Logic synthesis flow Technology independent mapping –Two level or multilevel optimization to optimize a coarse metric related to area/delay Technology.

Slides:



Advertisements
Similar presentations
Gregory Shklover, Ben Emanuel Intel Corporation MATAM, Haifa 31015, Israel Simultaneous Clock and Data Gate Sizing Algorithm with Common Global Objective.
Advertisements

Figure 4.1. The function f (x1, x2, x3) =  m(0, 2, 4, 5, 6).
ELEC Digital Logic Circuits Fall 2014 Logic Synthesis (Chapters 2-5) Vishwani D. Agrawal James J. Danaher Professor Department of Electrical and.
Buffer and FF Insertion Slides from Charles J. Alpert IBM Corp.
ELEN 468 Lecture 261 ELEN 468 Advanced Logic Design Lecture 26 Interconnect Timing Optimization.
Improving Placement under the Constant Delay Model Kolja Sulimma 1, Ingmar Neumann 1, Lukas Van Ginneken 2, Wolfgang Kunz 1 1 EE and IT Department University.
FPGA Technology Mapping Dr. Philip Brisk Department of Computer Science and Engineering University of California, Riverside CS 223.
Logical Effort A Method to Optimize Circuit Topology Swarthmore College E77 VLSI Design Adem Kader David Luong Mark Piper December 6, 2005.
ECE 667 Synthesis and Verification of Digital Systems
Circuit Retiming with Interconnect Delay CUHK CSE CAD Group Meeting One Evangeline Young Aug 19, 2003.
➢ Performing Technology Mapping and Optimization by DAG Covering: A Review of Traditional Approaches Evriklis Kounalakis.
Modern VLSI Design 2e: Chapter4 Copyright  1998 Prentice Hall PTR.
Technology Mapping.
Technology Mapping 2 Outline Goal Reading Tree Pattern Matching
EDA (CS286.5b) Day 2 Covering. Why covering now? Nice/simple cost model problem can be solved well (somewhat clever solution) general/powerful technique.
Penn ESE535 Spring DeHon 1 ESE535: Electronic Design Automation Day 2: January 23, 2008 Covering.
Penn ESE535 Spring DeHon 1 ESE535: Electronic Design Automation Day 16: March 23, 2009 Covering.
Technology Mapping 1 Outline –What is Technology Mapping? –Rule-Based Mapping –Tree Pattern Matching Goal –Understand technology mapping –Understand mapping.
COE 561 Digital System Design & Synthesis Library Binding Dr. Muhammad Elrabaa Computer Engineering Department King Fahd University of Petroleum & Minerals.
Charles Kime & Thomas Kaminski © 2008 Pearson Education, Inc. (Hyperlinks are active in View Show mode) Chapter 3 – Combinational Logic Design Part 1 –
ELEN 468 Lecture 271 ELEN 468 Advanced Logic Design Lecture 27 Interconnect Timing Optimization II.
Reinventing The Wheel: Developing a New Standard-Cell Synthesis Flow Alan Mishchenko University of California, Berkeley.
 2000 M. CiesielskiPTL Synthesis1 Synthesis for Pass Transistor Logic Maciej Ciesielski Dept. of Electrical & Computer Engineering University of Massachusetts,
Technology Mapping Outline Goal What is Technology Mapping?
FPGA Technology Mapping. 2 Technology mapping:  Implements the optimized nodes of the Boolean network to the target device library.  For FPGA, library.
ECE Synthesis & Verification, Lecture 17 1 ECE 697B (667) Spring 2006 ECE 697B (667) Spring 2006 Synthesis and Verification of Digital Systems Technology.
1 A Method for Fast Delay/Area Estimation EE219b Semester Project Mike Sheets May 16, 2000.
Overview Part 1 – Design Procedure 3-1 Design Procedure
May 2012Dynamic Programming1 Dynamic Programming and Some VLSI CAD Applications Shmuel Wimer Bar Ilan Univ. Eng. Faculty Technion, EE Faculty.
1 VLSI CAD Flow: Logic Synthesis, Lecture 13 by Ajay Joshi (Slides by S. Devadas)
Modern VLSI Design 4e: Chapter 4 Copyright  2008 Wayne Wolf Topics n Interconnect design. n Crosstalk. n Power optimization.
Digitaalsüsteemide verifitseerimise kursus1 Formal verification: BDD BDDs applied in equivalence checking.
Charles Kime & Thomas Kaminski © 2004 Pearson Education, Inc. Terms of Use (Hyperlinks are active in View Show mode) Terms of Use Lecture 12 – Design Procedure.
VLSI Backend CAD Konstantin Moiseev – Intel Corp. & Technion Shmuel Wimer – Bar Ilan Univ. & Technion.
An Efficient Clustering Algorithm For Low Power Clock Tree Synthesis Rupesh S. Shelar Enterprise Microprocessor Group Intel Corporation, Hillsboro, OR.
Shantanu Dutt ECE Dept. UIC
Modern VLSI Design 3e: Chapter 4 Copyright  1998, 2002 Prentice Hall PTR Topics n Combinational network delay. n Logic optimization.
Logical Effort and Transistor Sizing Digital designs are usually expected to operate at high frequencies, thus designers often have to choose the fastest.
Introduction to CMOS VLSI Design Lecture 5: Logical Effort GRECO-CIn-UFPE Harvey Mudd College Spring 2004.
Modern VLSI Design 3e: Chapter 4 Copyright  1998, 2002 Prentice Hall PTR Topics n Interconnect design. n Crosstalk. n Power optimization.
Topics Combinational network delay.
FPGA-Based System Design: Chapter 3 Copyright  2004 Prentice Hall PTR Circuit design for FPGAs n Static CMOS gate vs. LUT n LE output drivers n Interconnect.
ICS 252 Introduction to Computer Design Lecture 12 Winter 2004 Eli Bozorgzadeh Computer Science Department-UCI.
Technology Mapping. 2 Technology mapping is the phase of logic synthesis when gates are selected from a technology library to implement the circuit. Technology.
Curtis A. Nelson 1 Technology Mapping of Timed Circuits Curtis A. Nelson University of Utah September 23, 2002.
Courtesy RK Brayton (UCB) and A Kuehlmann (Cadence) 1 Logic Synthesis Multi-Level Logic Synthesis.
Give qualifications of instructors: DAP
Modern VLSI Design 4e: Chapter 4 Copyright  2008 Wayne Wolf Topics n Combinational network delay. n Logic optimization.
Static Timing Analysis
ELEC Digital Logic Circuits Fall 2015 Logic Synthesis (Chapters 2-5) Vishwani D. Agrawal James J. Danaher Professor Department of Electrical and.
1 CS151: Digital Design Chapter 3: Combinational Logic Design 3-2 Beginning Hierarchical Design 3-3 Technology Mapping.
1 Timing Closure and the constant delay paradigm Problem: (timing closure problem) It has been difficult to get a circuit that meets delay requirements.
CALTECH CS137 Fall DeHon 1 CS137: Electronic Design Automation Day 2: September 28, 2005 Covering.
Inexact and Approximate Circuits for Error Tolerant Applications IcySoc RTD 2013 Jérémy Schlachter, Vincent Camus, Christian Enz Ecole polytechnique fédérale.
Reducing Structural Bias in Technology Mapping
Overview Part 1 – Design Procedure Part 2 – Combinational Logic
Delay Optimization using SOP Balancing
Standard-Cell Mapping Revisited
Alan Mishchenko University of California, Berkeley
CSE 370 – Winter Combinational Implementation - 1
ELEC 7770 Advanced VLSI Design Spring 2014 Technology Mapping
ELEC 7770 Advanced VLSI Design Spring 2016 Technology Mapping
Sungho Kang Yonsei University
ECE 667 Synthesis and Verification of Digital Systems
Reinventing The Wheel: Developing a New Standard-Cell Synthesis Flow
Technology Mapping I based on tree covering
VLSI CAD Flow: Logic Synthesis, Placement and Routing Lecture 5
Delay Optimization using SOP Balancing
Reinventing The Wheel: Developing a New Standard-Cell Synthesis Flow
Presentation transcript:

Logic synthesis flow Technology independent mapping –Two level or multilevel optimization to optimize a coarse metric related to area/delay Technology dependent mapping –More “concrete” synthesis: map logic function to a given library, typically with characterized areas and delays –Represent circuit by a subject graph –Represent library cells by a set of pattern graphs –(These graphs are always directed acyclic graphs (DAG’s) for combinational circuits) –Find optimal mapping of pattern graphs to the subject graph to optimize a cost function

Example library and its pattern graphs INV (2) NAND2 (4) NAND4 (8) NAND3 (6) AOI21 (6) AOI22 (8) XOR2 (12) Nodes: 2NAND NOT (input permutations not shown) or Cost of using the gate, taken here as the number of transistors

Example circuit Subject graph

Two possible mappings Cost = 2  Cost(NAND2)+2  Cost(INV) +Cost(NAND4)+Cost (NAND3) = 2   = 26 Cost = Cost(NAND2)+Cost(INV) +Cost(AOI21)+2  Cost (NAND3) =  6 = 24 NAND2 INV NAND3 NAND4 NAND3 INV NAND2 AOI21

Tree mapping Applicable to “fanout-free” regions and provably optimal Simple objective: minimize area –Cost is fanout-independent –Things are more tricky when costs are fanout-dependent Basic idea –Traverse from inputs to outputs –At each node Enumerate all mappings of the node and find cost Choose lowest cost solution based on node cost and sum of input costs –Optimal substructure property: any optimal solution can be formed by using the optimal solutions at previous nodes

Tree mapping: an example Cost = 4 Cost = 2+4=6 Cost = 4 Cost = min(10,14) = 10 Cost = 2+10=12 Cost = min(4+12+0, , ) = 12 NAND2 NAND3 NAND4 Mapped as NAND2 Cost = Cost(NAND2)+  Cost(fanins) = = 14 Mapped as NAND3 Cost = Cost(NAND3)+  Cost(fanins) = = 10

Finer points Handling inversions better –Introduce a buffer element (two inverters) with cost 0 into library –Replace inverter-free edges in subject graph by a pair of inverters –Make appropriate (similar) minor changes to pattern graphs –Map as before –Advantage: allows inversions to be considered; any inversions not used are replaced by buffers with cost zero during mapping Mapping DAGs –DAG mapping is NP-complete for area objectives –Intuition: at multi-fanout points, contradictory optimal choices can be made; can only be resolved by logic duplication –Heuristic: decompose DAG into a forest of trees and map each optimally

Circuit decomposition Quality of mapping depends on decomposition of circuit –2NAND/NOT and AND/OR/NOT decompositions used most often –Number of such decompositions is huge! Clever way of considering all decompositions proposed by Lehman, Watanabe, Grodstein and Harkness, IEEE Transactions on CAD 8/1997, pp

Other objective functions Minimizing area considered – seen to be easy Minimizing delay is more complex, since delay is fanout- dependent Basic procedure for tree-mapping –Traverse graph as before –Store several best solutions at each node, parameterized by load capacitance values (i.e., store best solution for several load caps) –This also allows mapper to use various power levels available for each cell in the library

Other objective functions (contd.) Minimize Area subject to Delay  D spec Basic dynamic programming procedure –Traverse tree as before –Find all (Area,Delay) choices at each node –(For fanout-dependent delay, parameterize delay by fanout cap) –Remove provably suboptimal (Area,Delay) subsolutions If Area 1  Area 2 and Delay 1  Delay 2 (clearly one inequality must be strict!), then any optimal solution will prefer (Area 1,Delay 1 ) to (Area 2,Delay 2 ) Can prune the latter from the list Pictorially Delay Area This point is provably worst than this one

Final note on synthesis Deep submicron technologies (< 0.25 micron or so): –Wire capacitances are significant –Interconnect significantly affects circuit performance Traditional wire-load models –Fanout load depends on # fanouts, calculated statistically –Invalid in deep submicron since fanout RC’s depend on locations and not just number of fanouts Physical synthesis –Concurrently perform placement and synthesis –Common approach Create “physical prototype” – coarse placement Perform synthesis according to this placement; refine placement as you go along