UT DA GAN-SRAF: Sub-Resolution Assist Feature Generation using - - PowerPoint PPT Presentation

UT DA

Mohamed Baker Alawieh, Yibo Lin, Zaiwei Zhang, Meng Li, Qixing Huang and David Z. Pan The University of Texas at Austin

Work funded in part by NSF

GAN-SRAF: Sub-Resolution Assist Feature Generation using Generative Adversarial Networks

1

Motivation

t With the IC technology scaling, resolution enhancement

techniques are becoming indispensable

t Sub-Resolution Assist Feature (SRAF) generation is used to

improve the lithographic process window of target patterns

2

https://slideplayer.com/slide/9416386/

Conventional Approaches

t Rule-Based approaches:

› Work well for simple designs with regular patterns › Cannot handle complex shapes

t Model-Based (MB) approaches:

› Achieve high quality results › Suffer from exorbitant computational cost

t Machine Learning (ML) Based approach:

› Achieves results quality similar to MB › Results in 10X reduction in runtime

3

Conventional Approaches

t Rule-Based approaches:

› Work well for simple designs with regular patterns › Cannot handle complex shapes

t Model-Based (MB) approaches:

› Achieve high quality results › Suffer from exorbitant computational cost

t Machine Learning (ML) Based approach:

› Achieves results quality similar to MB › Results in 10X reduction in runtime

Can we do better?!

4

ML Based Approach

t Proposes local sampling scheme with a classification model t On a 2D grid, the classifier predicts the presence of SRAF

in each grid

0 1 2 N%1 sub%sampling0point

Target pattern SRAF SRAF box OPC region SRAF region

SRAF%label:%0 SRAF%label:%1

5

[Xu et al, ISPD’16, TCAD’17]

ML Based Approach

6

t While achieving 10X runtime improvement, this approach

has large room for further enhancement [Xu et al, ISPD’16, TCAD’17]

t Proposes local sampling scheme with a classification model t On a 2D grid, the classifier predicts the presence of SRAF

in each grid

ML Based Approach

7

t While achieving 10X runtime improvement, this approach

has large room for further enhancement

› Do we need a 2D grid and local sampling?

[Xu et al, ISPD’16, TCAD’17]

t Proposes local sampling scheme with a classification model t On a 2D grid, the classifier predicts the presence of SRAF

in each grid

ML Based Approach

8

t While achieving 10X runtime improvement, this approach

has large room for further enhancement

› Do we need a 2D grid and local sampling? › Can we avoid the feature extraction step?

[Xu et al, ISPD’16, TCAD’17]

t Proposes local sampling scheme with a classification model t On a 2D grid, the classifier predicts the presence of SRAF

in each grid

ML Based Approach

9

t While achieving 10X runtime improvement, this approach

has large room for further enhancement

› Do we need a 2D grid and local sampling? › Can we avoid the feature extraction step? › Most importantly, with all advancements in Computer Vision, can we recast this problem to leverage these advancement?

[Xu et al, ISPD’16, TCAD’17]

t Proposes local sampling scheme with a classification model t On a 2D grid, the classifier predicts the presence of SRAF

in each grid

CGAN for Image Translation

10

t GANs have been proposed to produce images similar to those in

training data set

t CGAN, takes as an input a picture in one domain and translates it

to a new one

› During training it sees pairs of matched images

[Isola et al, CVPR’18]

SRAF Generation & Image Translation

t What does SRAF generation have to do with Image translation?!

11

t What does SRAF generation have to do with Image translation?! t Can we define the problem as translating images from the Target

Domain (DT) to the SRAF Domain (DS)?

12

SRAF Generation & Image Translation

Challenges

t Layout images have sharp edges which pose a challenge to GANs

› Model is not guaranteed to generate polygon SRAF shapes › Sharp edges can complicate gradient propagation

t Generated images need ultimately be changed to layout format

› Images cannot be directly mapped to ‘GDS’ format › Post-processing step should not be time consuming

13

Challenges

t Layout images have sharp edges which pose a challenge to GANs

› Model is not guaranteed to generate polygon SRAF shapes › Sharp edges can complicate gradient propagation

t Generated images need ultimately be changed to layout format

› Images cannot be directly mapped to ‘GDS’ format › Post-processing step should not be time consuming

t Hence, a proper encoding is needed to address these challenges!

14

Multi-Channel Heatmap Encoding

t Key Idea: encode each type of object on a separate channel in the

image

› Channel index carries object description (type, size,...) › Excitations on the channel carry objects location

15

Original Layout Encoded Layout

Challenges Revisited

t Layout images have sharp edges which pose a challenge to GANs

› Model is not guaranteed to generate polygon SRAF shapes › Polygon shapes are not needed, the objective of model is to predict locations on different channels › Sharp edges can complicate gradient propagation › No sharp edges in encoded image

t Generated images need ultimately be changed to layout format

› Images cannot be directly mapped to ‘GDS’ format › Decoding is straight forward, it suffices to detect excitation location on each channel to get full GDS information

16

t Generator:

› Trained to produce images in DS based

• n input from DT

› Tries to fool the Discriminator

t Discriminator:

› Trained to detect ‘fakes’ generated by the Generator

t The two networks are jointly trained

until convergence

CGAN Approach

17

Encoder Decoder Generator Discriminator Real Input Fake Diff Fake/Real

t Generator:

› Encoder: Downsampling › Decoder: Upsampling

CGAN Approach

18

Encoder Decoder Generator Discriminator Real Input Fake Diff Fake/Real

t Generator:

› Encoder: Downsampling › Decoder: Upsampling

t Discriminator:

› CNN trained as a classifier

t After training, only the generator is

used

CGAN Approach

19

Encoder Decoder Generator Discriminator Real Input Fake Diff Fake/Real

Results Decoding

t Decoding the generated layout images consists of two steps:

› Thresholding & Excitation detection

20

Results Decoding

t Decoding the generated layout images consists of two steps:

› Thresholding & Excitation detection

21

Results Decoding

t Decoding the generated layout images consists of two steps:

› Thresholding & Excitation detection

22

Results Decoding

t Decoding the generated layout images consists of two steps:

› Thresholding & Excitation detection

23 SRAF location Isolated pixel

Results Decoding

t Decoding the generated layout images consists of two steps:

› Thresholding & Excitation detection

24 SRAF location Isolated pixel

Decoding scheme is fast è GPU accelerated

Sample Results

25

• LS_SVM: Xu et al, ISPD’16, TCAD’17
• MB: Model-Based Approach - Calibre

MB

Sample Results

26

• LS_SVM: Xu et al, ISPD’16, TCAD’17
• MB: Model-Based Approach - Calibre

GAN-SRAF MB A post processing legalization step is applied

Sample Results

27

MB SRAF Target Pattern LS_SVM Final CGAN Final LS_SVM Prediction CGAN Prediction

• LS_SVM: Xu et al, ISPD’16, TCAD’17
• MB: Model-Based Approach - Calibre

LS_SVM GAN-SRAF MB

0.0020 0.0025 0.0030 0.0035 0.0040 1000 2000 3000 MB CGAN LS SVM NO SRAF −6 −4 −2 2 2000 4000 6000 MB CGAN LS SVM NO SRAF

Lithography Compliance Checks

28

Histogram of PV (um2) Histogram of EPE (nm)

• LS SVM: Xu et al, ISPD’16, TCAD’17
• MB: Model-Based Approach - Calibre

Comparison Summary

t The proposed CGAN based approach can achieve comparable

results with LS_SVM and MB with 14.6X and 144X reduction in runtime

29

No SRAF MB LS_SVM CGAN PV Band (um2) 0.00335 0.002845 0.00301 0.00291 EPE (nm) 3.9287 0.5270 0.5066 0.541 Run time (s)

• 6910

700 48

Conclusions

t GAN-SRAF, a novel SRAF generation framework, is presented

featuring:

› Novel problem formulation as image translation › Smart heatmap encoding scheme and GPU accelerated decoding

t Results demonstrate significant speedup when compared to ML

and MB

› While achieving comparable lithography performance

30

Thank You!

31