Energy-Adaptive%20Display%20System%20Designs%20for%20Future%20Mobile%20Environments%20%20(by%20Subu%20Iyer,%20Lu%20Luo,%20Robert%20Mayo,%20Parthasarathy%20Ranganathan)

1
Energy-Adaptive Display System Designs for Future
Mobile Environments(by Subu Iyer, Lu Luo,
Robert Mayo, Parthasarathy Ranganathan)

• Presented by Huitao, Rean, Alex

2
Main Contributions

• First detailed analysis and quantification of
  display usage
• Energy-adaptive display system design
• Hardware Organic LED (OLED)
• Software Dark-Windows (software that leverages
  OLED technology to reduce power consumption)

3
Motivations

• Increased acceptance and use of mobile devices
• Limited battery capacity
• Energy consumed by display sub-system
• Half on laptop system
• 61 on handheld device
• Relatively invariant across technology shrinks

4
Related Work

• Previous approaches
• Aggressively turning off the entire display
• Lower-quality or smaller sized displays
• New technology Organic LED (OLED)
• Built from small organic molecules that that
  efficiently emit light when stimulated by an
  electric field
• require no backlighting and use energy
  proportional to the overall light output of the
  display

5
Related Work (cont.)

• Varying display characteristics (Choi et al.)
• refresh rate
• color depth
• backlight luminance
• Hardware-level and other tradeoffs (Kamijoh et
  al.)
• Various standby and idle configurations of
  hardware
• Control the number of pixels turned on or off
• Reduce the duty factor of the display
• OS and applications modifications (Flinn and
  Satyanarayan)
• Reduce computation with lowered fidelity of
  images
• Zoned backlighting
• Other components in display sub-system (Xiong et
  al.)
• Focus on the power consumed by the display
  controller and driver

6
Outline

• Study of display usage
• Suggested Technology
• Prototyping and Evaluation Results
• Conclusion

7
User Study Objectives

• To understand screen usage patterns and exploit
  opportunities for power reduction

8
User Traces
9
Active Screen Usage
10
Correlating Screen Usage and Applications
11
User study results summary

• Users use 60 of the screen available
• Large fraction of smaller windows associated with
  system related/application control messages
  independent of user preferences
• Significant mismatches between actual screen
  usage and display attributes e.g. resolution,
  color, brightness etc. Most of the smaller
  windows could be displayed with lower resolution
  equivalently
• These insights pave the way for energy-adaptive
  system designs to significantly reduce the energy
  consumption of the display sub-system

12
Technology proposed approach

• Hardware OLED
• Allows for variability in the display power
  output
• Power proportional to brightness and color
• Software Dark Windows
• Use window of focus as users active area
• Dims the background and background windows

13
Dimming Strategies
14
Prototyping Dark-Windows Xvnc

• Changed pixel values in the frame buffer
• Used X objects to track window of focus
• Performance overhead negligible

15
Evaluating Energy Benefits Synthetic Trace

• Modeled behavior observed in user study and
  replayed it on the prototype
• Average Window of Focus size (59)
• Applications similar to the ones in user study
• Background teal (pixel RGB 0,131,131)
• Trace Duration 1000 seconds
• Used software power model to calculate energy
  savings by OLED and Dark Windows

16
Power Model

• For 15 display
• Pcontroller 0.5W, Pdriver 1W
• For 1024 X 768 resolution display
• Pred 4.3µW, Pgreen 2.2µW, Pblue 4.3µW

17
Energy Savings

• OLED achieves 25 of power reduction
• Fully-Dimmed achieves 20 reduction on top of
  OLED (43 vs. LCD)

18
Extreme Windows/Background Combinations

• Energy savings partially depend on user settings

19
Evaluation Summary

• OLED design achieves 25 reduction in energy
• Dark Windows achieves further optimizations
• Reduction depends on user-specific settings
  (background, window color window sizes)
• 20 more reduction achieved in user trace
• Informal Study Reasonable acceptance of
  dark-windows display

20
Alternatives to Dark Windows

• Follow area around the cursor
• Application specific dimming (e.g. highlight the
  current slide in PowerPoint)
• Aged dimming of background windows
• Explore other properties (resolution, refresh
  rate)

21
Conclusion

• Study quantifies opportunities in energy-savings
  with adaptive displays
• On average 60 of display used by windows of
  focus
• Contents can often be displayed with smaller
  resolution and less brightness
• Simulation of OLED and Dark Windows shows
  reduction in power consumption

22
Study Shortcomings

• Small Sample used (17 users)
• Measurements performed on specific settings (e.g.
  teal background)
• Windows of focus may not be the only areas of
  focus
• Study does not include hand-held devices