Data Wordlength Reduction for Low-Power Signal Processing Software

Slides:



Advertisements
Similar presentations
FINITE WORD LENGTH EFFECTS
Advertisements

CENG536 Computer Engineering department Çankaya University.
Digital to Analog and Analog to Digital Conversion
Aug 23, ‘021Low-Power Design Minimum Dynamic Power Design of CMOS Circuits by Linear Program Using Reduced Constraint Set Vishwani D. Agrawal Agere Systems,
Fixed-Point Arithmetics: Part I
Computer ArchitectureFall 2008 © August 25, CS 447 – Computer Architecture Lecture 3 Computer Arithmetic (1)
+ CS 325: CS Hardware and Software Organization and Architecture Integers and Arithmetic Part 4.
Computer ArchitectureFall 2007 © August 29, 2007 Karem Sakallah CS 447 – Computer Architecture.
Digital System Design EEE344 Lecture 3 Introduction to Verilog HDL Prepared by: Engr. Qazi Zia, Assistant Professor EED, COMSATS Attock1.
+ CS 325: CS Hardware and Software Organization and Architecture Exam 1: Study Guide.
ELEN 5346/4304 DSP and Filter Design Fall Lecture 12: Number representation and Quantization effects Instructor: Dr. Gleb V. Tcheslavski Contact:
Data Representation – Binary Numbers
Computer Arithmetic Nizamettin AYDIN
Chapter 6-2 Multiplier Multiplier Next Lecture Divider
Fixed-Point Arithmetics: Part II
Data Acquisition Systems
07/19/2005 Arithmetic / Logic Unit – ALU Design Presentation F CSE : Introduction to Computer Architecture Slides by Gojko Babić.
1 Sign Bit Reduction Encoding for Low Power Applications Hsin-Wei Lin Saneei, M. Afzali-Kusha, A. and Navabi, Z. Sign Bit Reduction Encoding for Low Power.
1 CS103 Guest Lecture Number Systems & Conversion Bitwise Logic Operations.
Low-Power Multipliers with Data Wordlength Reduction Kyungtae Han Brian L. Evans Earl E. Swartzlander, Jr.
ESPL 1 Wordlength Optimization with Complexity-and-Distortion Measure and Its Application to Broadband Wireless Demodulator Design Kyungtae Han and Brian.
COE 202 Introduction to Verilog Computer Engineering Department College of Computer Sciences and Engineering King Fahd University of Petroleum and Minerals.
Digital to Analog Converters (DAC) 1 Technician Series ©Paul Godin March 2015.
MECH1500 Chapter 3.
CS 151: Digital Design Chapter 4: Arithmetic Functions and Circuits
DEPARTMENTT OF ECE TECHNICAL QUIZ-1 AY Sub Code/Name: EC6502/Principles of digital Signal Processing Topic: Unit 1 & Unit 3 Sem/year: V/III.
Number Representation and Arithmetic Circuits
Seok-jae, Lee VLSI Signal Processing Lab. Korea University
Low Power IP Design Methodology for Rapid Development of DSP Intensive SOC Platforms T. Arslan A.T. Erdogan S. Masupe C. Chun-Fu D. Thompson.
Data Word Length Reduction for Low- Power DSP Software Kyungtae Han March 24, 2004.
Digital Systems and Information Dr. Sanjay P. Ahuja, Ph.D. FIS Distinguished Professor of CIS ( ) School of Computing, UNF.
1 Summary Lecture: Part 1 Sensor Readout Electronics and Data Conversion Discovering Sensor Networks: Applications in Structural Health Monitoring.
William Stallings Computer Organization and Architecture 8th Edition
CHAPTER 5: Representing Numerical Data
Unit IV Finite Word Length Effects
Unit 1 Introduction Number Systems and Conversion.
Addition and Subtraction
Chapter 2: Integers and Mathematical Induction
Introduction Molecules: made up of atoms of individual elements.
Integer Representations and Arithmetic
Number System conversions
Multipliers Multipliers play an important role in today’s digital signal processing and various other applications. The common multiplication method is.
William Stallings Computer Organization and Architecture 7th Edition
DESIGN AND IMPLEMENTATION OF DIGITAL FILTER
James K Beard, Ph.D. April 20, 2005 SystemView 2005 James K Beard, Ph.D. April 20, 2005 April 122, 2005.
Lecture 10: Floating Point, Digital Design
Automating Transformations from Floating Point to Fixed Point for Implementing Digital Signal Processing Algorithms Kyungtae Han Ph.D. Defense Committee.
CDA 3101 Spring 2016 Introduction to Computer Organization
Introduction to Scientific Computation
CDA 3101 Summer 2007 Introduction to Computer Organization
ECEG-3202 Computer Architecture and Organization
On-line arithmetic for detection in digital communication receivers
Finite Wordlength Effects
DEPARTMENT OF INFORMATION TECHNOLOGY DIGITAL SIGNAL PROCESSING UNIT 4
Chapter 8 Computer Arithmetic
University of Texas at Austin
C Model Sim (Fixed-Point) -A New Approach to Pipeline FFT Processor
Automatic Floating-Point to Fixed-Point Transformations
DEPARTMENT OF INFORMATION TECHNOLOGY DIGITAL SIGNAL PROCESSING UNIT 4
Binary Numbers Material on Data Representation can be found in Chapter 2 of Computer Architecture (Nicholas Carter) CSIT 301 (Blum)
Introduction SYSC5603 (ELG6163) Digital Signal Processing Microprocessors, Software and Applications Miodrag Bolic.
Introduction SYSC5603 (ELG6163) Digital Signal Processing Microprocessors, Software and Applications Miodrag Bolic.
On-line arithmetic for detection in digital communication receivers
COE 202 Introduction to Verilog
Presentation transcript:

Data Wordlength Reduction for Low-Power Signal Processing Software Kyungtae Han May 3rd, 2004 The University of Texas at Austin May 3, 2004

Low-Power at Software Level Portable wireless computing demands minimizing power dissipation due to limited power Minimizing power consumption Reduce supply voltage Decrease switching activity Software can reduce power consumption Ordering of operation Changing of number representation Reducing of data wordlength May 3, 2004

Objectives and Problems Power analysis of digital signal processing blocks for data word length in low-power software Software level power minimization with reducing data wordlength while not changing the hardware structures Problems How much can shorter data wordlength reduce power consumption? Need formulation of power consumption according to reduced data wordlength for optimum design May 3, 2004

Power Consumption Average power consumption Switching power consumption May 3, 2004

Multiply Unit for Digital Wireless Transceivers Multiply unit is usually a major source of power consumption in typical DSP applications Many digital blocks are required multiply unit for digital wireless communications Digital filters, equalizers, FFT/IFFT, digital down/up converter, etc. Multiplier core in DSP TI 64 uses a Wallace tree with 3-2 compressors TI 62 uses a Booth multiplier May 3, 2004

Data Wordlength Reduction Fixes the size of multiplication Decreases input data wordlength Signed right shift (Least significant bit side) Move toward LSB (i.e. left shift) Arithmetic right shift (if negative) Truncation (Most significant bit side) Make LSB side to be zero values May 3, 2004

Example of Array Multiplier Transition Counts Array multiplier with data flow model 16 x16 bit array multiplier 10,000 random data Node transitions counted Without delay information Data Flow Modeling 16 x 16 bit array multiplier transition counts Input data Max Min Std Mean 16 bits 3152 383 319 1863 May 3, 2004

Wallace Multiplier Transition Counts Simulation Discrete Event Simulation: Verilog Compiler Simulator Average transition counts of all gate output All gates are assumed to have a unit gate delay Variable size Wallace multiplier Wallace Size (bits) 8 16 32 Avg. transition counts 670 4177 22021 May 3, 2004

Wallace Multiplier Transition Counts Fixed-size Wallace multiplier (32 x 32 bits) n2 rate reduction in truncation method Average transition counts for data wordlength reduction Input wordlength Truncation Signed Right Shift 32 22021 16 6138 28634 8 1383 31364 4 404 32996 May 3, 2004

Conclusion and Future Work Data wordlength reduction in MSB side decreases transition counts and power consumption In Wallace multiplier, transition counts decrease n2 rate Future work Find optimum wordlength for power and precision Comparison with Booth Radix-4 Apply for digital signal processing blocks Wordlength(n) Power (p) Error (d) [1/performance] Optimum wordlength May 3, 2004