Frame stalls during buffer growth / region changes: no long full-GPU waits on the main thread #699
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#699
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?
Several growth/shrink paths wait for the whole GPU (device idle / fence waits) on the main thread: geometry buffer rebuilds, rigged buffer regrowth (several waits in one call), temporal / probe target re-creates, VRAM ladder rungs that rebuild targets, and synchronous pipeline-compile retries (~3.5 s each). Under memory pressure (Windows paging) these can stretch long enough to cause a region disconnect, and they tend to cluster during region changes.
Scope
Shipped in 1.3.0 Beta 7 (
4eb78641f8GPU wait accounting and redundant device-idle elision;453256ffdbno synchronous pipeline retry ladder on the render thread). Closing -- reopen or comment if you still see this.