From: Junio C Hamano <gitster@pobox.com>
Date: Sun, 01 Mar 2009 07:31:55 GMT
Subject: Re: [PATCH 3/4] diffcore-pickaxe: further refactor count_match()
Message-ID: <7viqmtaec4.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <49A937B8.1030205@lsrfire.ath.cx>

René Scharfe <rene.scharfe@lsrfire.ath.cx> writes:

> I get this (Ubuntu 8.10 x64, Fedora 10 x64 using the same Linux repo,
> Windows Vista x64 using a different Linux repo with the same HEAD on
> NTFS and msysgit, numbers are the elapsed time in seconds, best of five
> runs):
>
>                            Ubuntu  Fedora  Windows
>    v1.6.2-rc2                8.14    8.16    9.236
>    v1.6.2-rc2+[1-4]          2.43    2.45    2.995
>    v1.6.2-rc2+[1-4]+memmem   1.31    1.25    2.917
>    v1.6.2-rc2+[1-3]+memmem   1.51    1.16    8.455
>
> Ubuntu has glibc 2.8, while Fedora 10 has glibc 2.9, with a new and more
> efficient memmem() implementation.  On Windows, we use our own naive
> memmem() implementation.

Shoot, I was not being careful enough.

> So using memmem() is worthwhile.  And providing a better fall-back
> version in compat/ can speed up this particular case to the point where
> the fourth patch becomes moot.

Are these numbers telling me that with a good memmem() implementation,
patch 4/4 is not just moot but actively detrimental?

With a long enough needle, it entirely is possible that scanning the whole
image with sublinear string search algorithm would perform much better
than the preprocessing patch 4/4 does which has to scan all the bytes in
the common parts.