Repository navigation
stock_animation_open_first_frame_average regressed #42564
Description
This looked fairly noisy before but never got back below the selected baseline after #42278.
{{ message }}
This looked fairly noisy before but never got back below the selected baseline after #42278.
Activity
chinmaygarde commentedon Oct 12, 2019
cc @gspencergoog
gspencergoog commentedon Oct 12, 2019
Dang, looks like I just moved the regression from one benchmark to another. I might need some help finding something other ways to speed this up: I've already tried a lot of things.
Optimize focus operations by caching descendants and ancestors. (#42683)
gspencergoog commentedon Oct 15, 2019
OK, I'm officially stumped. I spent all day yesterday optimizing this code, and thought I had found a combination that almost eliminated this regression (1.3% instead of 5.1%), so I checked it in, but there has been no change in the benchmark.
I have a spreadsheet where I tracked my changes and their effect, and it showed a clear improvement. I ran the benchmark 10x for each data point in my spreadsheet, and took the median (and the mean, just to sanity check). I'm running it on a Moto G4, set up the same as our benchmark devices.
If I can't measure the impact of my changes, it's going to be pretty hard to optimize this...
gspencergoog commentedon Oct 15, 2019
After talking with others about this, I am going to rebaseline this as a fixed cost for the feature of keyboard shortcuts and focus traversal. I've done quite a lot of investigation and optimization of this code, and in reality there are probably more impactful optimizations that could be done in other areas of the code.
Optimize focus operations by caching descendants and ancestors. (flut…
github-actions commentedon Aug 27, 2021
This thread has been automatically locked since there has not been any recent activity after it was closed. If you are still experiencing a similar issue, please open a new bug, including the output of
flutter doctor -vand a minimal reproduction of the issue.