Coordinates-based generative adversarial networks for generating synthetic physical design layout patterns

Patent metadata

Field	Value
Publication number	US-10606975-B2
Application number	US-201815994598-A
Country	US
Kind code	B2
Filing date	May 31, 2018
Priority date	May 31, 2018
Publication date	Mar 31, 2020
Grant date	Mar 31, 2020

Abstract

A method for generating physical design layout patterns includes the step of selecting one or more physical design layouts, a given physical design layout comprising a set of physical design layout patterns for features in at least one layer of a given patterned structure. The method also includes the step of converting the physical design layout patterns into coordinate arrays, a given coordinate array comprising feature center coordinates for the features in a given one of the physical design layout patterns. The method further includes the step of training, utilizing the coordinate arrays, a generative adversarial network (GAN) comprising discriminator and generator neural networks. The method further includes the step of generating one or more synthetic coordinate arrays utilizing the trained generator neural network of the GAN, a given one of the synthetic coordinate arrays comprising feature center coordinates of features for a new physical design layout pattern.

First claim

What is claimed is:

1. A method for generating physical design layout patterns, comprising steps of:
selecting one or more physical design layouts, a given one of the physical design layouts for a given patterned structure comprising a set of physical design layout patterns for features in at least one layer of the given patterned structure;
converting the physical design layout patterns into coordinate arrays, a given coordinate array comprising feature center coordinates for the features in a given one of the physical design layout patterns;
training, utilizing the coordinate arrays, a generative adversarial network (GAN) comprising a discriminator neural network and a generator neural network; and
generating one or more synthetic coordinate arrays utilizing the trained generator neural network of the GAN, a given one of the synthetic coordinate arrays comprising feature center coordinates of features for a new physical design layout pattern;
wherein converting the physical design layout patterns into coordinate arrays comprises generating the given coordinate array by populating designated entries of the given coordinate array with center coordinates for one or more features in a given field of view of the given physical design layout represented by the given physical design layout; and
wherein the method is performed by at least one processing device comprising a processor coupled to a memory.

2. The method of claim 1 , wherein the features comprise vias of the given patterned structure.

3. The method of claim 1 , wherein the features comprise lines for the given patterned structure.

4. The method of claim 1 , wherein the given coordinate array has a size of M*N, where M is a number of values for representing the center coordinates of respective ones of the features in a designated coordinate system and Nis a maximum number of features in the given field of view of the given physical design layout represented by the given physical design layout pattern.

5. The method of claim 4 , wherein when the given field of view comprises fewer than N features, the given coordinate array is padded with null values outside a range of possible coordinates for the given field of view of the given physical design layout pattern.

6. The method of claim 1 , wherein the features comprise fixed-size rectangular features.

7. The method of claim 6 , wherein the given coordinate array comprises designated entries for coordinates of respective ones of the centers of the fixed-size rectangular features.

8. The method of claim 7 , wherein the coordinates of the centers of the fixed-size rectangular features comprise coordinates in a Cartesian coordinate system, wherein a first designated set of entries of the given coordinate array provide x-coordinates of the centers of the fixed-size rectangular features, and wherein a second designated set of entries of the given coordinate array provide y-coordinates of the centers of the fixed-size rectangular features.

9. The method of claim 1 , wherein the given physical design layout comprises features of at least two different sizes.

10. The method of claim 9 , wherein the coordinates of the centers of the features comprise coordinates in a Cartesian coordinate system, wherein a first designated set of entries of the given coordinate array provide x-coordinates of the centers of features of a first size, wherein a second designated set of entries of the given coordinate array provide y-coordinates of the centers of the features of the first size, wherein a third designated set of entries of the given coordinate array provide x-coordinates of the centers of features of a second size different than the first size, and wherein a fourth designated set of entries of the given coordinate array provide y-coordinates of the centers of the features of the second size.

11. The method of claim 1 , wherein at least one of the discriminator and generator neural networks comprise a series of fully connected neural network layers.

12. The method of claim 1 , further comprising applying post-processing to the generated synthetic coordinate arrays using one or more design rule checks.

13. The method of claim 1 , further comprising converting the generated synthetic coordinate arrays to a format used in an electronic design automation (EDA) software.

14. The method of claim 1 , further comprising utilizing the given synthetic coordinate array to evaluate manufacturability of the new physical design layout pattern.

15. A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by at least one computing device to cause the at least one computing device to perform steps of:
selecting one or more physical design layouts, a given one of the physical design layouts for a given patterned structure comprising a set of physical design layout patterns for features in at least one layer of the given patterned structure;
converting the physical design layout patterns into coordinate arrays, a given coordinate array comprising feature center coordinates for the features in a given one of the physical design layout patterns;
training, utilizing the coordinate arrays, a generative adversarial network (GAN) comprising a discriminator neural network and a generator neural network; and
generating one or more synthetic coordinate arrays utilizing the trained generator neural network of the GAN, a given one of the synthetic coordinate arrays comprising feature center coordinates of features for a new physical design layout pattern;
wherein converting the physical design layout patterns into coordinate arrays comprises generating the given coordinate array by populating designated entries of the given coordinate array with center coordinates for one or more features in a given field of view of the given physical design layout represented by the given physical design layout.

16. The computer program product of claim 15 , wherein the given coordinate array has a size of M*N, where M is a number of values for representing the center coordinates of respective ones of the features in a designated coordinate system and N is a maximum number of features in the given field of view of the given physical design layout represented by the given physical design layout pattern.

17. The computer program product of claim 15 , wherein the program instructions further cause the at least one computing device to perform the step of utilizing the given synthetic coordinate array to evaluate manufacturability of the new physical design layout pattern.

18. An apparatus comprising:
a memory; and
at least one processor coupled to the memory and configured for:
selecting one or more physical design layouts, a given one of the physical design layouts for a given patterned structure comprising a set of physical design layout patterns for features in at least one layer of the given patterned structure;
converting the physical design layout patterns into coordinate arrays, a given coordinate array comprising feature center coordinates for the features in a given one of the physical design layout patterns;
training, utilizing the coordinate arrays, a generative adversarial network (GAN) comprising a discriminator neural network and a generator neural network; and
generating one or more synthetic coordinate arrays utilizing the trained generator neural network of the GAN, a given one of the synthetic coordinate arrays comprising feature center coordinates of features for a new

Assignees

• IBM

Inventors

• Sha Jing
• Guillorn Michael A
• Dunn Derren N

Classifications

• G06F30/394

  Routing (G06F30/396 takes precedence) · CPC title

• G06F30/27

  using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model · CPC title

• G06F30/392 Primary

  Floor-planning or layout, e.g. partitioning or placement · CPC title

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title

• G06F17/5077

  Physics · mapped topic