

Date: 2012-01-04T11:53:04Z

Lists: pgsql-hackers

2012/1/4 Simon Riggs <simon@2ndquadrant.com>:

> On Wed, Jan 4, 2012 at 9:20 AM, Andres Freund <andres@anarazel.de> wrote:
>
>> I wonder if CRC32c wouldn't be a good alternative given more and more cpus
>> (its in SSE 4.2) support calculating it in silicon.
>
> We're trying to get something that fits in 16bits for this release.
> I'm guessing CRC32c doesn't?

What happens to the problem-detecting performance of a 16 bit part of
a CRC32c vs. a real 16 bit checksum? If it is still as good, it might
make sense to use the former, assuming that there is a way to easily
trigger the silicon support and enough CPUs support it.

Nicolas

-- 
A. Because it breaks the logical sequence of discussion.
Q. Why is top posting bad?