1801463 - Js1k demo (https://js1k.com/2015-hypetrain/demo/2287) is very slow

Status: RESOLVED FIXED

(Core :: SVG, enhancement)

Description

[Mayank Bansal]

Go to https://js1k.com/2015-hypetrain/demo/2287

https://share.firefox.dev/3tP2wgJ

Comment 1

[Jonathan Kew [:jfkthame]]

The profile here is dominated by FlattenedPath::ComputePointAtLength, which is getting called repeatedly to position each glyph along the path. With a long path, this gets really expensive as we iterate through the path operations from the beginning each time.

It looks like we could significantly improve the performance here with a simple "cursor" in FlattenedPath, so that when iterating forwards to position each glyph, we continue from the previous position instead of starting over.

Comment 2

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1801463 - Use a cursor in FlattenedPath to accelerate successive calls to ComputePointAtLength for SVG text-on-a-path layout. r=#gfx-reviewers

This enables my local build to achieve 60fps on the js1k demo linked from the bug,
whereas without the patch I get barely 10fps.

Note: it's still possible for ComputePointAtLength would perform poorly if
the caller is iterating backwards or doing random access to a long path.
A potential mitigation for that would be to add an mLength field to FlatPathOp,
storing the length-so-far of the path, so that ComputePointAtLength could do
a binary search instead of linear accumulation. But this would add significant
memory overhead, and may not be worth doing; the low-overhead cursor here
appears to be enough to make text-on-a-path perform much better.

Comment 3

[Mayank Bansal]

bug 1801307 appears to be using an analogous approach in gpu-canvas.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/100b45bcf622
Use a cursor in FlattenedPath to accelerate successive calls to ComputePointAtLength for SVG text-on-a-path layout. r=gfx-reviewers,nical

Comment 5

[Norisz Fay [:noriszfay]]

https://hg.mozilla.org/mozilla-central/rev/100b45bcf622

Comment 6

[Mayank Bansal]

This is much better now : https://share.firefox.dev/3lcNOiI
Thanks very much for the fix!