Bug 89547 – Add a performance test for paths in SVG

• Component: SVG
• Status: RESOLVED FIXED

[Philip Rogers]

We need better test coverage in SVG to track performance changes.

[Philip Rogers]

Created attachment 148502 [details]
Adds a performance test for paths in SVG

[Ryosuke Niwa]

Comment on attachment 148502 [details]
Adds a performance test for paths in SVG

View in context: https://bugs.webkit.org/attachment.cgi?id=148502&action=review

> PerformanceTests/ChangeLog:8
> +        This change adds the first performance test for SVG paths.
> +        In the test we modify complex cubic paths in several ways, testing:
> +        transformations, clipping, d attribute changes, stroke properties, and opacity.

Can we include sample outputs?

> PerformanceTests/ChangeLog:10
> +        Reviewed by NOBODY (OOPS!).

This line should appear before the long description.

> PerformanceTests/SVG/SvgCubics.html:858
> +      }, 2, 100, function() {

Do we really need 100 samples? How long does it take to run this test on, say, Mac mini?

[Philip Rogers]

Created attachment 148590 [details]
Update per reviewer comments

[Philip Rogers]

(In reply to comment #2)
> (From update of attachment 148502 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=148502&action=review
> 
> > PerformanceTests/ChangeLog:8
> > +        This change adds the first performance test for SVG paths.
> > +        In the test we modify complex cubic paths in several ways, testing:
> > +        transformations, clipping, d attribute changes, stroke properties, and opacity.
> 
> Can we include sample outputs?

Done (listed in the ChangeLog).

For reference, the result will look something like:
https://docs.google.com/file/d/0B4rN2u5hCJiUZmg0NTBob0Q0Ymc/edit?pli=1

> 
> > PerformanceTests/ChangeLog:10
> > +        Reviewed by NOBODY (OOPS!).
> 
> This line should appear before the long description.

Done.

> 
> > PerformanceTests/SVG/SvgCubics.html:858
> > +      }, 2, 100, function() {
> 
> Do we really need 100 samples? How long does it take to run this test on, say, Mac mini?

100 was not necessary as the sample output using just 20 shows that the std dev is quite reasonable. On my linux desktop this test now takes less than 1 second to run, and 2 seconds total including starting up the test script.

[Ryosuke Niwa]

Comment on attachment 148590 [details]
Update per reviewer comments

View in context: https://bugs.webkit.org/attachment.cgi?id=148590&action=review

> PerformanceTests/ChangeLog:14
> +            median= 32.0 ms, stdev= 2.92403830344 ms, min= 22.0 ms, max= 37.0 ms

We probably want to tweak the test so that's median is well over 50ms or even 100ms in the case things get faster in the future :)

[Philip Rogers]

Created attachment 148672 [details]
Slow test down to ~70ms per run

Add even more path data to slow the test down to ~70ms per run which should get us out of the range of accuracy of JavaScript timers.

[WebKit Review Bot]

Comment on attachment 148672 [details]
Slow test down to ~70ms per run

Clearing flags on attachment: 148672

Committed r120892: <http://trac.webkit.org/changeset/120892>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.