Skip to content

Perf B5c: mulberry-river tree draws (dropped offline) #101

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hunterh37 merged 2 commits into from
Oct 17, 2026
Merged

Perf B5c: mulberry-river tree draws (dropped offline) #101

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ded5026
Perf B5c: record dropped (cell size 48 reaches 671 draws, tris over b…
hunterh37 Oct 17, 2026
2dae1a0
QUEUE: B5c PR url
hunterh37 Oct 17, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions docs/perf-roadmap.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -317,6 +317,10 @@ Items, in order. Each lists the measure that proves it and the sim bench that cl
- B5c mulberry-river tree fields (oak, shortleaf, sycamore, river-birch) are 784 of 857 draws: 640 draws
from 4 species because each LOD level splits into per-material parts. Merge bark+leaf parts per level
or cut LOD3 draws. Close: mulberry-river < 500 draws (phase 2).
Status (offline; sim validation pending): DROPPED. Surfaces of one level carry different materials (bark,
leaf, autumn leaf), so they cannot merge into one draw. Tried tree cellSize 24 -> 48 in MulberryVegetation:
draws 857 -> 671 (still > 500), tris 2044k -> 2184k (over the 2147k budget), coarser LOD switching. Reverted.
Remaining option: a single-surface LOD3 (needs a bark+leaf atlas material) or far-cell coalescing in RealInstancing.
- B5d mulberry-river rocks (limestone-algae x70 212k with 13872 LOD0 tris, ozark-limestone x70 88k): LOD1
is 19 % of LOD0 already, so the in-view cost is LOD0 near the camera; shorten rock switch distances or
cull. Close: rock fields < 150k in-view.
Expand Down
2 changes: 2 additions & 0 deletions notes/QUEUE.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,3 +5,5 @@ B7 | perf-b7 | https://github.com/hunterh37/RealityHD/pull/96 | hoverSwapsRestPr
B4 | perf-b4 | https://github.com/hunterh37/RealityHD/pull/97 | dropped offline: alpine in-view 2734k->1977k, diff 2.47% (need <1.5M, <0.1%) | none
B5b | perf-b5b | https://github.com/hunterh37/RealityHD/pull/100 | dropped offline: grass.tall 497k->141k (-72%), diff 10.8% (need halved, <0.1%) | none
B5a | perf-b5a | https://github.com/hunterh37/RealityHD/pull/99 | in-view 2225k->2183k, shadow 2122k->2068k, draws 330 | sim bench lakeside-camp
B5c | perf-b5c | PRURL | dropped: cellSize 48 gave 857->671 draws, tris over budget | none
B5c | perf-b5c | https://github.com/hunterh37/RealityHD/pull/101 | dropped: cellSize 48 gave 857->671 draws, tris over budget | none
Loading