Carnegie Mellon A System Design And Build Course On Wearable Computers Dan Siewiorek and Asim Smailagic Carnegie Mellon University MSE ‘01, Las Vegas,

Slides:



Advertisements
Similar presentations
HARDWARE Rashedul Hasan..
Advertisements

VADA Lab.SungKyunKwan Univ. 1 L3: Lower Power Design Overview (2) 성균관대학교 조 준 동 교수
Daniel Schall, Volker Höfner, Prof. Dr. Theo Härder TU Kaiserslautern.
Introduction Computer Hardware Jess 2006 EXPANSION CARDS BUS ARCHITECTURE AND CONNECTORS.
The physical parts of Computer
Basic Input Output System
Parts & Functions of a Computer. 2 Functions of a Computer.
CMU Wearable Computers and Pervasive Computing Asim Smailagic Institute for Complex Engineered Systems Carnegie Mellon June 28, 2001.
IT Infrastructure: Software September 18, LEARNING GOALS Identify the different types of systems software. Explain the main functions of operating.
Lesson 15 – INSTALL AND SET UP NETWARE 5.1. Understanding NetWare 5.1 Preparing for installation Installing NetWare 5.1 Configuring NetWare 5.1 client.
Exploring The Green Blade Ken Lutz University of California, Berkeley LoCal Retreat, June 8, 2009.
Optimization Of Power Consumption For An ARM7- BASED Multimedia Handheld Device Hoseok Chang; Wonchul Lee; Wonyong Sung Circuits and Systems, ISCAS.
Chapter 4 The Components of the System Unit
Intel ® Research mote Ralph Kling Intel Corporation Research Santa Clara, CA.
UNIT 9 Computer architecture
Embedded Systems Programming Introduction to the course.
Introduction to Embedded Development. What is an Embedded System ? An embedded system is a computer system embedded in a device with a dedicated function.
A Flexible Architecture for Simulation and Testing (FAST) Multiprocessor Systems John D. Davis, Lance Hammond, Kunle Olukotun Computer Systems Lab Stanford.
Prepared by Careene McCallum-Rodney Hardware specification of a computer system.
TM Freescale, the Freescale logo, AltiVec, C-5, CodeTest, CodeWarrior, ColdFire, C-Ware, mobileGT, PowerQUICC, StarCore, and Symphony are trademarks of.
Handheld Devices (portable but still explicit usage) Laptops Personal Digital Assistants (Palm, PocketPC) TabletPC Smart Phones.
XP Practical PC, 3e Chapter 17 1 Upgrading and Expanding your PC.
Basic Computer Structure and Knowledge Project Work.
Chapter 5 Basic Input/Output System (BIOS)
The Components of the System Unit Chapter 4 By: Janice Colon.
1 A+ Guide to Managing and Maintaining Your PC, Fifth Edition Hardware Needs Software to Work Hardware  Physical components of a computer  Visible part.
Understanding Computers, Ch.31 Chapter 3 The System Unit: Processing and Memory.
Chipset Introduction The chipset is commonly used to refer to a set of specialized chips on a computer's motherboard or.
A+ Guide to Managing and Maintaining Your PC Fifth Edition Chapter 1 Introducing Hardware.
© Janice Regan, CMPT 300, May CMPT 300 Introduction to Operating Systems Principles of I/0 hardware.
McGraw-Hill Technology Education © 2006 by the McGraw-Hill Companies, Inc. All rights reserved. 66 CHAPTER THE SYSTEM UNIT.
REXAPP Bilal Saqib. REXAPP  Radio EXperimentation And Prototyping Platform Based on NOC  REXAPP Compiler.
Mayuresh Varerkar ECEN 5613 Current Topics Presentation March 30, 2011.
EWatch: A Wearable Sensor and Notification Platform Paper By: Uwe Maurer, Anthony Rowe, Asim Smailagic, Daniel P. Siewiorek Presenter: Ke Gao.
Sogang University Advanced Computing System Chap 1. Computer Architecture Hyuk-Jun Lee, PhD Dept. of Computer Science and Engineering Sogang University.
The Mechanics Of Computers The Operating System (OS) & Hardware.
Discovering Computers Fundamentals, Third Edition CGS 1000 Introduction to Computers and Technology Spring 2007.
Lesson 3 — How a Computer Processes Data Unit 1 — Computer Basics.
Copyright 2003 The McGraw-Hill Companies, Inc CHAPTER The System Unit computing ESSENTIALS    
Chapter 1 Computer Abstractions and Technology. Chapter 1 — Computer Abstractions and Technology — 2 The Computer Revolution Progress in computer technology.
Types of computers Hardware. 8/3/12 Hardware - the tangible, physical parts of the computer which work together to input, process, store and output data.
A Systematic Approach to the Design of Distributed Wearable Systems Urs Anliker, Jan Beutel, Matthias Dyer, Rolf Enzler, Paul Lukowicz Computer Engineering.
CHAPTER Microcomputer as a Communication Device. Chapter Objectives Examine the components of the motherboard that relate to communication Describe a.
Survey of Program Compilation and Execution Bangor High School Ali Shareef 2/28/06.
Morgan Kaufmann Publishers
Computer Hardware. Lally School of M&T- Microcomputing and Info Systems Lecture Topics 1. Data Representation 2. Data Metrics 3. Central processing Unit.
Introduction to Computers Rabie A. Ramadan, PhD. 2 Class Information Website ses/2011/summer/intro/
CompTIA A+ Guide to Managing & Maintaining Your PC By: JEAN ANDREW Computer Maintenance Chapter 1: Computer Basics Knowledge.
BMTS 242: Computer and Systems Lecture 4: Computer Hardware and Ports Yousef Alharbi Website
Module 12: Configuring Windows XP Professional for Mobile Computing.
JouleTrack - A Web Based Tool for Software Energy Profiling Amit Sinha and Anantha Chandrakasan Massachusetts Institute of Technology June 19, 2001.
MOBILE DEVICES AND COMPUTER BUS Batch - HYD10 / 1516 Learning Group - HIS 45 Team Nagashravani K L Anirban Bhattacharjee
CS203 – Advanced Computer Architecture
SEPTEMBER 8, 2015 Computer Hardware 1-1. HARDWARE TERMS CPU — Central Processing Unit RAM — Random-Access Memory  “random-access” means the CPU can read.
COMPUTER SCIENCE AND THE FOUNDATION OF KNOWLEDGE NURSING INFORMATICS CHAPTER 5 1.
Embedded Software Design Week III Processor Basics Raspberry Pi -> Blinking LEDs & pushing buttons.
Evaluation of Advanced Power Management for ClassCloud based on DRBL Rider Grid Technology Division National Center for High-Performance Computing Research.
Identify internal hardware devices (e. g
Computer systems is a 10-credit unit
I/O SYSTEMS MANAGEMENT Krishna Kumar Ahirwar ( )
UNIT 9 Computer architecture
Atoll Solutions PVT LTD
Green cloud computing 2 Cs 595 Lecture 15.
Components of Computer
CS111 Computer Programming
Realizing the potential of mobile devices as experimental devices: Human computer interface and performance considerations Chiung Ching Ho & C. Eswaran.
Green Software Engineering Prof
المحور 3 : العمليات الأساسية والمفاهيم
Journey: Introduction to Embedded Systems
Presentation transcript:

Carnegie Mellon A System Design And Build Course On Wearable Computers Dan Siewiorek and Asim Smailagic Carnegie Mellon University MSE ‘01, Las Vegas, June 2001

Carnegie Mellon Overview 1.Introduction 2.Approach 3.Electronic Design in a Multidisciplinary Project 4.Design Methodology 5.Power Measurements 6.Evaluation 7.Conclusions

Carnegie Mellon Introduction A System - Level design approach to the power and performance of CMU’s wearable computers dedicated to speech processing - the Speech Translator Smart Modules Power consumption has to be considered in all phases of system design

Carnegie Mellon Introduction We examine the impact of processor speed, memory size, and type of secondary storage on power consumption and performance We see ahead a world of near-zero energy / weight / cost mobile systems

Carnegie Mellon Approach Experimental framework includes a family of CMU wearable computers dedicated to speech processing - Smart Modules They perform speech recognition, language translation, and speech synthesis

Carnegie Mellon

Multidisciplinary Project Five major factors in a portable electronic system include: Functionality User Interface Physical Form Factor Power Sensors

Carnegie Mellon Multidisciplinary Project Decisions made in one design discipline affect decisions in another discipline The impact can be measured if the cost of the design decision can be reduced to a common “currency” In mobile electronic systems that “currency” is power consumption

Carnegie Mellon Major Factors in Portable Electronic Systems and their Relationship to the Design Disciplines

Carnegie Mellon Smart Module Architecture The core of the smart module is the Cardio processor card, combining the processor and motherboard chips The Cardio also supports two serial ports for communication between the modules and a VGA interface

Carnegie Mellon Smart Module Hardware Diagram CARDIO -Processor -Memory -Chipsets Keyboard Mouse VGA PCMCIA HD ESS 1888 IDEISA Serial Ports Communications To other Modules (For debugging only)

Carnegie Mellon Smart Module Functional Prototype

Carnegie Mellon Optimized Smart Module

Carnegie Mellon Approach The code was profiled and tuned Profiling identified “hot spots” for HW and SW acceleration, and places to reduce computation and storage requirements

Carnegie Mellon Power Measurements Power is consumed by Processor Memory Disk Sound Chip Serial Ports

Carnegie Mellon Power Measurements The power and performance measurements were taken using a body of 10 English test sentences for the Speech Recognizer and Language Translator, and their 10 Croatian translations for the Speech Synthesizer

Carnegie Mellon Power Measurements A power profile for the Smart Modules included: Idle Mode: processor is at nearly 0% usage Full On Mode: processor is at nearly 100% usage Each state transition has an associated latency value

Carnegie Mellon Power Measurements States for spinning disk drive: Power Down Mode: power to the disk drive is shut off - when the module is in Suspend Mode High Spin Mode: disk drive is being accessed Low Spin Mode: power-saving mode

Carnegie Mellon Power Measurements State diagrams for both processor and disk were combined to produce a new model for power consumption

Carnegie Mellon New Model of Power Consumption Over Time Power (W) Time Suspend Idle & High Spin Full on & High Spin Idle & Low Spin Full On & Low Spin

Carnegie Mellon Power Consumption Profile Power consumption of the Speech Recognizer module over time, using the 586-based Cardio and a spinning disk drive, was measured

Carnegie Mellon Power Consumption of Speech Recognition Module Over Time Suspend Spin Up Full Spin Spin Down Spin Up Spin Down Close TCP Link & Spin Up

Carnegie Mellon

Performance Comparison The metric for comparison is proportional to the processing power, and inversely proportional to the product of volume, weight and power consumption

Carnegie Mellon NameSpecInt Volume (in 3 ) Weight (lbs)Power (watts)R (V*W*P)Log of Normalized TI TIA-P TIA-O SR-SM OPT-SM Performance Values for Wearable Computers

Carnegie Mellon Composite Performance of Speech Recognition Wearable Computers Evolution of Systems Norm Log [perf/(vol*wt*power)] Normalized Performance Goal: 2.8

Carnegie Mellon Spot: SA-1110 Based Wearable Computer New wearable computer as a research platform, including context aware computing Low power design 233 MHz StrongARM MB DRAM Mobile, throttleable platform that can trade performance for energy

Carnegie Mellon Spot Specifications

Carnegie Mellon Spot Wearable Computer

Carnegie Mellon Spot

Carnegie Mellon Conclusions Results show that there are orders of magnitude improvement from one generation of wearable computers to the next A system-level approach to power / performance optimization improved the metric by over a factor of 300 through the four generations, and over 400 through the five generations

Carnegie Mellon Conclusions Peak demand by an application can often determine the battery life rather than average demand Audio-centric interfaces exhibit high demand “spikes,” potentially causing significantly reduced battery life

Carnegie Mellon Summary The options for the design space included: –Two processor architectures (Intel 486,586) –Three different processor speeds –Two different main memory sizes –Two different secondary storage types The best configuration for performance is the 586/100/32 Flash, and the 486/75/16 Flash for power

Carnegie Mellon Power Consumption of Speech Recognition Module Our research indicates that the peak demand of an application can often determine the battery life rather than the average demand Audio-centric interfaces exhibit high demand spikes By filling in the valleys, it would be possible to cut the peak demand in half and thus significantly extend battery life