From: Jeff King <peff@peff•net>
To: Taylor Blau <me@ttaylorr•com>
Cc: git@vger•kernel.org, Junio C Hamano <gitster@pobox•com>,
Elijah Newren <newren@gmail•com>,
Derrick Stolee <stolee@gmail•com>
Subject: Re: [PATCH 3/8] pack-bitmap: reuse stored selected bitmaps
Date: Fri, 29 May 2026 02:00:42 -0400 [thread overview]
Message-ID: <20260529060042.GA1106035@coredump.intra.peff.net> (raw)
In-Reply-To: <ahcCX+xAKFOL8HcW@nand.local>
On Wed, May 27, 2026 at 10:40:31AM -0400, Taylor Blau wrote:
> > > Teach `fill_bitmap_commit()` to notice that case. For non-root commits in
> > > the walk, look for a stored selected bitmap and OR it into the bitmap
> > > being built. If one exists, skip the commit, its tree, and its parents.
> >
> > I feel like this _shouldn't_ be necessary, because the idea of the
> > current writing code is to go from the roots up, following inverted
> > parent pointers, and passing the bitmap up as we go. So whenever we
> > visit a commit we should in theory have all of the ancestor's bits set
> > in that bitmap. But I remember that the simple-and-stupid approach ended
> > up being too memory hungry, so we pick some focal points in the graph
> > and then fill them independently.
>
> It's sharing within the non-first parent history that is killing us
> here. I think what you said is true in a completely linear repository
> with no merges. But since we only pass commit masks from commits to
> their first parents, we don't reuse any already-generated bitmaps for
> common points in history not shared between commits' first parents.
Ah, that makes sense. I had forgotten exactly how the maximal-commit
selection worked, and what we compromised versus the original naive
"build from the bottom up" strategy.
> Yeah, these were for my own curiosity as much as anything. I had written
> them as a temporary measure in order to write the "[...] there are 1,261
> commits selected for bitmap coverage, and 1,382 maximal commits induced
> [...]" portion of the commit message above.
>
> Once I had written it, I found the result useful enough to keep around.
Makes sense. It might help us (or even some very clueful user) debug or
fine-tune parameters down the road.
-Peff
next prev parent reply other threads:[~2026-05-29 6:00 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-05-19 16:12 [PATCH 0/8] pack-bitmap-write: speed up bitmap generation Taylor Blau
2026-05-19 16:12 ` [PATCH 1/8] pack-bitmap: pass object position to `fill_bitmap_tree()` Taylor Blau
2026-05-27 8:57 ` Jeff King
2026-05-27 14:36 ` Taylor Blau
2026-05-19 16:12 ` [PATCH 2/8] pack-bitmap: check subtree bits before recursing Taylor Blau
2026-05-27 9:03 ` Jeff King
2026-05-27 14:36 ` Taylor Blau
2026-05-19 16:12 ` [PATCH 3/8] pack-bitmap: reuse stored selected bitmaps Taylor Blau
2026-05-27 9:24 ` Jeff King
2026-05-27 14:40 ` Taylor Blau
2026-05-29 6:00 ` Jeff King [this message]
2026-05-19 16:12 ` [PATCH 4/8] pack-bitmap: consolidate `find_object_pos()` success path Taylor Blau
2026-05-20 14:42 ` SZEDER Gábor
2026-05-20 17:12 ` Taylor Blau
2026-05-27 9:27 ` Jeff King
2026-05-19 16:12 ` [PATCH 5/8] pack-bitmap: cache object positions during fill Taylor Blau
2026-05-27 9:45 ` Jeff King
2026-05-27 14:46 ` Taylor Blau
2026-05-19 16:12 ` [PATCH 6/8] pack-bitmap: sort bitmaps before XORing Taylor Blau
2026-05-27 10:04 ` Jeff King
2026-05-27 16:56 ` Taylor Blau
2026-05-29 8:26 ` Jeff King
2026-05-19 16:12 ` [PATCH 7/8] pack-bitmap: remember pseudo-merge parents Taylor Blau
2026-05-19 16:12 ` [PATCH 8/8] pack-bitmap: build pseudo-merge bitmaps after regular bitmaps Taylor Blau
2026-05-27 10:25 ` Jeff King
2026-05-27 19:24 ` Taylor Blau
2026-05-29 8:33 ` Jeff King
2026-05-27 10:27 ` [PATCH 0/8] pack-bitmap-write: speed up bitmap generation Jeff King
2026-05-27 19:55 ` [PATCH v2 " Taylor Blau
2026-05-27 19:55 ` [PATCH v2 1/8] pack-bitmap: pass object position to `fill_bitmap_tree()` Taylor Blau
2026-05-27 19:55 ` [PATCH v2 2/8] pack-bitmap: check subtree bits before recursing Taylor Blau
2026-05-27 19:55 ` [PATCH v2 3/8] pack-bitmap: reuse stored selected bitmaps Taylor Blau
2026-05-27 19:55 ` [PATCH v2 4/8] pack-bitmap: consolidate `find_object_pos()` success path Taylor Blau
2026-05-27 19:56 ` [PATCH v2 5/8] pack-bitmap: cache object positions during fill Taylor Blau
2026-05-27 19:56 ` [PATCH v2 6/8] pack-bitmap: sort bitmaps before XORing Taylor Blau
2026-05-27 19:56 ` [PATCH v2 7/8] pack-bitmap: remember pseudo-merge parents Taylor Blau
2026-05-27 19:56 ` [PATCH v2 8/8] pack-bitmap: build pseudo-merge bitmaps after regular bitmaps Taylor Blau
2026-05-29 8:34 ` [PATCH v2 0/8] pack-bitmap-write: speed up bitmap generation Jeff King
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
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260529060042.GA1106035@coredump.intra.peff.net \
--to=peff@peff$(echo .)net \
--cc=git@vger$(echo .)kernel.org \
--cc=gitster@pobox$(echo .)com \
--cc=me@ttaylorr$(echo .)com \
--cc=newren@gmail$(echo .)com \
--cc=stolee@gmail$(echo .)com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox