[Webkit-unassigned] [Bug 259538] New: Potential Optimization - Avoid multiple recursions through the tree when calculating percent height

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=259538

            Bug ID: 259538
           Summary: Potential Optimization - Avoid multiple recursions
                    through the tree when calculating percent height
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
               URL: https://jsfiddle.net/stwre9j6/show
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: Layout and Rendering
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: ahmad.saleem792 at gmail.com
                CC: bfulgham at webkit.org, simon.fraser at apple.com,
                    zalan at apple.com

HI Team,

Based on discussion with Alan (on Slack 1-1), I am exploring whether we can do this merge:

Blink Merge: https://chromium.googlesource.com/chromium/src.git/+/74e95b29247f21620cb94fdf72e7e1051fff80c2

WebKit Source: https://searchfox.org/wubkat/source/Source/WebCore/rendering/RenderBox.cpp#731

Performance Test: https://jsfiddle.net/stwre9j6/show

_____

Safari Technology Preview 174:

Description: Measures performance of nested divs with percent values for height and max-height.

Time:
values 2.1645021645021645, 2.1645021645021636, 2.1668472372697734, 2.1691973969631246, 2.1810250817884405, 2.185792349726774, 2.183406113537118, 2.185792349726776, 2.185792349726776, 2.18579234972678, 2.1857923497267717, 2.183406113537118, 2.181025081788445, 2.178649237472767, 2.1834061135371137, 2.1739130434782696, 2.183406113537118, 2.183406113537118, 2.178649237472767, 2.1739130434782608 runs/s
avg 2.178910800251782 runs/s
median 2.182215597662779 runs/s
stdev 0.007440810283025579 runs/s
min 2.1645021645021636 runs/s
max 2.18579234972678 runs/s

and

Chrome Canary 117

Description: Measures performance of nested divs with percent values for height and max-height.

Time:
values 11472.47119099994, 11770.888071650446, 11833.069202510645, 11837.759282350935, 11977.542108546475, 12051.21767520943, 11902.231668720548, 11945.073990466863, 11837.759280952629, 11671.225739175741, 11867.549668639947, 11853.419764191982, 11746.198217140805, 11733.892090649078, 11733.8920904201, 11643.924626380767, 11462.197774693923, 11594.202897924757, 11478.349986619643, 11466.598413454549 runs/s
avg 11743.97318703496 runs/s
median 11758.543144395626 runs/s
stdev 178.66304514479208 runs/s
min 11462.197774693923 runs/s
max 12051.21767520943 runs/s

and Firefox Nightly 117 is similar ballpark to Chrome.

_________

Alan raised concern along the lines that the code can raise potential regressions, so I think  I will continue to explore this locally but if someone has better and faster way to do it. Go ahead and knock it out of the park.

Thanks!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20230726/97afa71c/attachment-0001.htm>