github/flutter
7
0
Fork 0
mirror of https://github.com/flutter/flutter.git synced 2025-06-03 00:51:18 +00:00
Code Issues Packages Projects Releases Wiki Activity
b7a9e60850
flutter/dev/benchmarks/microbenchmarks
History
Danny Tuppeny cea1d6f055
Unpin DDS and roll pub packages (#147925) 
DDS was temporarily pinned to 4.1.0 because 4.2.0 triggered some test
failures (see https://github.com/flutter/flutter/pull/147250). Those
failures should be fixed by vm_service 14.2.2, so this unpins DDS and
rolls both of these packages (along with devtools_shared, which is a DDS
dependency).

(If the bot updates vm_service before this is done, I can rebase over
that will reduce the size of this PR to just a few files)
2024-05-08 10:35:28 +01:00
..
android Reland "Update gradle lockfiles template (#140115)" (#143081) 2024-02-08 16:22:15 +00:00
ios Migrate Xcode projects last version checks to Xcode 15.1 (#140256) 2024-01-03 23:05:46 +00:00
lib Reland "Add a microbenchmark for text intrinsic height layout (#145007)" (#145037) 2024-03-13 00:52:24 +00:00
money_asset_manifest.json Add asset manifest parsing benchmark (#112836) 2022-10-07 13:39:35 +00:00
pubspec.yaml Unpin DDS and roll pub packages (#147925) 2024-05-08 10:35:28 +01:00
README.md Added missing code block language in docs (#147481) 2024-05-01 14:44:27 +00:00

README.md

microbenchmarks

To run these benchmarks on a device, first run `flutter logs' in one window to see the device logs, then, in a different window, run any of these:

flutter run --release lib/gestures/velocity_tracker_bench.dart
flutter run --release lib/gestures/gesture_detector_bench.dart
flutter run --release lib/stocks/animation_bench.dart
flutter run --release lib/stocks/build_bench.dart
flutter run --release lib/stocks/layout_bench.dart

The results should be in the device logs.

Avoid changing names of the benchmarks

Each microbenchmark is identified by a name, for example, "catmullrom_transform_iteration". Changing the name of an existing microbenchmarks will effectively remove the old benchmark and create a new one, losing the historical data associated with the old benchmark in the process.