git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH v3 1/4] connected: do not sort input revisions

From
Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date
Aug 4, 2021, 11:01 UTC
Message-ID
<87r1f9mq1j.fsf@evledraar.gmail.com>
In-Reply-To
<YQkDQY9Qyqi7FCg6@ncase>
On Tue, Aug 03 2021, Patrick Steinhardt wrote:
Show 34 quoted lines
> [[PGP Signed Part:Undecided]]
> On Mon, Aug 02, 2021 at 02:49:29PM +0200, Ævar Arnfjörð Bjarmason wrote:
>> On Mon, Aug 02 2021, Patrick Steinhardt wrote:
> [snip]
>> > Introduce a new "--unsorted-input" flag to git-rev-list(1) which will
>> > cause it to not sort the commits and adjust the connectivity check to
>> > always pass the flag. This results in the following speedups, executed
>> > in a clone of gitlab-org/gitlab [1]:
>> >
>> >     Benchmark #1: git rev-list  --objects --quiet --not --all --not $(cat newrev)
>> >       Time (mean ± σ):      7.639 s ±  0.065 s    [User: 7.304 s, System: 0.335 s]
>> >       Range (min … max):    7.543 s …  7.742 s    10 runs
>> >
>> >     Benchmark #2: git rev-list --unsorted-input --objects --quiet --not --all --not $newrev
>> >       Time (mean ± σ):      4.995 s ±  0.044 s    [User: 4.657 s, System: 0.337 s]
>> >       Range (min … max):    4.909 s …  5.048 s    10 runs
>> >
>> >     Summary
>> >       'git rev-list --unsorted-input --objects --quiet --not --all --not $(cat newrev)' ran
>> >         1.53 ± 0.02 times faster than 'git rev-list  --objects --quiet --not --all --not $newrev'
>> 
>> Just bikeshedding for a potential re-roll, perhaps --unordered-input, so
>> that it matches/rhymes with the existing "git cat-file --unordered",
>> which serves the same conceptual purpose (except this one's input, that
>> one's output).
>
> Yeah, I wasn't quite sure how to name it myself either. Internally, we
> typically use "unsorted" instead of "unordered", and there's also the
> preexisting "--no-walk=(sorted|unsorted)" flag for git-rev-list(1). With
> the latter in mind, I think that "unsorted" fits a bit better given that
> we already use it in the same command, but I don't particularly mind.
>
> For now, I'll keep "--unsorted-input", but please feel free to give me
> another shout if you disagree with my reasoning.

Sounds good. I didn't mean it as a strong suggestion, unsorted does indeed sounds better in this context, just a gentle poke to check if you knew about that similar option...

Thanks.
Previous: Patrick SteinhardtNext: Junio C Hamano
Message 27 of 64 in “Speed up connectivity checks via bitmaps”
  1. 0/3 Speed up connectivity checks via bitmapsPatrick Steinhardt, Jun 28, 2021
  2. 1/3 p5400: add perf tests for git-receive-pack(1)Patrick Steinhardt, Jun 28, 2021
  3. Ævar Arnfjörð BjarmasonJun 28, 2021
  4. Patrick SteinhardtJun 29, 2021
  5. Ævar Arnfjörð BjarmasonJun 29, 2021
  6. 2/3 receive-pack: skip connectivity checks on delete-only commandsPatrick Steinhardt, Jun 28, 2021
  7. Ævar Arnfjörð BjarmasonJun 28, 2021
  8. Ævar Arnfjörð BjarmasonJun 28, 2021
  9. Patrick SteinhardtJun 29, 2021
  10. Jeff KingJun 30, 2021
  11. Jeff KingJun 30, 2021
  12. Patrick SteinhardtJun 30, 2021
  13. 3/3 connected: implement connectivity check using bitmapsPatrick Steinhardt, Jun 28, 2021
  14. Taylor BlauJun 28, 2021
  15. Taylor BlauJun 29, 2021
  16. Jeff KingJun 30, 2021
  17. Taylor BlauJun 30, 2021
  18. Jeff KingJun 30, 2021
  19. Taylor BlauJul 2, 2021
  20. Jeff KingJul 2, 2021
  21. Jeff KingJun 30, 2021
  22. Patrick SteinhardtJul 20, 2021
  23. 0/4 Speed up connectivity checksPatrick Steinhardt, Aug 2, 2021
  24. 1/4 connected: do not sort input revisionsPatrick Steinhardt, Aug 2, 2021
  25. Ævar Arnfjörð BjarmasonAug 2, 2021
  26. Patrick SteinhardtAug 3, 2021
  27. Ævar Arnfjörð BjarmasonAug 4, 2021
  28. Junio C HamanoAug 2, 2021
  29. Patrick SteinhardtAug 3, 2021
  30. Junio C HamanoAug 3, 2021
  31. 2/4 revision: stop retrieving reference twicePatrick Steinhardt, Aug 2, 2021
  32. Ævar Arnfjörð BjarmasonAug 2, 2021
  33. 3/4 revision: avoid loading object headers multiple timesPatrick Steinhardt, Aug 2, 2021
  34. Ævar Arnfjörð BjarmasonAug 2, 2021
  35. Patrick SteinhardtAug 5, 2021
  36. Junio C HamanoAug 2, 2021
  37. Patrick SteinhardtAug 3, 2021
  38. Patrick SteinhardtAug 6, 2021
  39. 4/4 revision: avoid hitting packfiles when commits are in commit-graphPatrick Steinhardt, Aug 2, 2021
  40. Junio C HamanoAug 2, 2021
  41. Patrick SteinhardtAug 3, 2021
  42. Junio C HamanoAug 3, 2021
  43. Patrick SteinhardtAug 5, 2021
  44. Junio C HamanoAug 5, 2021
  45. Ævar Arnfjörð BjarmasonAug 4, 2021
  46. 0/6 Speed up connectivity checksPatrick Steinhardt, Aug 5, 2021
  47. 1/6 revision: separate walk and unsorted flagsPatrick Steinhardt, Aug 5, 2021
  48. Junio C HamanoAug 5, 2021
  49. 2/6 connected: do not sort input revisionsPatrick Steinhardt, Aug 5, 2021
  50. Junio C HamanoAug 5, 2021
  51. Patrick SteinhardtAug 6, 2021
  52. Junio C HamanoAug 6, 2021
  53. 3/6 revision: stop retrieving reference twicePatrick Steinhardt, Aug 5, 2021
  54. 4/6 revision: avoid loading object headers multiple timesPatrick Steinhardt, Aug 5, 2021
  55. 5/6 commit-graph: split out function to search commit positionPatrick Steinhardt, Aug 5, 2021
  56. 6/6 revision: avoid hitting packfiles when commits are in commit-graphPatrick Steinhardt, Aug 5, 2021
  57. 0/5 Speed up connectivity checksPatrick Steinhardt, Aug 9, 2021
  58. Patrick SteinhardtAug 9, 2021
  59. 0/5 Speed up connectivity checksPatrick Steinhardt, Aug 9, 2021
  60. 1/5 revision: separate walk and unsorted flagsPatrick Steinhardt, Aug 9, 2021
  61. 2/5 connected: do not sort input revisionsPatrick Steinhardt, Aug 9, 2021
  62. 3/5 revision: stop retrieving reference twicePatrick Steinhardt, Aug 9, 2021
  63. 4/5 commit-graph: split out function to search commit positionPatrick Steinhardt, Aug 9, 2021
  64. 5/5 revision: avoid hitting packfiles when commits are in commit-graphPatrick Steinhardt, Aug 9, 2021

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.