Re: Tuple sampling - Mailing list pgsql-patches

From Bruno Wolff III
Subject Re: Tuple sampling
Date
Msg-id 20040524035743.GA337@wolff.to
Whole thread Raw
In response to Re: Tuple sampling  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-patches
On Sun, May 23, 2004 at 17:32:36 -0400,
  Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> slower.  (The datatype conversions induced inside get_next_S are likely
> to outweigh any savings from counting by ints, on most modern hardware.)

You aren't even guarenteed that integer operations are faster. On the Cray
XMP integer math was often done using floating point operations because
they were faster. This limited them to 48 bits instead of instead of 64,
but normally 48 bits was enough. I don't remember if just multiplication
was faster or if all operations were faster.

pgsql-patches by date:

Previous
From: Neil Conway
Date:
Subject: Re: WIP list rewrite
Next
From: Tom Lane
Date:
Subject: Re: WIP list rewrite