WIP: docs(vk): record the Hyrule Field slowdown and A/B plan #363

Draft

Buasfe wants to merge 1 commit from Buasfe/Longshot:docs/issue-348-renderer-ab into main

pull from: Buasfe/Longshot:docs/issue-348-renderer-ab

merge into: MagicBeanDev:main

MagicBeanDev:main

Conversation 0 Commits 1 Files changed 1 +93

Buasfe commented 2026-11-14 04:27:46 +00:00

First-time contributor

• Copy link
• Copy Markdown

Why

Riding through Hyrule Field on Longshot v3.0.0-rc.3 led to a large FPS drop on my RTX 3070, even without changing areas. I put the timing evidence and a repeatable A/B plan in one place so we can check whether current Longshot still behaves the same way.

What changed

• docs: docs/hyrule-field-dispatch-investigation.md records the rc.3 dispatch timings, what the Nsight review suggests, and how to compare early and late runs with the same camera and workload.
• The older Hookshot memory, eviction, upload and rollback results are included as background. The diff changes documentation only.

Flags and settings

None. The protocol uses existing logging and profiling switches.

Measurements

main: cf605c85 (2026-11-14)
this PR: docs/issue-348-renderer-ab @ cf6f7ded

Baseline: rc.3 @ 0c74253d at 1,800 presents. These columns compare stages of the saved diagnostic run.

Saved diagnostic run: Longshot v3.0.0-rc.3 (release tag 0c74253d, log source fingerprint cb01b5a3). RTX 3070 8 GB, driver 621.14, Ryzen 7 5800X3D, 32 GB DDR4, Windows 11. Adult Hyrule Field on Epona, 3440×1440, upscaling and frame generation off. One inspected run with RX_VK_PASS_TIMING=1. Each pass timing report covers 600 presents.

Metric	rc.3 @ 0c74253d at 1,800 presents	rc.3 @ 0c74253d at 2,400 presents	rc.3 @ 0c74253d at 3,000 presents
Dispatch 26fe0c70ab9fc547	9.225 ms/frame	10.412 ms/frame	21.499 ms/frame
Nearby VRAM status, used / allowance	6.4 / 6.9 GB	6.3 / 6.9 GB	6.1 / 6.8 GB
Nearby 120-present FPS	33.1	20.0	18.2

vk:   m   9.225 ms     0.9 x  10288.1 us       0.9 items    0.00 MB  dispatch  prog 26fe0c70ab9fc547
vk:   m  10.412 ms     0.9 x  11633.2 us       0.9 items    0.00 MB  dispatch  prog 26fe0c70ab9fc547
vk:   m  21.499 ms     1.0 x  21498.7 us       1.0 items    0.00 MB  dispatch  prog 26fe0c70ab9fc547

These figures compare stages of one rc.3 run. They provide no current-main comparison or estimate of an FPS improvement. The timestamps affect overlap, and the VRAM status is a periodic sample. The repeat instructions are in the new document.

Testing

Tested: On Windows, 22 tests passed in tests/test_check_pr_description.py and 32 in tests/test_forge_pr.py. Checked all 21 data values in the three timing rows, the total instrumented span time and both development-guide links against local source/evidence. git diff --check passed.

Not tested: Full game-free suite deferred to avoid running the compiler-heavy checks during gameplay. No fresh build, gameplay run, current-main A/B or new GPU capture. The original Nsight observations are reported as hypotheses and supporting evidence.

Needs other testers:

• NVIDIA 8 GB on current main: repeat the Field loop and matched early/late static views first with pass timing off, then in a diagnostic-only repeat. Check whether the FPS drop and corresponding dispatch-time increase reproduce. Please report the exact commit, hardware and settings (request to open).

Screenshots

Nothing on screen.

Risk and rollback

Documentation only. Reverting the commit removes the protocol and historical summary.

Reviewer notes

The current tree includes sparse mirrors #250 and host memory accounting changes #324. Mirror sync #326 targets a separately timed path. The first question is whether the Field slowdown persists after these baseline changes. I'd keep any counter placement change or legacy trim rollback in its own PR once it has a current reproduction.

I collected the gameplay observations, logs and Nsight evidence. OpenAI Codex assisted with the prior Hookshot source investigation, patches and test automation, and with this Longshot source review, log verification, documentation and A/B plan. The branch commit carries the attribution trailer.

Issues

See #348

Checklist

For everyone

• One topic.
• The title is a Conventional Commit (type(scope): description).
• Rebased on current main.
• python -I -B tests/run_all.py run and passing (any failure that also fails on main is named above). Full suite deferred. Targeted verification passed.
• Tested in the game, or what wasn't is named under Testing.
• Testing filled in: what was tested, what wasn't, and what needs other testers (with test requests). The test request is pending.
• Measured: main and this PR at named commits, on the same PC, scene and settings, at least 2 interleaved runs each, the median reported. No current-main test was performed. Historical rc.3 measurements are listed above.
• 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.

If applicable

• AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
• Docs updated.

## Why Riding through Hyrule Field on Longshot v3.0.0-rc.3 led to a large FPS drop on my RTX 3070, even without changing areas. I put the timing evidence and a repeatable A/B plan in one place so we can check whether current Longshot still behaves the same way. ## What changed - docs: `docs/hyrule-field-dispatch-investigation.md` records the rc.3 dispatch timings, what the Nsight review suggests, and how to compare early and late runs with the same camera and workload. - The older Hookshot memory, eviction, upload and rollback results are included as background. The diff changes documentation only. ## Flags and settings None. The protocol uses existing logging and profiling switches. ## Measurements main: `cf605c85` (2026-11-14) this PR: `docs/issue-348-renderer-ab` @ `cf6f7ded` Baseline: rc.3 @ 0c74253d at 1,800 presents. These columns compare stages of the saved diagnostic run. **Saved diagnostic run:** Longshot v3.0.0-rc.3 (release tag `0c74253d`, log source fingerprint `cb01b5a3`). RTX 3070 8 GB, driver 621.14, Ryzen 7 5800X3D, 32 GB DDR4, Windows 11. Adult Hyrule Field on Epona, 3440×1440, upscaling and frame generation off. One inspected run with `RX_VK_PASS_TIMING=1`. Each pass timing report covers 600 presents. | Metric | rc.3 @ 0c74253d at 1,800 presents | rc.3 @ 0c74253d at 2,400 presents | rc.3 @ 0c74253d at 3,000 presents | |---|---:|---:|---:| | Dispatch `26fe0c70ab9fc547` | 9.225 ms/frame | 10.412 ms/frame | 21.499 ms/frame | | Nearby VRAM status, used / allowance | 6.4 / 6.9 GB | 6.3 / 6.9 GB | 6.1 / 6.8 GB | | Nearby 120-present FPS | 33.1 | 20.0 | 18.2 | ```text vk: m 9.225 ms 0.9 x 10288.1 us 0.9 items 0.00 MB dispatch prog 26fe0c70ab9fc547 vk: m 10.412 ms 0.9 x 11633.2 us 0.9 items 0.00 MB dispatch prog 26fe0c70ab9fc547 vk: m 21.499 ms 1.0 x 21498.7 us 1.0 items 0.00 MB dispatch prog 26fe0c70ab9fc547 ``` These figures compare stages of one rc.3 run. They provide no current-main comparison or estimate of an FPS improvement. The timestamps affect overlap, and the VRAM status is a periodic sample. The repeat instructions are in the new document. ## Testing **Tested:** On Windows, 22 tests passed in `tests/test_check_pr_description.py` and 32 in `tests/test_forge_pr.py`. Checked all 21 data values in the three timing rows, the total instrumented span time and both development-guide links against local source/evidence. `git diff --check` passed. **Not tested:** Full game-free suite deferred to avoid running the compiler-heavy checks during gameplay. No fresh build, gameplay run, current-main A/B or new GPU capture. The original Nsight observations are reported as hypotheses and supporting evidence. **Needs other testers:** - NVIDIA 8 GB on current main: repeat the Field loop and matched early/late static views first with pass timing off, then in a diagnostic-only repeat. Check whether the FPS drop and corresponding dispatch-time increase reproduce. Please report the exact commit, hardware and settings (request to open). ## Screenshots Nothing on screen. ## Risk and rollback Documentation only. Reverting the commit removes the protocol and historical summary. ## Reviewer notes The current tree includes [sparse mirrors #250](https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/250) and [host memory accounting changes #324](https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/324). [Mirror sync #326](https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/326) targets a separately timed path. The first question is whether the Field slowdown persists after these baseline changes. I'd keep any counter placement change or [legacy trim rollback](https://github.com/NotRexed/Hookshot/pull/2825) in its own PR once it has a current reproduction. I collected the gameplay observations, logs and Nsight evidence. OpenAI Codex assisted with the prior Hookshot source investigation, patches and test automation, and with this Longshot source review, log verification, documentation and A/B plan. The branch commit carries the attribution trailer. ## Issues See #348 ## Checklist ### For everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`). - [x] Rebased on current `main`. - [ ] `python -I -B tests/run_all.py` run and passing (any failure that also fails on `main` is named above). Full suite deferred. Targeted verification passed. - [x] Tested in the game, or what wasn't is named under Testing. - [ ] Testing filled in: what was tested, what wasn't, and what needs other testers (with test requests). The test request is pending. - [ ] Measured: main and this PR at named commits, on the same PC, scene and settings, at least 2 interleaved runs each, the median reported. No current-main test was performed. Historical rc.3 measurements are listed above. - [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. ### If applicable - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [x] Docs updated.

Buasfe added 1 commit 2026-11-14 04:27:46 +00:00

docs(vk): record the Hyrule Field slowdown and A/B plan ... cf6f7ded50

See #348.

Assisted-by: OpenAI Codex

mooki added the

kind

docs

priority

low

labels 2026-11-14 04:31:57 +00:00

Some checks are pending

Hide checks

CI / Conventional Commit title (pull_request) Blocked by required conditions

Details

CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Blocked by required conditions

Details

CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Blocked by required conditions

Details

CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Blocked by required conditions

Details

CI / Workflow lint (actionlint) (pull_request) Blocked by required conditions

Details

Third-party parts / As their notes say (pull_request) Blocked by required conditions

Details

Third-party parts / The same as upstream (pull_request) Blocked by required conditions

Details

PR description / Description follows the template (pull_request_target) Successful in 11s

Details

Website / Website on Cloudflare (pull_request_target) Has been skipped

Details

This pull request is marked as a work in progress.

Some workflows are waiting to be reviewed.

This branch is out-of-date with the base branch

View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.

git fetch -u 'docs/issue-348-renderer-ab:Buasfe-docs/issue-348-renderer-ab'

git switch 'Buasfe-docs/issue-348-renderer-ab'

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

area/accessibility

Barriers for people with disabilities

  

area/audio

Sound and music

  

area/input

Controllers, keyboard, mouse

  

area/install

Installing, updating, building from your game

  

area/menu

In-game menu and settings

  

area/renderer

Vulkan renderer, shaders, effects

  

area/saves

Save files

  

area/upscaling

DLSS, FSR, frame generation

  

good first issue

A good start for a new contributor

  

gpu/amd

AMD GPUs

  

gpu/igpu

Integrated GPUs

  

gpu/intel

Intel GPUs

  

gpu/nvidia

NVIDIA GPUs

  

help wanted

Extra hands or knowledge wanted

  

kind

bug

Something doesn't work as it should

  

kind

chore

Build, CI, dependencies, tidying

  

kind

docs

Documentation

  

kind

feature

A new feature or an improvement

  

kind

performance

Frame rate, frame time, loading, memory

  

kind

question

A question rather than a change

  

needs measurements

Waiting for before/after measurements

  

platform/linux

Linux only

  

platform/steam-deck

Steam Deck

  

platform/windows

Windows only

  

preview

Maintainers: deploy a preview of this PR's website (read its changes first). Removed after the run.

  

priority

critical

Crashes, data loss, or blocks most players

  

priority

high

Important: next in line

  

priority

low

Nice to have

  

status

blocked

Waits on something else first

  

status

confirmed

Reproduced or agreed on

  

status

duplicate

Already reported or proposed

  

status

invalid

Not a problem in this project

  

status

needs-info

Waiting for more information from the reporter

  

status

wontfix

Will not be worked on

  

testing wanted

Wanted: a test on other hardware (posts a test request in #test-requests)

No labels

area/accessibility

area/audio

area/input

area/install

area/menu

area/renderer

area/saves

area/upscaling

good first issue

gpu/amd

gpu/igpu

gpu/intel

gpu/nvidia

help wanted

kind

bug

kind

chore

kind

docs

kind

feature

kind

performance

kind

question

needs measurements

platform/linux

platform/steam-deck

platform/windows

preview

priority

critical

priority

high

priority

low

status

blocked

status

confirmed

status

duplicate

status

invalid

status

needs-info

status

wontfix

testing wanted

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

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!363

No description provided.