Top Banner
Logic Synthesis Improving XOR-Dominated Circuits by Exploiting Dependencies between Operands Zhongkai Chen Electrical and Computer Engineering
15

Background Motivation Implementation Conclusion 2.

Jan 04, 2016

Download

Documents

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Background Motivation Implementation Conclusion 2.

Logic SynthesisImproving XOR-Dominated Circuits by

Exploiting Dependencies between Operands

Zhongkai Chen

Electrical and Computer Engineering

Page 2: Background Motivation Implementation Conclusion 2.

Outline

Background

Motivation

Implementation

Conclusion

2

Page 3: Background Motivation Implementation Conclusion 2.

Unlike FPGA, XOR Gate in ASIC Design has more delay and more area than other gates.

We should reduce the number of XOR.

Background

3

# of XOR ↑ # of XOR ↓

Page 4: Background Motivation Implementation Conclusion 2.

Computer Arithmetic

4

Background

Which is the last area where people still manually optimize circuits, even though logic synthesis has made enormous progress.

AdderIs domination of a good arithmetic, however it contains a LARGE number of XORs.

Page 5: Background Motivation Implementation Conclusion 2.

5

But…Synthesizer seldom rewrites XORs…

Automatically rewriting XORs is necessary for synthesizer! It is time to find a solution!

Background

Page 6: Background Motivation Implementation Conclusion 2.

MotivationWhy not rewrite all XORs?

6

1. Increase size of input expression exponentially. Only being used when number of XORs is small.

Page 7: Background Motivation Implementation Conclusion 2.

MotivationWhy not rewrite all XORs?

7

2. Not all XORs are suitable for expansion:

Before: D: 0.37ns A: 138.2um2

After: D:0.26ns A: 146.9um2

30% Faster! A Small Area Cost.

Before: D: 0.22ns A: 58.8um2

After: D:0.27ns A: 221.2um2

Both Delay and Area Increases!

Page 8: Background Motivation Implementation Conclusion 2.

Implementation

8

Selective ExpansionLocal Correlation

Global Correlation

Page 9: Background Motivation Implementation Conclusion 2.

Implementation

9

Local Correlation

Two Extreme Cases:AB=0 => A B=A+B⊕A+B=1 => A B=AB⊕So we consider AB and (A+B) as two important factors. We evaluate the delay and area of these two factors.

Page 10: Background Motivation Implementation Conclusion 2.

Implementation

10

Local Correlation

Page 11: Background Motivation Implementation Conclusion 2.

Implementation

11

Global Correlation

Page 12: Background Motivation Implementation Conclusion 2.

Implementation

12

Global Correlation

Much FasterLess Area

Optimized Result

Page 13: Background Motivation Implementation Conclusion 2.

Conclusion

13

2 Correlations MergedOptimisation results for all benchmarks

Page 14: Background Motivation Implementation Conclusion 2.

Conclusion

14

2 Correlations MergedComparison of bitwise delays of the multiplier

Less Delay

Page 15: Background Motivation Implementation Conclusion 2.

Thank You

15