Re: New CRC algorithm: Slicing by 8 - Mailing list pgsql-hackers

From Gregory Maxwell
Subject Re: New CRC algorithm: Slicing by 8
Date
Msg-id e692861c0610211922g7216d74fr21084095942a6251@mail.gmail.com
Whole thread Raw
In response to Re: New CRC algorithm: Slicing by 8  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: New CRC algorithm: Slicing by 8  ("Gurjeet Singh" <singh.gurjeet@gmail.com>)
List pgsql-hackers
On 10/21/06, Tom Lane <tgl@sss.pgh.pa.us> wrote:
[snip]
> It hasn't even been tested.  One thing I'd want to know about is the
> performance effect on non-Intel machines.

On Opteron 265 his test code shows SB8 (the intel alg) is 2.48x faster
for checksum and 1.95x faster for verify for the 800 * 1024 blocks of
1 KB each workload. For 100000 blocks of 8k I got simmlar results as
well.

It looks like the new code may have a larger cache footprint, so
actual performance may differ from the microbenchmark.


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: New CRC algorithm: Slicing by 8
Next
From: Volkan YAZICI
Date:
Subject: Re: estimated_count() implementation