Re: [PATCH v1] diffcore-rename: speed up register_rename_src

From: Junio C Hamano <gitster@pobox.com>
Date: Apr 19, 2017, 02:45 UTC
Message-ID: <xmqqd1c9cdzi.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20170419013214.q35jarvmk5jhqdyi@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 21 quoted lines

> On Tue, Apr 18, 2017 at 07:44:21PM +0000, git@jeffhostetler.com wrote:
>
>> From: Jeff Hostetler <jeffhost@microsoft.com>
>> 
>> Teach register_rename_src() to see if new file pair
>> can simply be appended to the rename_src[] array before
>> performing the binary search to find the proper insertion
>> point.
>
> I guess your perf results show some minor improvement. But I suspect
> this is because your synthetic repo does not resemble the real world
> very much. You're saving a few strcmps, but for each of those files
> you're potentially going to have actually zlib inflate the object
> contents and do similarity analysis.
>
> So "absurd number of files doing 100% exact renames" is the absolute
> best case, and it saves a few percent.
>
> I dunno. It is not that much code _here_, but I'm not excited about the
> prospect of sprinkling this same "check the last one" optimization all
> over the code base. I wonder if there's some way to generalize it.

When adding many things, we often just append and then sort at the end after we finished adding. I wonder if recent "check the last one and append" optimization beats that strategy.