- Update Flame to v18D_nightly_v4
- Change memory to 512MB
- Use latest v2.2 build, October 7th
- Checkout Gaia v2.2 branch
- Reset Gaia using
GAIA_OPTIMIZE=1 DEVICE_DEBUG=1 PERF_LOGGING=1 RAPTOR=1 NO_LOCK_SCREEN=1 NOFTU=1 SCREEN_TIMEOUT=0 make reset-gaia
- Install workload[1]:
make reference-workload-light
- Ensure using Raptor v3.x
- Run Raptor for each app:
raptor test coldlaunch --app <app> --runs 30
- Record results for
visuallyLoaded
anduss
95% Bound
[1] Except for Calendar! Test Calendar first, then other apps using the workload!
- GC at
fullyLoaded
before memory capture - 10 seconds between application launches
- B2G restart between application tests