Repository navigation
On iOS the first GPU out-of-memory failure ends the game: an iPad Pro went down whole two failures into Ocarina of Time's 8,523 of 8,192 MB - #1486
Open
aksmith-git wants to merge 1 commit into
Conversation
github-actions
Bot
added
nvn
… went down whole two failures into Ocarina of Time's 8,523 of 8,192 MB res_top_once ended the process after 32 Insufficient Memory failures within a minute. On an iPad Pro (M5, iOS 27) Ocarina of Time logged the residency report (8,523 MB against Metal's 8,192 MB working set), failed one more command buffer, and the whole device went down: the log stops there. A run that went on drew pink or broken frames of 120-600 ms of GPU each. On iOS the first failure now ends the game, after the report; the Mac keeps its 32 (the process ends at the 33rd, as before).
aksmith-git
force-pushed
the
pr/ios-gpu-oom
branch
from
November 6, 2026 12:23
af9c053 to
7dff607
Compare
john62036203-ops
added a commit
to john62036203-ops/NXNativeWindows
that referenced
this pull request
Nov 6, 2026
… the game) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013SzykpcKm2eZPvPjGPKx1C
brad-n
added a commit
to brad-n/TVPort
that referenced
this pull request
Nov 9, 2026
… RG11B10F non-finite clamp (white bloom), seedds#1876, DM_NVN_LEDGER
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What and why
res_top_once(nvn_mtl.m) ends the process after 32Insufficient Memorycommand buffer failures within a minute (Repeated GPU out-of-memory failures end the game). An iPad does not get that far. On an iPad Pro (iPad17,3, M5, iOS 27), Ocarina of Time (0400486024DEC000v131072) logged the report at its first failure:One more command buffer failed, and then the whole device went down. The log stops there. In another run on the same iPad the game went on after three failures, with pink or broken frames that took 120-600 ms of GPU each, and did not recover.
On iOS the first failure now ends the game, after that report. The Mac is unchanged: the process ends at the 33rd failure within a minute, as before.
This is a stop, not a fix: what fills the working set is #1436, and #1484 takes 334 MB out of it. #1391's guard deliberately sets no stop level on iOS. This change acts only once Metal has already refused a command buffer.
Checked
make -C dm/host GAME=0400486024DEC000 dmhostbuilds on main 0f510fd;dm/ios/tools/compile_check.py(261 files for the iPhone SDK, 0 failed);test_docs.py.regress.pynot run: the Mac path is unchanged.Docs