From: Stefan Beller <sbeller@google.com>
To: Derrick Stolee <dstolee@microsoft.com>
Cc: git <git@vger.kernel.org>, "Jeff King" <peff@peff.net>,
"Jakub Narębski" <jnareb@gmail.com>
Subject: Re: [RFC PATCH 13/13] commit-reach: use can_all_from_reach
Date: Fri, 29 Jun 2018 16:21:06 -0700 [thread overview]
Message-ID: <CAGZ79kbVvto1uV0Rk2UbqLTeL2s21efdcwkEfEPZG+3pD5gwXg@mail.gmail.com> (raw)
In-Reply-To: <20180629161223.229661-14-dstolee@microsoft.com>
Hi Derrick,
On Fri, Jun 29, 2018 at 9:13 AM Derrick Stolee <dstolee@microsoft.com> wrote:
>
> The is_descendant_of method previously used in_merge_bases() to check if
> the commit can reach any of the commits in the provided list. This had
> two performance problems:
>
> 1. The performance is quadratic in worst-case.
>
> 2. A single in_merge_bases() call requires walking beyond the target
> commit in order to find the full set of boundary commits that may be
> merge-bases.
>
> The can_all_from_reach method avoids this quadratic behavior and can
> limit the search beyond the target commits using generation numbers. It
> requires a small prototype adjustment to stop using commit-date as a
> cutoff, as that optimization is no longer appropriate here.
>
> Performance was meausured on a copy of the Linux repository using the
> 'test-tool reach is_descendant_of' command using this input:
>
> A:v4.9
> X:v4.10
> X:v4.11
> X:v4.12
> X:v4.13
> X:v4.14
> X:v4.15
> X:v4.16
> X:v4.17
> X.v3.0
>
> Note that this input is tailored to demonstrate the quadratic nature of
> the previous method, as it will compute merge-bases for v4.9 versus all
> of the later versions before checking against v4.1.
>
> Before: 0.31 s
> After: 0.27 s
>
> Since we previously used the is_descendant_of method in the ref_newer
> method, we also measured performance there using
> 'test-tool reach ref_newer':
>
> Before: 0.12 s
> After: 0.11 s
>
> Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
> ---
>
> One thing I know is missing from this commit is a special-case to use
> the old logic when there is no commit-graph present. The
> can_all_from_reach() algorithm can be worse when we do not have good
> generation number cutoffs. In the previous case of
> can_all_from_reach_with_flags(), we already had an established pattern
> of using commit date as a cutoff, so the generation number is only a
> second cutoff and the algorithm cannot walk more commits than before.
I like this series,
Thanks for writing it!
Stefan
next prev parent reply other threads:[~2018-06-29 23:21 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-06-29 16:12 [RFC PATCH 00/13] Consolidate reachability logic Derrick Stolee
2018-06-29 16:12 ` [RFC PATCH 01/13] commit-reach: move walk methods from commit.c Derrick Stolee
2018-06-29 21:35 ` Stefan Beller
2018-06-29 21:52 ` Junio C Hamano
2018-06-29 16:12 ` [RFC PATCH 02/13] commit-reach: move ref_newer from remote.c Derrick Stolee
2018-06-29 16:12 ` [RFC PATCH 03/13] commit-reach: move commit_contains from ref-filter Derrick Stolee
2018-06-29 21:38 ` Stefan Beller
2018-06-30 1:32 ` Derrick Stolee
2018-06-29 22:00 ` Junio C Hamano
2018-06-29 16:12 ` [RFC PATCH 04/13] upload-pack: make reachable() more generic Derrick Stolee
2018-06-29 22:05 ` Junio C Hamano
2018-06-29 16:12 ` [RFC PATCH 05/13] upload-pack: refactor ok_to_give_up() Derrick Stolee
2018-06-29 21:44 ` Stefan Beller
2018-06-29 16:12 ` [RFC PATCH 06/13] commit-reach: move can_all_from_reach_with_flag() Derrick Stolee
2018-06-29 21:47 ` Stefan Beller
2018-06-30 1:35 ` Derrick Stolee
2018-06-29 16:12 ` [RFC PATCH 07/13] test-reach Derrick Stolee
2018-06-29 21:54 ` Stefan Beller
2018-06-30 1:40 ` Derrick Stolee
2018-06-29 16:12 ` [RFC PATCH 08/13] test-reach: test reduce_heads() Derrick Stolee
2018-06-29 22:06 ` Stefan Beller
2018-06-29 16:12 ` [RFC PATCH 09/13] commit-reach: test can_all_from_reach Derrick Stolee
2018-06-29 16:12 ` [RFC PATCH 10/13] commit-reach: test is_descendant_of Derrick Stolee
2018-06-29 16:13 ` [RFC PATCH 11/13] commit-reach: make can_all_from_reach... linear Derrick Stolee
2018-06-29 23:18 ` Stefan Beller
2018-06-29 16:13 ` [RFC PATCH 12/13] commit-reach: use is_descendant_of for ref_newer Derrick Stolee
2018-06-29 16:13 ` [RFC PATCH 13/13] commit-reach: use can_all_from_reach Derrick Stolee
2018-06-29 23:21 ` Stefan Beller [this message]
2018-06-29 17:33 ` [RFC PATCH 00/13] Consolidate reachability logic Derrick Stolee
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAGZ79kbVvto1uV0Rk2UbqLTeL2s21efdcwkEfEPZG+3pD5gwXg@mail.gmail.com \
--to=sbeller@google.com \
--cc=dstolee@microsoft.com \
--cc=git@vger.kernel.org \
--cc=jnareb@gmail.com \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).