A.Olokhtonov
|
b961504b54
|
Separate stroke rendering into function. Reuse said function for static
and dynamic draw.
|
5 months ago |
|
A.Olokhtonov
|
5bb8b0a7ee
|
Further tune LOD levels to optimize triangle counts. Removed LOD level 7
|
5 months ago |
|
A.Olokhtonov
|
d618913111
|
Fix uninitialized return when no strokes are submitted to a thread
|
5 months ago |
|
A.Olokhtonov
|
500fd71d3e
|
Batching is working! (only in singethreaded mode though)
|
5 months ago |
|
A.Olokhtonov
|
777772530f
|
Batching is close to working. Not quite though
Also possibly fixed a very nasty bug, but probably not
|
5 months ago |
|
A.Olokhtonov
|
ce824a8e31
|
Pass stroke widths to wasm (not actually writing any values right now,
just some busywork to allocate memory and pass pointers)
|
5 months ago |
|
A.Olokhtonov
|
55f390e55b
|
Support different LOD level per batch of instances. Mostly OpenGL
busywork
|
5 months ago |
|
A.Olokhtonov
|
eb66ffbcad
|
Significantly improve stroke smoothing and pressure handling. "Fix" issue with underallocation of WASM memory by doubling the size
|
12 months ago |
|
A.Olokhtonov
|
c3f66d966c
|
Determinitstic zooming
|
1 year ago |
|
A.Olokhtonov
|
6d9791d3e0
|
Cache full stroke lists in BVH levels (massive speedup!). Remove "line_threshold" array, since it has very little effect.
|
1 year ago |
|
A.Olokhtonov
|
f6573e7bb9
|
Correcty complute total point count. Only call glClear once workers have finished LOD. Only allow next draw() call after we finished rendering frame
|
1 year ago |
|
A.Olokhtonov
|
4f84509b16
|
Multithreading works! Kinda sorta (slows down in chrome, out of bounds accesses on phone)
|
1 year ago |
|
A.Olokhtonov
|
2c01d315d6
|
A small test WASM module to demonstrate how to use shared memory and run multiple threads in WASM without Emscripten
|
1 year ago |
|
A.Olokhtonov
|
e949860279
|
Change extracts of pressure byutes to a f32x4_make. Change andnot masking to a bitselect. Rename readme.md to readme.txt to hopefully fix gitea formatting
|
1 year ago |
|
A.Olokhtonov
|
37ff1e55bc
|
Oops, actually enable SIMD
|
1 year ago |
|
A.Olokhtonov
|
84a5859541
|
SIMD version of rdp_find_max. 16->12ms cpu frametime firefox, 16->9ms chrome
|
1 year ago |
|
A.Olokhtonov
|
6f78c0ae21
|
AoS -> SoA for point coordinates
|
1 year ago |
|
A.Olokhtonov
|
58f2a60295
|
Fix wrong wasm allocation, remove unused js LOD code, take radius of stroke into account when doing LOD, reduce EPS for LOD
|
1 year ago |
 aolo2
|
dfecad2c99
|
Fix strokes being added to storage even when skipped by SN logic
|
1 year ago |
|
aolo2
|
704c833c16
|
First working draft of pressure hanlding
|
1 year ago |
|
aolo2
|
71f1016a40
|
Store WASM-processed data in WASM memory
|
1 year ago |
|
A.Olokhtonov
|
3bf812df39
|
Move point writes to WASM, already a lot faster!!!!
|
1 year ago |
|
A.Olokhtonov
|
1960bdebe9
|
LOD generation has been wassembled! (a little borked for now though)
|
1 year ago |
|
A.Olokhtonov
|
cf11f6c701
|
Successfull sum function in wasm :D
|
1 year ago |
|
A.Olokhtonov
|
28fec7f306
|
Redraw HTML on canvas move, first draft of wasm LOD core
|
1 year ago |
