Iterative Deletion Routing Algorithm

Similar documents
Power-Driven Flip-Flop p Merging and Relocation. Shao-Huan Wang Yu-Yi Liang Tien-Yu Kuo Wai-Kei Tsing Hua University

Power Reduction Approach by using Multi-Bit Flip-Flops

DUE to the popularity of portable electronic products,

Sample: A small part of a lot or sublot which represents the whole. A sample may be made up of one or more increments or test portions.

ORF 307: Lecture 14. Linear Programming: Chapter 14: Network Flows: Algorithms

Achieving Faster Time to Tapeout with In-Design, Signoff-Quality Metal Fill

PAGE HEADERS AND FOOTERS

A Fast Constant Coefficient Multiplier for the XC6200

16B CSS LAYOUT WITH GRID

Latch-Based Performance Optimization for FPGAs. Xiao Teng

Headend Systems. Series 9900 RF Signal Manager Modules. Introduction. Description

Lecture 23 Design for Testability (DFT): Full-Scan

Post-Routing Layer Assignment for Double Patterning

K.T. Tim Cheng 07_dft, v Testability

Overview: Logic BIST

VLSI Design: 3) Explain the various MOSFET Capacitances & their significance. 4) Draw a CMOS Inverter. Explain its transfer characteristics

QCTool. PetRos EiKon Incorporated

Low Power VLSI Circuits and Systems Prof. Ajit Pal Department of Computer Science and Engineering Indian Institute of Technology, Kharagpur

Lecture 23 Design for Testability (DFT): Full-Scan (chapter14)

11. Sequential Elements

Designing VeSFET-based ICs with CMOS-oriented EDA Infrastructure

Testability: Lecture 23 Design for Testability (DFT) Slide 1 of 43

The Art of Engineering

ISPD 2015 Detailed Routing-Driven Placement Contest with Fence Regions and Routing Blockages

Flip-flop Clustering by Weighted K-means Algorithm

Switching Circuits & Logic Design, Fall Final Examination (1/13/2012, 3:30pm~5:20pm)

Good Evening! Welcome!

Novel Pulsed-Latch Replacement Based on Time Borrowing and Spiral Clustering

Investigation of Look-Up Table Based FPGAs Using Various IDCT Architectures

Clock Tree Power Optimization of Three Dimensional VLSI System with Network

COPY RIGHT. To Secure Your Paper As Per UGC Guidelines We Are Providing A Electronic Bar Code

C8000. switch over & ducking

Data Sheet. HDSP-70xE 17.3 mm (0.68 inch) General Purpose 5x7 Dot Matrix Alphanumeric Displays

VISSIM Tutorial. Starting VISSIM and Opening a File CE 474 8/31/06

Cisco ONS Exposed Faceplate Mux/Demux 48-Channel Extended Bandwidth Patch Panel and Splitter Coupler Module

This paper is a preprint of a paper accepted by Electronics Letters and is subject to Institution of Engineering and Technology Copyright.

Introduction to CMOS VLSI Design (E158) Lab 3: Datapath and Zipper Assembly

Project 6: Latches and flip-flops

In-System Testing of Configurable Logic Blocks in Xilinx 7-Series FPGAs

EEC 116 Fall 2011 Lab #5: Pipelined 32b Adder

Sequencing. Lan-Da Van ( 范倫達 ), Ph. D. Department of Computer Science National Chiao Tung University Taiwan, R.O.C. Fall,

DC Ultra. Concurrent Timing, Area, Power and Test Optimization. Overview

CMOS VLSI Design. Lab 3: Datapath and Zipper Assembly

CIS Pre Test. Multiple Choice Identify the choice that best completes the statement or answers the question.

Sharif University of Technology. SoC: Introduction

ORF 307 Network Flows: Algorithms

CS 7643: Deep Learning

Chapter 12. Synchronous Circuits. Contents

imso-104 Manual Revised August 5, 2011

Design of Fault Coverage Test Pattern Generator Using LFSR

PCIe: EYE DIAGRAM ANALYSIS IN HYPERLYNX

Design for Testability

Clocking Spring /18/05

Design for Testability Part II

A Module Area Estimator for VLSI Layout*

Section 6.8 Synthesis of Sequential Logic Page 1 of 8

A Survey on Post-Placement Techniques of Multibit Flip-Flops

EL302 DIGITAL INTEGRATED CIRCUITS LAB #3 CMOS EDGE TRIGGERED D FLIP-FLOP. Due İLKER KALYONCU, 10043

Wire Delay and Switch Logic

Implementation of an MPEG Codec on the Tilera TM 64 Processor

MTO 22.1 Examples: Carter-Ényì, Contour Recursion and Auto-Segmentation

Harvatek International 2.0 5x7 Dot Matrix Display HCD-88442

Innovative Fast Timing Design

imso-104 Manual Revised July 19, 2012

BIST-Based Diagnostics of FPGA Logic Blocks

MATH& 146 Lesson 11. Section 1.6 Categorical Data

LED Display Product Data Sheet LTP-1557TBE Spec No.: DS Effective Date: 06/16/2010 LITE-ON DCC RELEASE

MVP: Capture-Power Reduction with Minimum-Violations Partitioning for Delay Testing

VISSIM TUTORIALS This document includes tutorials that provide help in using VISSIM to accomplish the six tasks listed in the table below.

High Performance Carry Chains for FPGAs

Australian Journal of Basic and Applied Sciences. Design of SRAM using Multibit Flipflop with Clock Gating Technique

Route optimization using Hungarian method combined with Dijkstra's in home health care services

A New 4MW LHCD System for EAST

Monolithic Amplifier GVA-60+ Flat Gain, High IP to 5 GHz. The Big Deal

Chapter 4. Displaying Quantitative Data. Copyright 2007 Pearson Education, Inc. Publishing as Pearson Addison-Wesley

Quantifying Academic Placer Performance on Custom Designs

Interconnect Planning with Local Area Constrained Retiming

Automatic Transistor-Level Design and Layout Placement of FPGA Logic and Routing from an Architectural Specification

CEDES. Functional description CEDES 7-Segment Double Digit Indicator (CAN) Functional description 7-Segment Double Digit Indicator (CAN)

Lecture 10: Sequential Circuits

Jin-Fu Li Advanced Reliable Systems (ARES) Laboratory. National Central University

The Effect of Wire Length Minimization on Yield

TKK S ASIC-PIIRIEN SUUNNITTELU

Model Features and Groups RDS-303 RDS RDS-304.2

Exercise 1-2. Digital Trunk Interface EXERCISE OBJECTIVE

OPTIMALITY AND STABILITY STUDY OF TIMING-DRIVEN PLACEMENT ALGORITHMS. Jason Cong, Michail Romesis, Min Xie

Computer Architecture and Organization

LoopBack Relay. SGLB363 Series. With Built-in AC Bypass Capacitors / DC LoopBack Relay

ESE534: Computer Organization. Previously. Today. Previously. Today. Preclass 1. Instruction Space Modeling

Running head: [SHORTENED TITLE UP TO 50 CHARACTERS] 1. [Title Here, up to 12 Words, on One to Two Lines]

Pattern Discovery and Matching in Polyphonic Music and Other Multidimensional Datasets

Installation instructions Roxtec RM ES systems

Amon: Advanced Mesh-Like Optical NoC

Traditional RF Splitter/Combiner and Directional Coupler User Manual

Advanced Synchronization Techniques for Data Acquisition

Statistics for Engineers

FinFETs & SRAM Design

Minimailer 4 OMR SPECIFICATION FOR INTELLIGENT MAILING SYSTEMS. 1. Introduction. 2. Mark function description. 3. Programming OMR Marks

Boolean, 1s and 0s stuff: synthesis, verification, representation This is what happens in the front end of the ASIC design process

Chapter 7 Memory and Programmable Logic

Transcription:

Iterative Deletion Routing Algorithm Perform routing based on the following placement Two nets: n 1 = {b,c,g,h,i,k}, n 2 = {a,d,e,f,j} Cell/feed-through width = 2, height = 3 Shift cells to the right, each cell contains self-feed-through Practical Problems in VLSI Physical Design Iterative Deletion Routing (1/18)

Feed-through Insertion Add one edge with min-weight at a time Continue until we form a spanning forest Our spanning forest needs 4+5 edges (why?) Use K = 0.5 Break ties in alphabetical order Place feed-throughs right below top gate Practical Problems in VLSI Physical Design Iterative Deletion Routing (2/18)

Feed-through Insertion (cont) First step: build net connection graph Union of individual complete graphs Practical Problems in VLSI Physical Design Iterative Deletion Routing (3/18)

Feed-through Insertion (cont) Edge weight computation w(a,d) = 2 + 0.5 0 = 2 w(c,i) = 13 + 0.5 (21 + 21) = 34 Practical Problems in VLSI Physical Design Iterative Deletion Routing (4/18)

Feed-through Insertion (cont) Sorted edge list (increasing order) Practical Problems in VLSI Physical Design Iterative Deletion Routing (5/18)

Iterative Addition Adding first 7 edges Based on increasing order of edge weight (should not form cycle) Edge weight changes if feed-through is added No feed-through is used for the first 7 edges, so no update Practical Problems in VLSI Physical Design Iterative Deletion Routing (6/18)

Iterative Addition (cont) Adding 8 th edge Choose (e,j): does not create a cycle Need a feed-through (= x) in third row (= R 3 ) Some edges will have new weights (details in next slide) Practical Problems in VLSI Physical Design Iterative Deletion Routing (7/18)

Iterative Addition (cont) Edge weight update after adding 8 th edge All edges intersecting with R 3 All edges connecting to cell h (because h is shifted) Practical Problems in VLSI Physical Design Iterative Deletion Routing (8/18)

Iterative Addition (cont) Adding 9 th (= last) edge Skip (d,f) (= creates a cycle), so add (c,h) Need a feed-through (= y) in R 2 Practical Problems in VLSI Physical Design Iterative Deletion Routing (9/18)

Iterative Addition (cont) Final Result Two feed-throughs are inserted: already have routing solutions Why do we need iterative deletion then? Improve congestion Practical Problems in VLSI Physical Design Iterative Deletion Routing (10/18)

Iterative Deletion Step 1: obtain simplified net connection graph Form cliques among pins in the same channel Remove edges that connect non-adjacent pins (= dotted lines) Practical Problems in VLSI Physical Design Iterative Deletion Routing (11/18)

Iterative Deletion (cont) Step 2: compute channel density (= congestion) Number of edges passing, beginning, or ending at each column Density of channel 1/2/3 is 4/6/2 (= max value) Practical Problems in VLSI Physical Design Iterative Deletion Routing (12/18)

Iterative Deletion (cont) Step 3: delete edges in G Continue until we obtain spanning forest of G Should not isolate any node Delete edges with max-weight first w(e) = d(e) / d(c e ) Break ties: delete edges With longer x-span first With higher edge density, d(e) From bottom-most channel Lexicographically Practical Problems in VLSI Physical Design Iterative Deletion Routing (13/18)

Iterative Deletion (cont) Deleting first edge Choose (x,f): does not isolate any node Density of channel 2 reduces to 5: weights of all edges in channel 2 to change Practical Problems in VLSI Physical Design Iterative Deletion Routing (14/18)

Iterative Deletion (cont) Edge weight update after deleting first edge all edges in channel 2 to change Practical Problems in VLSI Physical Design Iterative Deletion Routing (15/18)

Iterative Deletion (cont) Practical Problems in VLSI Physical Design Iterative Deletion Routing (16/18)

Iterative Deletion (cont) Final result Practical Problems in VLSI Physical Design Iterative Deletion Routing (17/18)

Iterative Addition vs Deletion Density of channel (= congestion) improved Reduced from 3 to 2 in channel 1 Practical Problems in VLSI Physical Design Iterative Deletion Routing (18/18)