docs(tested): the cloth gate with every lifted chunk at -O2, against -O1, six interleaved pairs (#1485) #1630

Open
enum wants to merge 1 commit from enum/Longshot:tested/cloth-o2-runs into main
Contributor

Type: docs · Risk: low · Tested on: Windows 11, RTX 4070 · Needs: none

TL;DR

Twelve measured runs for the tested configurations catalogue: the game with all of its translated code compiled at -O2, against the usual -O1, on an RTX 4070 with a Ryzen 9 7900X on Windows. In six interleaved timed pairs the -O2 build was 10.3% faster in the Market and 6.9% faster in Hyrule Field, and the same within the spread at the Zora's Domain arrival. Records only: no code changes.

Breakdown

  • Who it helps: whoever decides whether players' builds should compile the game at -O2, and readers of the tested configurations catalogue.
  • What was wrong: the team's performance notes had -O2/-O3 numbers from Linux only, where they showed no fps change.
  • What changes: twelve files in docs/tested/, one a run.
  • What does not change: everything else; no code, no default.
  • How it was checked: tests/test_tested_configurations.py (the record format).
  • What is still open: nothing for this PR; the analysis is on the wiki's Performance: what to fix next, item 4.
  • Risk: low: records only.

Why

Item 4 of the wiki's performance page proposes the cloth and physics code at -O2, as an estimate; its "ruled out" list has "all game code at -O3 on its own: no fps" from a Linux RTX 3090. These runs measure the Windows side for #1485 (Zora's Domain slows down near King Zora). See #1485, #1493.

What changed

Twelve run records in the tested configurations' format, one file a run, six for each side. Nothing else changes.

  • docs: docs/tested/2026-11-18-cloth60-o2-enum-a1.md to -a6.md (the -O1 side: what players run) and -b1.md to -b6.md (every lifted chunk at -O2, LONGSHOT_CHUNK_OPT=all, with HOS_FPS_CLOTH_60=1 from #1628).

The build the records name, perf/cloth-above-60 @ 8aadf550, is #1628's code before its documentation and switch-table row were added (#1628's 0663d20f has the same code). The gate's own runs (-O1, the gate on against off) are on #1628 and show no fps change, so the difference here is the compiler setting.

Measurements

Verdict: all lifted chunks at -O2: +10.3% in the Market and +6.9% in Hyrule Field (5 of 6 runs above the best -O1 run at both), within the spread at the Zora's Domain arrival.

Hardware: Windows 11 Pro (26300), RTX 4070 12 GB (NVIDIA 616.56), Ryzen 9 7900X, 64 GB
Scene: teleport arrivals: Market (child), Hyrule Field (adult), Zora's Domain (child); output 2560x1440 in a 1280x720 window, Quality preset Auto (tier 4, the game's own upscaler), frame generation off, cap 240
Method: one game session a run in the job queue's timing mode, each spot settled 12 s and measured over 20 s by the game's frame counter, hands off; interleaved A1 B1 ... A6 B6 after a warm-up of each; the median of six. Runs 4 to 6 record the CPU busy outside the game (1.0 to 1.8 cores, both sides).
How to reproduce: dev_packs.py pack of #1628's branch twice over one base, the second with LONGSHOT_CHUNK_OPT=all; the owner's save; the control bridge's teleport to the three destinations.
Baseline: the before column is the same branch built as players build it (-O1, HOS_FPS_CLOTH_60 unset); main has no -O2 build to compare.

fps Before: -O1, gate off @ 8aadf550 After: all -O2, gate on @ 8aadf550
Market (child) 67.9, 70.7, 69.7, 68.4, 70.4, 70.0 (median 69.8) 75.8, 69.9, 74.5, 79.1, 78.5, 78.2 (median 77.0)
Hyrule Field (adult) 93.7, 93.3, 94.9, 97.7, 95.2, 97.7 (median 95.0) 101.1, 94.6, 98.9, 103.1, 102.8, 102.2 (median 101.6)
Zora's Domain (child) 89.4, 88.7, 91.9, 91.7, 91.3, 92.3 (median 91.5) 94.5, 89.9, 91.2, 95.6, 95.0, 92.0 (median 93.3)
build: LONGSHOT_CHUNK_OPT=all: every lifted chunk (93) at -O2
build: LONGSHOT_CHUNK_OPT: chunks cut into parts of about 6 MB (338 part files)
the build took 0:07:54, compiled in 0:05:52

Flags and settings

None.

Testing

Tested: python -I -B tests/test_tested_configurations.py: the record checks pass; its one failure (test_cli_ascii_pipe_emits_the_arrow_containing_source_as_utf8, a Windows line-ending case) fails on main too. The runs themselves are what the records describe.

Not tested: other PCs; the King Zora stairs as timed runs (the bridge can't walk there; hand readings are on the wiki, marked as such).

Must be tested:

  • None

Risks

What could break: nothing; records only.

Turning it off: revert the commit.

Compatibility: Nothing changes.

Issues

See #1485

Checklist

Everyone

  • One topic.
  • The title is a Conventional Commit (type(scope): description) of the status line's type.
  • Why gives the reason plainly and links the issues that fit (the tracker searched).
  • Tested on my own PC where it runs (in the game, for a change the game runs), as Testing's Tested says; what
    can't be run by hand is in Testing's Not testable by hand, with why.
  • Measurements has a verdict, or one line saying why there is nothing to measure.
  • Nothing gets worse that I know of: any cost (speed, latency, memory, a setup that worked before) is measured and
    in Risks.
  • Testing names what was tested, what wasn't, and what must still be tested, where and how.
  • Rebased on current main.
  • python -I -B tests/run_all.py passes (any failure that also fails on main is named). (Docs only: the check that reads these files, tests/test_tested_configurations.py, was run; its one failure also fails on main.)
  • No game files, dumps, build output or generated files.
  • third_party/ untouched (a new folder with a PROVENANCE.md instead).
  • The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled.

When it applies

  • AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
  • Numbers from the same PC, scene and settings, at least 2 interleaved runs each, the median reported.
  • A file in docs/tested/ for each new measured run (format in docs/tested-configurations.md).

🤖 Generated with Claude Code

**Type:** docs · **Risk:** low · **Tested on:** Windows 11, RTX 4070 · **Needs:** none ## TL;DR Twelve measured runs for the tested configurations catalogue: the game with all of its translated code compiled at `-O2`, against the usual `-O1`, on an RTX 4070 with a Ryzen 9 7900X on Windows. In six interleaved timed pairs the `-O2` build was 10.3% faster in the Market and 6.9% faster in Hyrule Field, and the same within the spread at the Zora's Domain arrival. Records only: no code changes. ## Breakdown - **Who it helps:** whoever decides whether players' builds should compile the game at `-O2`, and readers of the tested configurations catalogue. - **What was wrong:** the team's performance notes had `-O2`/`-O3` numbers from Linux only, where they showed no fps change. - **What changes:** twelve files in `docs/tested/`, one a run. - **What does not change:** everything else; no code, no default. - **How it was checked:** `tests/test_tested_configurations.py` (the record format). - **What is still open:** nothing for this PR; the analysis is on the wiki's [Performance: what to fix next](https://git.magicbean.dev/MagicBeanDev/Longshot/wiki/Performance%3A-what-to-fix-next), item 4. - **Risk:** low: records only. ## Why Item 4 of the wiki's performance page proposes the cloth and physics code at `-O2`, as an estimate; its "ruled out" list has "all game code at `-O3` on its own: no fps" from a Linux RTX 3090. These runs measure the Windows side for #1485 (Zora's Domain slows down near King Zora). See #1485, #1493. ## What changed Twelve run records in the tested configurations' format, one file a run, six for each side. Nothing else changes. - docs: `docs/tested/2026-11-18-cloth60-o2-enum-a1.md` to `-a6.md` (the `-O1` side: what players run) and `-b1.md` to `-b6.md` (every lifted chunk at `-O2`, `LONGSHOT_CHUNK_OPT=all`, with `HOS_FPS_CLOTH_60=1` from #1628). The build the records name, `perf/cloth-above-60 @ 8aadf550`, is #1628's code before its documentation and switch-table row were added (#1628's `0663d20f` has the same code). The gate's own runs (`-O1`, the gate on against off) are on #1628 and show no fps change, so the difference here is the compiler setting. ## Measurements **Verdict:** all lifted chunks at `-O2`: +10.3% in the Market and +6.9% in Hyrule Field (5 of 6 runs above the best `-O1` run at both), within the spread at the Zora's Domain arrival. Hardware: Windows 11 Pro (26300), RTX 4070 12 GB (NVIDIA 616.56), Ryzen 9 7900X, 64 GB Scene: teleport arrivals: Market (child), Hyrule Field (adult), Zora's Domain (child); output 2560x1440 in a 1280x720 window, Quality preset Auto (tier 4, the game's own upscaler), frame generation off, cap 240 Method: one game session a run in the job queue's timing mode, each spot settled 12 s and measured over 20 s by the game's frame counter, hands off; interleaved A1 B1 ... A6 B6 after a warm-up of each; the median of six. Runs 4 to 6 record the CPU busy outside the game (1.0 to 1.8 cores, both sides). How to reproduce: `dev_packs.py pack` of #1628's branch twice over one base, the second with `LONGSHOT_CHUNK_OPT=all`; the owner's save; the control bridge's `teleport` to the three destinations. Baseline: the before column is the same branch built as players build it (`-O1`, `HOS_FPS_CLOTH_60` unset); main has no `-O2` build to compare. | fps | Before: -O1, gate off @ 8aadf550 | After: all -O2, gate on @ 8aadf550 | |---|---|---| | Market (child) | 67.9, 70.7, 69.7, 68.4, 70.4, 70.0 (median 69.8) | 75.8, 69.9, 74.5, 79.1, 78.5, 78.2 (median 77.0) | | Hyrule Field (adult) | 93.7, 93.3, 94.9, 97.7, 95.2, 97.7 (median 95.0) | 101.1, 94.6, 98.9, 103.1, 102.8, 102.2 (median 101.6) | | Zora's Domain (child) | 89.4, 88.7, 91.9, 91.7, 91.3, 92.3 (median 91.5) | 94.5, 89.9, 91.2, 95.6, 95.0, 92.0 (median 93.3) | ``` build: LONGSHOT_CHUNK_OPT=all: every lifted chunk (93) at -O2 build: LONGSHOT_CHUNK_OPT: chunks cut into parts of about 6 MB (338 part files) the build took 0:07:54, compiled in 0:05:52 ``` ### Flags and settings None. ## Testing **Tested:** `python -I -B tests/test_tested_configurations.py`: the record checks pass; its one failure (`test_cli_ascii_pipe_emits_the_arrow_containing_source_as_utf8`, a Windows line-ending case) fails on main too. The runs themselves are what the records describe. **Not tested:** other PCs; the King Zora stairs as timed runs (the bridge can't walk there; hand readings are on the wiki, marked as such). **Must be tested:** - None ## Risks **What could break:** nothing; records only. **Turning it off:** revert the commit. **Compatibility:** Nothing changes. ## Issues See #1485 ## Checklist ### Everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`) of the status line's type. - [x] Why gives the reason plainly and links the issues that fit (the tracker searched). - [x] Tested on my own PC where it runs (in the game, for a change the game runs), as Testing's Tested says; what can't be run by hand is in Testing's Not testable by hand, with why. - [x] Measurements has a verdict, or one line saying why there is nothing to measure. - [x] Nothing gets worse that I know of: any cost (speed, latency, memory, a setup that worked before) is measured and in Risks. - [x] Testing names what was tested, what wasn't, and what must still be tested, where and how. - [x] Rebased on current `main`. - [x] `python -I -B tests/run_all.py` passes (any failure that also fails on `main` is named). (Docs only: the check that reads these files, `tests/test_tested_configurations.py`, was run; its one failure also fails on main.) - [x] No game files, dumps, build output or generated files. - [x] `third_party/` untouched (a new folder with a `PROVENANCE.md` instead). - [x] The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled. ### When it applies - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [x] Numbers from the same PC, scene and settings, at least 2 interleaved runs each, the median reported. - [x] A file in `docs/tested/` for each new measured run (format in `docs/tested-configurations.md`). 🤖 Generated with [Claude Code](https://claude.com/claude-code)
docs(tested): the cloth gate with every lifted chunk at -O2, against -O1, six interleaved pairs (#1485)
Some checks are pending
CI / Conventional Commit title (pull_request) Blocked by required conditions
CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Blocked by required conditions
CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Blocked by required conditions
CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Blocked by required conditions
CI / Workflow lint (actionlint) (pull_request) Blocked by required conditions
Guide screenshots / Guide images and local UI source review (pull_request) Blocked by required conditions
Third-party parts / As their notes say (pull_request) Blocked by required conditions
Third-party parts / The same as upstream (pull_request) Blocked by required conditions
Website / Website on Cloudflare (pull_request) Blocked by required conditions
Website / Website on Cloudflare (pull_request_target) Has been skipped
PR description / Description follows the template (pull_request_target) Successful in 4s
cf737b53b6
RTX 4070, Ryzen 9 7900X, Windows 11; teleport arrivals in the child Market, the adult Hyrule Field and the child Zora's
Domain, 20 s each by the game's frame counter, in the job queue's timing mode. Medians of six: Market 69.8 -> 77.0 fps
(+10.3 %), Hyrule Field 95.0 -> 101.6 (+6.9 %), Zora's Domain 91.5 -> 93.3 (within the spread).

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Assisted-by: Claude Code (claude-opus-5-5)
enum changed title from WIP: docs(tested): the cloth gate with every lifted chunk at -O2, against -O1, six interleaved pairs (#1485) to docs(tested): the cloth gate with every lifted chunk at -O2, against -O1, six interleaved pairs (#1485) 2026-11-18 17:56:55 +00:00
Some checks are pending
CI / Conventional Commit title (pull_request) Blocked by required conditions
CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Blocked by required conditions
CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Blocked by required conditions
CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Blocked by required conditions
CI / Workflow lint (actionlint) (pull_request) Blocked by required conditions
Guide screenshots / Guide images and local UI source review (pull_request) Blocked by required conditions
Third-party parts / As their notes say (pull_request) Blocked by required conditions
Third-party parts / The same as upstream (pull_request) Blocked by required conditions
Website / Website on Cloudflare (pull_request) Blocked by required conditions
Website / Website on Cloudflare (pull_request_target) Has been skipped
PR description / Description follows the template (pull_request_target) Successful in 4s
This pull request doesn't have enough approvals yet. 0 of 1 approvals granted.
Some workflows are waiting to be reviewed.
This branch is out-of-date with the base branch
You are not authorized to merge this pull request.
View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u 'tested/cloth-o2-runs:enum-tested/cloth-o2-runs'
git switch 'enum-tested/cloth-o2-runs'
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
MagicBeanDev/Longshot!1630
No description provided.