Tsung-Hao Chen and Kuang-Ching Wang May

Slides:



Advertisements
Similar presentations
Mani Srivastava UCLA - EE Department Room: 6731-H Boelter Hall Tel: WWW: Copyright 2003.
Advertisements

ADSP Lecture2 - Unfolding VLSI Signal Processing Lecture 2 Unfolding Transformation.
* Bellman-Ford: single-source shortest distance * O(VE) for graphs with negative edges * Detects negative weight cycles * Floyd-Warshall: All pairs shortest.
Courtesy RK Brayton (UCB) and A Kuehlmann (Cadence) 1 Logic Synthesis Sequential Synthesis.
Chapter 4 Retiming.
Part2 AI as Representation and Search
VLSI Communication SystemsRecap VLSI Communication Systems RECAP.
FPGA Latency Optimization Using System-level Transformations and DFG Restructuring Daniel Gomez-Prado, Maciej Ciesielski, and Russell Tessier Department.
ECE734 VLSI Arrays for Digital Signal Processing Algorithm Representations and Iteration Bound.
Pipelining and Retiming 1 Pipelining  Adding registers along a path  split combinational logic into multiple cycles  increase clock rate  increase.
The max flow problem
CS294-6 Reconfigurable Computing Day 16 October 15, 1998 Retiming.
A Tool for Partitioning and Pipelined Scheduling of Hardware-Software Systems Karam S Chatha and Ranga Vemuri Department of ECECS University of Cincinnati.
VLSI DSP 2008Y.T. Hwang3-1 Chapter 3 Algorithm Representation & Iteration Bound.
Algorithmic Transformations
EDA (CS286.5b) Day 18 Retiming. Today Retiming –cycle time (clock period) –C-slow –initial states –register minimization.
Penn ESE535 Spring DeHon 1 ESE535: Electronic Design Automation Day 8: February 13, 2008 Retiming.
Lecture 9: Structure for Discrete-Time System XILIANG LUO 2014/11 1.
Using Dijkstra’s Algorithm to Find a Shortest Path from a to z 1.
CSC401: Analysis of Algorithms CSC401 – Analysis of Algorithms Chapter Dynamic Programming Objectives: Present the Dynamic Programming paradigm.
1 The Floyd-Warshall Algorithm Andreas Klappenecker.
CALTECH CS137 Winter DeHon CS137: Electronic Design Automation Day 7: February 3, 2002 Retiming.
Algorithms for the optimum communication spanning tree problem Prabha Sharma Ann Oper Res (2006) 143: /5/29 Presenter : Wei-Ni Chu, Chuan-Ju.
Dr. Elwin Chandra Monie Department of ECE, RMK Engineering College
ELEC692 VLSI Signal Processing Architecture Lecture 3
1 Chapter Equivalence, Order, and Inductive Proof.
Pipelining and Retiming
OPTIMIZING DSP SCHEDULING VIA ADDRESS ASSIGNMENT WITH ARRAY AND LOOP TRANSFORMATION Chun Xue, Zili Shao, Ying Chen, Edwin H.-M. Sha Department of Computer.
1 Motion Fuzzy Controller Structure(1/7) In this part, we start design the fuzzy logic controller aimed at producing the velocities of the robot right.
1 VLSI Algorithm & Computing Structures Chapter 1. Introduction to DSP Systems Younglok Kim Dept. of Electrical Engineering Sogang University Spring 2007.
Chapter 4 Structures for Discrete-Time System Introduction The block diagram representation of the difference equation Basic structures for IIR system.
Shortest Paths.
DSP Design – Lecture 7 Unfolding cont. & Folding Fredrik Edman fredrik
Vishwani D. Agrawal James J. Danaher Professor
Shortest Paths and Minimum Spanning Trees
Automatic Control Theory CSE 322
Adaptive Filters Common filter design methods assume that the characteristics of the signal remain constant in time. However, when the signal characteristics.
CS137: Electronic Design Automation
Algebra II Explorations Review ( )
By: Mohammadreza Meidnai Urmia university, Urmia, Iran Fall 2014
102-1 Under-Graduate Project Techniques in VLSI design
Pipelined Adaptive Filters
Equalization in a wideband TDMA system
James D. Z. Ma Department of Electrical and Computer Engineering
Digital Filter Design Tools
ESE535: Electronic Design Automation
Subject Name: Information Theory Coding Subject Code: 10EC55
لجنة الهندسة الكهربائية
Lecture 7 All-Pairs Shortest Paths
Modeling and Simulation NETW 707
ESE535: Electronic Design Automation
101-1 Under-Graduate Project Techniques in VLSI design
Multiplier-less Multiplication by Constants
Shortest Path Algorithms
Shortest Paths and Minimum Spanning Trees
2.2 Shortest Paths Def: directed graph or digraph
CS184a: Computer Architecture (Structures and Organization)
All pairs shortest path problem
ESE535: Electronic Design Automation
Signal Processing First
Algorithms (2IL15) – Lecture 7
EE5900 Advanced Embedded System For Smart Infrastructure
A Dynamic System Analysis of Simultaneous Recurrent Neural Network
Trevor Brown DC 2338, Office hour M3-4pm
Zhongguo Liu Biomedical Engineering
UNINFORMED SEARCH -BFS -DFS -DFIS - Bidirectional
Lecture 22 IIR Filters: Feedback and H(z)
Implementation of a De-blocking Filter and Optimization in PLX
Murugappan Senthilvelan May 4th 2004
All Pairs Shortest Path Examples While the illustrations which follow only show solutions from vertex A (or 1) for simplicity, students should note that.
Presentation transcript:

Tsung-Hao Chen and Kuang-Ching Wang May 10 2002 Spring 2002 ECE 734 Course Project An Integrated Loop Transformation and Optimization Tool for FIR/IIR filters Tsung-Hao Chen and Kuang-Ching Wang May 10 2002

Introduction Single-input single output FIR/IIR filters Generalized z-transformed polynomial representation Integrated tool suite for loop retiming, pipelining, unfolding, and look ahead transformation. GUI for configuration and presentation.

Project Scope

Direct Form II IIR/FIR Filters Matrix representation for weighted (delays) directed graph

Algorithms Loop retiming Pipelining Unfolding transformation [Parhi] Critical path analysis – modified Bellman Ford algorithm for longest path search. Iteration bound analysis – Longest Path Matrix (LPM) algorithm and DFG transformation. Retiming algorithm [Parhi] with Floyd Warshall shortest path algorithm Pipelining Delay injection and retiming Retiming algorithm’s [Parhi] limitation: allows only forward retiming. Unfolding transformation [Parhi] Look-ahead transformation Direct look-ahead transformation [Parhi]

Delay Injection and Retiming Feed forward cutsets on right leg only. Cutset pipelining can be done with delay injection in all entering branches of output node. Delays are expected to be pushed back (away from output). Retiming algorithm [Parhi] has limitation; delays can only be pushed toward output.

Retiming algorithm limitation Shortest path lengths must be less than or equal to zero. [Parhi] convention: r(i) is retiming value for node I, denoting the number of delays pushed backward. R(i) is never positive implies: solution space for retiming is not fully exploited.

Graphical User Interface A 2nd order IIR example.

Retiming Retiming reduces critical path to as little as iteration bound. In this example, critical path reduces from 7 to 4.

Pipelining for IIR Critical path limited by feedback loops (4). Feed-forward cutset retiming applicable on right leg only. Pipelining provides no improvement for IIR

Pipelining for FIR Critical path reduced to minimum operation latency.

Unfolding Filter DFG as shown. As stated in [Parhi], critical path can be reduced to iteration bound with unfolding alone. Observation suggests: not true. Retiming must be applied to achieve iteration bound. For IIR/FIR, retiming alone achieves iteration bound.

Look-ahead Transformation Filter DFG as shown. Filter coefficients calculated for reference.

Thank you.