EE244: Design Technology for Integrated Circuits and Systems Outline Lecture 9.2

11/14/97


Click here to start


Table of Contents

EE244: Design Technology for Integrated Circuits and Systems Outline Lecture 9.2

Role of VHDL

Languages versus Models: A Software Analogy

An Alternate View of Synthesis

An Alternate View of Synthesis

An Alternate View of Synthesis

An Alternate View of Synthesis

An Alternate View of Synthesis

Two Approaches to Combinational and Sequential Optimization

An Alternate View of Synthesis: Implications

Abstract Datatype Approach to Consistency: A Software Analogy

Simulated Annealing with Learning

High-Level Goals

PPT Slide

Rationality

Data Engineering

Our Proposal: Stochastic Optimizer with Learning

Overall Flow of Experiment

Selecting a Candidate Move

Parameters in the Model

Cost Estimate vs. TimberWolf Chip Area

Examples Used for Placement

Results for Trained Placement

Three Approaches

Performance of Trained Model: Apex6

Percentage Improvement for Placement

Variation of Parameter Model Weights with Temperature Zone

Netlist Partitioning

Examples for Partitioning

Examples for Partitioning

Netlist Partitioning

Dutt and Deng’s PROP

FM Algorithm

Changes in Cost During FM

Changes in Cost During FM

Note on Experimental Evaluation

Stochastic FM

Stochastic FM: Block Diagram

Experimental Results

Comparison: FM vs HP-FM vs S-FM

Comparison: Best-of-N FM vs long S-FM

Multiple Passes of FM

Visualizing Typical Passes

Cutoff Best-of-N FM

Cutoff Best-of-N FM

Best-of-N FM vs Cutoff Best-of-20

Cutoff Stochastic FM or DT-FM

DT-FM vs Best-of-N FM and PROP

Summary

Stochastic Optimization with Learning for Complex Problems

Summary

Author: Richard Newton

Email: rnewton@ic.eecs.berkeley.edu

Home Page: http://www-inst.eecs.berkeley.edu/~ee244