From: Marco Costalba <mcostalba@gmail.com>
Date: Sun, 30 Dec 2007 19:06:25 GMT
Subject: Re: [PATCH WIP] sha1-lookup: make selection of 'middle' less aggressive
Message-ID: <e5bfff550712301106l133dd38btd2cc4be02159387d@mail.gmail.com>
In-Reply-To: <7vtzm08l9w.fsf@gitster.siamese.dyndns.org>

On Dec 30, 2007 12:38 PM, Junio C Hamano <gitster@pobox.com> wrote:
>
> With this patch, we actually see slight improvements in
> execution time as well.

I have applied both your patch above and what we have discussed today,
the patch to speedup prefixcmp() and mine experimental one to avoid a
lookup loop in strbuf_expand().

I don't see big differences with or without you patch applied.

Just for document the profiling I have uploaded a snapshot of
KCachegrind profiling data on a run of git-log on the git tree:

http://digilander.libero.it/mcostalba/callgrind_git_log1.png

>From there you can see that pretty.c and strbuf.c, after all the
optimizations, account for less then 8% of total time.
The biggest part is that 86.64% that is due almost entirely to zlib.

In particular

st = inflate(&stream, Z_FINISH);

called from unpack_compressed_entry() in sha1_file.c accounts for 72%
of total time.


Marco