Skip to content

Ocarina of Time: prewarm list rows from 1 h 40 min of play (Hyrule Field, Lake Hylia, Castle Town) - #2531

Open
NicoPV94 wants to merge 2 commits into
seedds:mainfrom
NicoPV94:ocarina-field-prewarm
Open

NicoPV94 wants to merge 2 commits into
seedds:mainfrom
NicoPV94:ocarina-field-prewarm

Conversation

@NicoPV94

Copy link
Copy Markdown

This builds on #2518: its first commit is that PR's (the Deku Tree's 223 rows). Merging this one merges both.

On a Mac (M4 Max, 36 GB, macOS 27) running a build of main at f4b7e77 (chain 3856a9225d68) on v131072, a session of about 1 h 40 min built about 1,000 pipelines at their first draws:

  • 274 frames of 50 ms or more, 125 s of stalls in all;
  • worst at about 35-55 and 70-80 minutes into play, in areas new to the lists and at their loads.

The session played Hyrule Field, Lake Hylia and Castle Town, among other places. Its lists were the bundled pso_manifest.tsv and the 223 rows of #2518.

This appends the 11,061 rows that device's pso_user.tsv gained in that session to pso_manifest.tsv. None of them were already in the list:

Rows Count
S 6,475
R 4,002
C 452
M 132

The header gets a comment line for them, and the game note's pso_manifest.tsv paragraph says where they came from.

Checks:

  • Every shader key in the new rows is in shader_map.tsv.
  • test_docs.py and test_game_profile.py pass.
  • No host, tool or chain change, so no regress.py run.

Open:

  • Not measured yet with these rows bundled.
  • They lengthen a cold first start's prepare step and, since 0a126ac, the Mac app's prebuild of the prewarm lists. Neither has been measured with them.

🤖 Generated with Claude Code

seedds and others added 2 commits November 12, 2026 15:15
… in play into the Great Deku Tree: with the list before them its entry scene built 5 model task / mesh pipelines behind a5288d0a6c4971fd_b (128-141 ms each) and the render pipelines around them at their first draws, six frames of 110-482 ms within 5 s

33 S, 137 R, 48 C and 5 M rows, the device's pso_user.tsv after about 11 minutes of play on v131072 (chain 3856a9225d68), none of them in the list before; every shader key is in shader_map.tsv. Not measured yet with these rows bundled.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ist gained over 1 h 40 min of later play (Hyrule Field, Lake Hylia and Castle Town among it): that play built about 1,000 pipelines at their first draws, 274 frames of 50 ms or more and 125 s of stalls in all, worst in new areas and at their loads

6,475 S, 4,002 R, 452 C and 132 M rows from the device's pso_user.tsv on v131072 (chain 3856a9225d68), none of them in the list before; every shader key is in shader_map.tsv. Not measured yet with these rows bundled; they lengthen a cold first start's prepare step and the Mac app's prebuild.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions github-actions Bot added docs Documentation game: 0400486024DEC000 A game labels Nov 12, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

docs Documentation game: 0400486024DEC000 A game

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants