Skip to content
Snippets Groups Projects
Select Git revision
  • fourteen default
  • thirteen
2 results
Search by author
  • Any Author
  • Dhina17 Dhina17 Dhina17
  • electimon electimon electimon
  • Erfan Abdi Erfan Abdi erfanoabdi
  • LMO BuildBot LMO BuildBot lmo-buildbot
  • LMO GerritBot LMO GerritBot lmo-gerritbot
  • Mohammad Hasan Keramat J Mohammad Hasan Keramat J ikeramat
  • Nick Nick nift4
7 authors
  1. Jun 02, 2022
  3. Jun 01, 2022
  5. May 31, 2022
  7. May 27, 2022
    • Jimmy Shiu's avatar
      ADPF: Skip I signal and set Uclamp init value · 0dc4b0d8
      Jimmy Shiu authored 
      Skip I signal by setting kI to 0 and use UClampMinInit as the display
update boost value.

Bug: 198663407
Test: Manual Test
Change-Id: I309ae4f63c4d7f50dc320aec28fffbe1473eb6e5
      0dc4b0d8
    • Richard Hsu's avatar
      [Tuning] Performance turning for NNAPI ML benchmarks · 18797cfc
      Richard Hsu authored 
      Sets the ML_ACC boost value to 650. The medium core's capacity is 763, so this means our benchmark process will always be scheduled onto the medium or big cores. We experimented with different uclamp values, and 650 is a good balance between power and performance.

Measurements:
MLPerf IC model Latency (ms) Power (mW) Energy/inference (mJ) MLPerf Scores
Uclamp = 512 0.981 3539 3.47 826
Uclamp = 600 0.973 3683 3.58 872
Uclamp = 650 0.899 3917 3.52 970 <- good balance
Uclamp = 700 0.892 4067 3.63 976
Uclamp = 800 0.875 4440 3.89 997

https://docs.google.com/presentation/d/1zx7sLkhOClmuRTCrq8-l3N1mZrrv7f-CtcdMuzV0eaI/edit?usp=sharing

Bug: 232183574
Test: MLPerf score on P22 improved significantly.

Change-Id: Ib93ecee5af3f1cbe83c5a96dc14b3a537db5abc8
      18797cfc
    • Richard Hsu's avatar
      [Tuning] Disable CPU Idle in NNAPI workload with PMQoS CPU DMA Latency · 2c15fb2a
      Richard Hsu authored 
      To improve the return path latency, we want to keep the CPU at at least WFI state (Idle_1). The PMQos cpu_dma_latency knob prevents the CPU from falling below WFI state. This makes the return path CPU wakeup latency very good. Check with wvw@, the power impact shouldn't be too significant.

The average energy cost per inference dropped from 3.85 to 3.47 mJ. The reason why the power number is lower WITH disable-idle is that, due to better latency, we get to run more inferences given the same amount of time. This makes the average power consumption lower.

Measurement:

MLPerf IC model Latency (ms) Power (mW) Energy/inference (mJ) MLPerf scores
Default 1.35 2837 3.85 560
Disable CPU Idle 0.98 3539 3.47 826

https://docs.google.com/presentation/d/1zx7sLkhOClmuRTCrq8-l3N1mZrrv7f-CtcdMuzV0eaI/edit?pli=1#slide=id.g12dd9e50b4b_0_0

Bug: 232183574
Test: MLPerf on Android T. Performance improved. Verified on Perfetto.
Change-Id: Ia807bf0849e4d9b0b0e8c9510335129ca89e791f
      2c15fb2a
    • Carter Hsu's avatar
      audio: update FM tuning 20220527 · ee6a5a5e
      Carter Hsu authored 
      
FM table change list:
1. HE
    a) Enble BVE for BB on NB,WB, SWB
    b) Fine tune BB RX EQ on NB, WB.
2. HH
    a) Fine Tune RX for dual speaker
    b) Polishing SWB NS performance
3. HA
   a) Update TMOUS tuning for P10/C10 on NB/WB/SWB
   b) Fine tune C10 tuning setting for Generic on NB/WB/SWB
   c) Fine tune P10 tuning setting for Generic on NB
   d) Fine tune NS tuning for P10/C10 on SWB

Bug: 234079372
Test: build pass
Signed-off-by: default avatarCarter Hsu <carterhsu@google.com>
Change-Id: If380495da1e94fff33a2bdc9980f0e546d167e7f
      ee6a5a5e
  9. May 26, 2022
  11. May 25, 2022
  13. May 24, 2022
    • Sayanna Chandula's avatar
      thermal: add tpu/aur power hints for emergency severity · 14edfc76
      Sayanna Chandula authored 
      
Add power hints handling for EMERGENCY severity. Control temperature
and switchon temperature are lowered when the pwoerhint is received.
Remove power hint handling for isp thermal zones as PID control loop
is not enabled for isp as of now.

Bug: 229014824
Test: Build and boot on device. Check control/switchon temp when
thermal severity reaches EMERGENCY level

Signed-off-by: default avatarSayanna Chandula <sayanna@google.com>
Change-Id: I2e4bab9495d76e231df8277f3bf4ef68d5a141af
      14edfc76
    • Yichi Chen's avatar
      RRS: Apply the default config from persist prop · 493be28f
      Yichi Chen authored 
      vendor_config plays as another role to control the display config during
the boot time. To change the default configuration of the user selected
mode, we use persist config to store the value.

Bug: 232721840
Test: Boot w/ and w/o user selected configs and check the resolution
Change-Id: I8e6f551817e9b7cb747a7f5351430dd2ac9b4624
      493be28f
    • Jasmine Cha's avatar
      audio: improve mute duration. · 2a027324
      Jasmine Cha authored 
      
- preview of notification will be cut off slightly becuase of mute
  duration in framework is 160ms (40*40)
- set mute duration of primary output from 160 ms to 40 ms.

Bug: 233176336
Test: preview notification/end call sound on voice/VoIP(Line)

Signed-off-by: default avatarJasmine Cha <chajasmine@google.com>
Change-Id: I7f5225372c265cb22e74311b0805d16b1c68fd3d
      2a027324
    • linpeter's avatar
      display: update the indoor auto-brightness curve · 46546e2e
      linpeter authored 
      Modified the indoor range mapping in
the indoor auto-brightness curve.

Bug: 200227638
Test: check the default curve
Change-Id: I7b9219023fabdefe05cb910666c0e0ff998ed06c
      46546e2e
  15. May 23, 2022
    • Darren Hsu's avatar
      powerstats: update display resolution setting · 3c30a321
      Darren Hsu authored 
      
Since runtime resolution switching is enabled, the specific device
supports to run QHD displays at lower FHD resolutions. So we should
support both QHD/FHD resolution settings in PowerStats HAL.

Bug: 232313626
Test: dump power stats at QHD/FHD resolutions
Change-Id: I46a45c078da146a8ecbeb49af4765771a2c0b651
Signed-off-by: default avatarDarren Hsu <darrenhsu@google.com>
      3c30a321
  17. May 20, 2022
  19. May 19, 2022
  21. May 18, 2022
  23. May 17, 2022
  25. May 16, 2022
    • Jimmy Shiu's avatar
      ADPF: remove UclampMin_Granularity · 392b34f0
      Jimmy Shiu authored 
      No need to prevent uclamp.min update rate by Granularity.
There is no perf impact.

Bug: 232336917
Test: /data/nativetest64/libperfmgr_test/libperfmgr_test
Change-Id: I535a57b68da9e9b37f4126e46d0410ccb410be46
      392b34f0
  27. May 14, 2022