You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: content/blog/2025-08-05-tmil-19/index.md
+13-5Lines changed: 13 additions & 5 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -10,14 +10,22 @@ Linebender is an informal open-source organization working on various projects t
10
10
Vello is our GPU vector renderer.
11
11
It can draw large 2D scenes with high performance, using GPU compute shaders for most of the work.
12
12
13
-
<!-- TODO -->
14
-
-[vello#...][]:
13
+
-[vello#785][]: Strokes with a width of zero are no longer treated as fills.
14
+
-[vello#908][]: Updates to wgpu 25 release.
15
+
-[vello#1093][]: Disabled runtime checks in shaders using wgpu's new `create_shader_module_trusted` API, by sagudev.
15
16
16
17
This month's progress on the sparse strips renderers, a collaborative evolution of Vello, has been centered around adding support for NEON and WASM SIMD, as well as making further improvements to multi-threaded rendering.
17
18
18
-
<!-- TODO -->
19
-
-[vello#...][]:
19
+
-[vello#1064][], [vello#1086][]: Image rendering in Vello Hybrid.
20
+
-[vello#1078][]: Rewrote Vello CPU to be more SIMD-friendly.
21
+
-[vello#1092][]: Adds support for SIMD flattening.
22
+
-[vello#1103][]: Optimises alpha coverage calculation in strip rendering.
23
+
-[vello#1105][]: Ignores paths containing NaN points, giving a warning.
24
+
-[vello#1122][]: Adds opacity layers to Vello Hybrid.
25
+
-[vello#1134][]: Reuses FlattenCtx for paths.
26
+
-[kurbo#427][]: Contains a new stroke expander with significant performance improvements.
20
27
28
+
<!-- TODO: Clean up long sentences, hedge, ensure link text is meaningfuly, etc. -->
21
29
In order to get a better understanding of where we lie performance-wise, we included `tiny-skia` and `vello-cpu` in the [Blend2D benchmark harness](https://blend2d.com/performance.html), a comprehensive benchmarking tool that tests the performance of different parts of a 2D renderer and compares it against other renderers.
22
30
In order to do so, a [fork](https://github.com/LaurenzV/blend2d-apps/tree/benching) of the benchmark harness was created that uses C bindings to both libraries to include them in the harness.
23
31
To visualize the results, we created a chart, similar to how it is available on the official Blend2D website.
@@ -35,7 +43,7 @@ Nevertheless, by looking at this chart, it is clear that `vello-cpu` has very im
35
43
36
44
Our [working roadmap](https://docs.google.com/document/d/1ZquH-53j2OedTbgEKCJBKTh4WLE11UveM10mNdnVARY/edit?tab=t.0#heading=h.j3duh9pgdm94) outlines the planned timeline for work on the renderers into next year.
37
45
38
-
An integration of Vello as the backend for Servo's canvas rendering has landed in <!-- TODO -->(and also Vello CPU in <!-- TODO -->).
46
+
An integration of Vello as the backend for Servo's canvas rendering has landed in [servo#36821][](and also Vello CPU in [servo#38282][]).
0 commit comments