refactor(vk): 111 kill switches are deleted, the host's switches join the table, and the per-PC decisions get one place and one rule #1309
No reviewers
Labels
No labels
area/accessibility
area/audio
area/input
area/install
area/menu
area/renderer
area/saves
area/tooling
area/upscaling
beanstalk: approved
beanstalk: failed
beanstalk: needs approval
beanstalk: running
beanstalk: tested
ci: approved
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/macos
platform/steam-deck
platform/windows
preview
priority
critical
priority
high
priority
low
regression
risk
high
risk
low
risk
medium
status
blocked
status
confirmed
status
duplicate
status
invalid
status
needs-info
status
wontfix
ci: approved
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
MagicBeanDev/Longshot!1309
Loading…
Reference in a new issue
No description provided.
Delete branch "HumanBean/Longshot:refactor/kill-switches"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Type: refactor · Risk: medium · Tested on: Linux, RTX 3090 · Needs: Windows (AMD, Intel); RADV, AMDVLK and ANV on Linux
Why
Track F of the renderer rework plan: fewer switches, decided in one place, the same on every driver. 177 renderer switches only turned something off or back to an older way, each keeping a second path alive that nobody runs; the host's 216 switches were in no table; and 73 rules that go by the PC were spread over 14 files, asked several ways and parsed seven ways. This is the one pull request for that track; it took in #1325, #1326 and #1341. See #1300 (the table).
What changed
Read the commits in order; each stands alone.
GONErow that tells a run which still sets it what became of it.gpu/vk/switches.def); the start's log names every switch a run set and says aHOS_name that is none.gpu/vk/device_policy.h): 67 decisions worked out from facts filled once, with avk: policy:block in the log.=onis on):docs/development/device-policy.mdhas the table.Measurements
Verdict: no change in fps, as intended. This head against
main, one session: Kokiri Forest 66.6 fps (main 63.2 to 64.9), Hyrule Field 92.4 (90.3 to 90.7), the Market 70.0 (68.0 to 72.7).Testing
Tested:
python -I -B tests/run_all.py: 657 passed, 45 failed, 19 skipped. Onmainat8e9cfc3athis PC has 46 checks failing (648 pass, 19 skipped); none of them is new here.Not tested:
Must be tested:
@Beanstalk steamdeck: main against this PR, Kokiri Forest, Hyrule Field, the Market).@Beanstalk windows amd and windows intel: main against this PR, Kokiri Forest, Hyrule Field, the Market).Risks
What could break: a deletion that kept the wrong side changes behaviour for everyone on that path; a corrected rule misbehaves on the driver it now reaches or no longer reaches.
Turning it off: revert a commit; each corrected rule is one commented line in
device_policy()that says how to go back.Compatibility: a run that sets a removed switch gets the default and a log line. No cache is rebuilt,
MKWVK_API_VERSIONstays.Issues
See #1300
Checklist
Everyone
type(scope): description) of the status line's type.main.python -I -B tests/run_all.pypasses (any failure that also fails onmainis named).third_party/untouched (a new folder with aPROVENANCE.mdinstead).When it applies
Assisted-by:/Co-Authored-By:trailers.tests/that reads changed source updated in this PR, not weakened.🤖 Generated with Claude Code
https://claude.ai/code/session_017vw8dzCsAL3UdW7ft8ptL9
Tested in game: RTX 3090, NVIDIA 620.70, Linux: a dev pack over base
8e9cfc3a, nine sessions through Kokiri Forest, Hyrule Field and the Market (standing and turning, 2560x1440, Threaded rendering on) interleaved with five of main. Every scene drew with main's draw counts and GPU times, the frame rates overlap main's, no session ended early.WIP: refactor(vk): 93 kill switches and the older paths they kept are deleted from the four large renderer filesto WIP: refactor(vk): 111 kill switches and their older paths are deleted, and the host's switches join the table2e0743237dabbc397632WIP: refactor(vk): 111 kill switches and their older paths are deleted, and the host's switches join the tableto WIP: refactor(vk): 111 kill switches are deleted, the host's switches join the table, and the per-PC decisions get one place and one ruleWIP: refactor(vk): 111 kill switches are deleted, the host's switches join the table, and the per-PC decisions get one place and one ruleto refactor(vk): 111 kill switches are deleted, the host's switches join the table, and the per-PC decisions get one place and one rule