Games are Up for DVFS

1
(No Transcript)
2
Games are Up for DVFS

• Yan Gu Samarjit Chakraborty Wei Tsang Ooi
• Department of Computer Science
• National University of Singapore

3
Outline

• Goal of this work
• Explore the possibility of using DVFS for
  interactive games
• Characterize the workload of game applications
  and use it to propose a DVFS scheme for games

• Introduction
• Anatomy of a Game Engine
• A First Cut Reduce Frame Rates
• The Case for DVFS
• DVFS for Games
• Conclusion

4
Introduction

• Popularity of interactive games on
  battery-powered portable devices (e.g. mobile
  phones, PDAs, PSP, etc.)
• Voltage and frequency-scalable processors on
  portable devices
• Can DVFS algorithms developed for video decoding
  applications be applied to games?
• no interaction in video decoding, no buffering in
  games
• We show that
• It is meaningful to use DVFS in the context of
  games
• frame structure in game applications can be
  exploited to develop DVFS algorithms

5
Anatomy of a Game Engine
6
A First Cut Reduce Frame Rates

• Current game design principles
• higher frame rates imply the better game quality
• Recent study on frame rates Claypool et al. MMCN
  2006
• very high frame rates are not necessary, very low
  frame rates impact the game quality severely

7
A First Cut Reduce Frame Rates
time
Snapshots of animation Davis et al. Eurographics
2003
8
A First Cut Reduce Frame Rates

• Current game design principles
• higher frame rates imply the better game quality
• Recent study on frame rates Claypool et al. MMCN
  2006
• very high frame rates are not necessary, very low
  frame rates impact the game quality severely
• Obvious question Can the CPU be run at a
  constant but lower frequency (to reduce the frame
  rate)?

9
A First Cut Reduce Frame Rates
desired frame rate
10
A First Cut Reduce Frame Rates

• Current game design
• the higher frame rate, the better game quality
• Recent study of frame rate
• very high frame rates are not necessary, very low
  frame rates impact the game quality severely
• Obvious question Can the CPU be run at a
  constant but lower frequency (to reduce the frame
  rate)?
• However, can DVFS algorithms developed for video
  decoding applications be applied to games?
• unpredictable workload because of the interaction
  in game
• no frame structure in video decoding

11
A First Cut Reduce Frame Rates
Average system-level power consumption for
different processor frequencies
12
The Case for DVFS Snapshots of Game Maps
Outer Base game map in Quake II
Installation game map in Quake II
13
The Case for DVFS Workload in Games

• Game workload includes computational workload and
  rendering workload
• correspondence with the complexity of the scene
• Software renderer performs geometry,
  rasterization and texture processing on CPU
• portable devices without graphics accelerator
• Software rasterization workload is the workload
  of rasterizing objects on the screen
• correspondence with the complexity of the scene

14
The Case for DVFS Workload in Games
15
The Case for DVFS Workload in Games
Scene complexity
16
The Case for DVFSWorkload as a Function of Scene
Complexity
17
The Case for DVFSWorkload Characterization

• Each frame constitutes of the following objects
• brush model construct the world space
• Alias model model characters like monsters,
  soldiers and weapons
• texture give the appearance of the brush model
• light map generate lighting effect
• particles model small debris from gun shots
• ...

18
The Case for DVFSWorkload Characterization
19
The Case for DVFSWorkload Characterization

• Each frame constitutes of the following objects
• brush model construct the world space
• Alias model model characters like monsters,
  soldiers and weapons
• texture give the appearance of the brush model
• light map generate lighting effect
• particles model small debris from gun shots
• ...

• Computing workload of a frame
• Determine workload incurred in rasterizing each
  object offline
• Determine the number of occurrences of each
  object online

20
The Case for DVFSBrush Model

• Brush model
• parameter the number of polygons
  constitutingthe brush model

21
The Case for DVFSBrush Model

• Brush model

22
The Case for DVFSAlias Model

• Alias model
• parameters
• the number of pixels of triangles
• opaque or alpha blending mode of skin texture

23
The Case for DVFSAlias Model

• Alias model

24
The Case for DVFSTexture, light map and particles

• Texture
• parameter the number of surfaces
• Light map
• parameter the number of surfaces
• Particles
• parameters the number of pixels of 3D points

Total rasterization workload cycles for brush
X brush models
cycles for Alias X
Alias models
cycles for texture X
textures
cycles for light map X light maps

cycles for particle X particles
25
DVFS for Games
Poll players message
Compute visible objects
Obtain workload parameters of each object
Game loop
Online predict workload
Compute required CPU frequency
Scale to required CPU frequency
Render the frame
26
DVFS for Games

• Data structures to maintain the correlation
  between the workload parameters and the
  corresponding rasterization workload of each type
  of object
• Linear regression model to find the fitting
  functions between the workload parameters and the
  corresponding rasterization workload for each
  type of object
• Initial experiments show significant system-level
  power saving with our proposed framework

27
Conclusion

• Explore the possibility of using DVFS for
  interactive games
• Propose the workload characterization for games
• Outline DVFS algorithms with our proposed
  workload characterization

28
Games are Up for DVFS

• Thank You!