M1: a simulated loss brings back an empty scene with working UI; the old device is freed #1547
Labels
No labels
amd
avatars
bug
builder
cannot-reproduce
crash
epic
feature
from-viewer
gl-removal
known-limitation
monitoring
parity
perf
platform
priority
critical
priority
high
priority
low
priority
medium
provenance
rendering
rt
status
in-progress
status
testing
status
up-next
triage
ui
upstream
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
sky_line/Slipstream#1547
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
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?
RenderVKDebugSimulateDeviceLost(and on a real driver-reported loss), keep the SL session and main loop running, set the old epoch aside, and build a new device, swapchain and UI on the same instance.Done when 50 simulated losses in a row each return to a working UI with an empty or partial scene, with no leak of device memory (checked with the memory log) and no validation errors.
Part of #1545.
First real in-session GPU loss on the owner's RTX 5090 (2026-11-18 18:55): a TDR hang (graphics
g_gi_window, asyncc_defer_rtrefl_sel).What recovered (12.2 s): RT, async compute, frame generation and the scene all came back.
What did not:
unattributed 9594 MBand the process grew to 15.2 GB from 9.9 GB, so the viewer now hits the VRAM budget and drops resources. The reaper's "(released)" verdict was wrong: 9901 -> 9644 MB.The owner considers this unacceptable for recovery. Brief: agents/recovery_vram_leak_brief.md (free the epoch's allocations and objects before vkDestroyDevice, fix the verdict, re-init NGX).
1.5.6 ships the lost-device storage reclaim and NGX discovery preservation (
9239e99ddd). Leaving open: this milestone's remaining scope is tracked here.