Hotspot detection based on litho-aware machine learning
Aspects of the disclosed technology relate to techniques of hotspot detection. Pinching-type hotspot candidates and bridging-type hotspot candidates are first identified in the layout design based on predetermined criteria. Simulation is then performed to derive aerial image intensity values for a plurality of sites on each of the pinching-type and bridging-type hotspot candidates. Pinching-type hotspots are determined from the pinching-type hotspot candidates based on one or more machine learning models for pinching-type hotspots, and bridging-type hotspots are determined from the bridging-type hotspot candidates based on one or more machine learning models for bridging-type hotspots. The input vector for the machine learning models is the aerial image intensity values for the plurality of sites.
1. A method, executed by at least one processor of a computer, comprising:
receiving a layout design;
identifying pinching-type hotspot candidates and bridging-type hotspot candidates in the layout design based on predetermined criteria;
performing simulation to derive aerial image intensity values for a plurality of sites on each of the pinching-type hotspot candidates and for a plurality of sites on each of the bridging-type hotspot candidates;
determining pinching-type hotspots from the pinching-type hotspot candidates based on one or more machine learning models for pinching-type hotspots and bridging-type hotspots from the bridging-type hotspot candidates based on one or more machine learning models for bridging-type hotspots, the input for the one or more machine learning models for pinching-type hotspots being the aerial image intensity values for the plurality of sites on each of the pinching-type hotspot candidates, the input for the one or more machine learning models for bridging-type hotspots being the aerial image intensity values for the plurality of sites on each of the bridging-type hotspot candidates;
storing information of the pinching-type hotspots and the bridging-type hotspots; and
thereafter finalizing the layout design to repair the pinching-type hotspots and the bridging-type hotspots, and manufacturing a circuit according to the finalized layout design using a photolithographic process.
2. The method recited in claim 1 , wherein the one or more machine learning models for pinching-type hotspots comprise a machine learning model for horizontal pinching-type hotspots and a machine learning model for vertical pinching-type hotspots, the one or more machine learning models for bridging-type hotspots comprise a machine learning model for horizontal bridging-type hotspots and a machine learning model for vertical bridging-type hotspots, the pinching-type hotspot candidates are separated into a group of horizontal pinching-type hotspot candidates and a group of vertical pinching-type hotspot candidates, and the bridging-type hotspot candidates are separated into a group of horizontal bridging-type hotspot candidates and a group of vertical bridging-type hotspot candidates.
3. The method recited in claim 1 , wherein the predetermined criteria may be set according to design checking rules.
4. The method recited in claim 1 , wherein the one or more machine learning models for pinching-type hotspots and the one or more machine learning models for bridging-type hotspots are C-type Support Vector Machine models.
5. The method recited in claim 1 , wherein a number of the plurality of sites is greater than 8 and less than 20.
6. One or more non-transitory computer-readable media storing computer-executable instructions for causing one or more processors to perform a method, the method comprising:
receiving a layout design;
identifying pinching-type hotspot candidates and bridging-type hotspot candidates in the layout design based on predetermined criteria;
performing simulation to derive aerial image intensity values for a plurality of sites on each of the pinching-type hotspot candidates and for a plurality of sites on each of the bridging-type hotspot candidates;
determining pinching-type hotspots from the pinching-type hotspot candidates based on one or more machine learning models for pinching-type hotspots and bridging-type hotspots from the bridging-type hotspot candidates based on one or more machine learning models for bridging-type hotspots, the input for the one or more machine learning models for pinching-type hotspots being the aerial image intensity values for the plurality of sites on each of the pinching-type hotspot candidates, the input for the one or more machine learning models for bridging-type hotspots being the aerial image intensity values for the plurality of sites on each of the bridging-type hotspot candidates;
storing information of the pinching-type hotspots and the bridging-type hotspots; and
finalizing the layout design to repair the pinching-type hotspots and the bridging-type hotspots, wherein a circuit is thereafter manufactured according to the finalized layout design using a photolithographic process.
7. The one or more non-transitory computer-readable media recited in claim 6 , wherein the one or more machine learning models for pinching-type hotspots comprise a machine learning model for horizontal pinching-type hotspots and a machine learning model for vertical pinching-type hotspots, the one or more machine learning models for bridging-type hotspots comprise a machine learning model for horizontal bridging-type hotspots and a machine learning model for vertical bridging-type hotspots, the pinching-type hotspot candidates are separated into a group of horizontal pinching-type hotspot candidates and a group of vertical pinching-type hotspot candidates, and the bridging-type hotspot candidates are separated into a group of horizontal bridging-type hotspot candidates and a group of vertical bridging-type hotspot candidates.
8. The one or more non-transitory computer-readable media recited in claim 6 , wherein the predetermined criteria may be set according to design checking rules.
9. The one or more non-transitory computer-readable media recited in claim 6 , wherein the one or more machine learning models for pinching-type hotspots and the one or more machine learning models for bridging-type hotspots are C-type Support Vector Machine models.
10. The one or more non-transitory computer-readable media recited in claim 6 , wherein a number of the plurality of sites is greater than 8 and less than 20.
11. A system, comprising:
one or more processors, the one or more processors programmed to perform a method, the method comprising:
receiving a layout design;
identifying pinching-type hotspot candidates and bridging-type hotspot candidates in the layout design based on predetermined criteria;
performing simulation to derive aerial image intensity values for a plurality of sites on each of the pinching-type hotspot candidates and for a plurality of sites on each of the bridging-type hotspot candidates;
determining pinching-type hotspots from the pinching-type hotspot candidates based on one or more machine learning models for pinching-type hotspots and bridging-type hotspots from the bridging-type hotspot candidates based on one or more machine learning models for bridging-type hotspots, the input for the one or more machine learning models for pinching-type hotspots being the aerial image intensity values for the plurality of sites on each of the pinching-type hotspot candidates, the input for the one or more machine learning models for bridging-type hotspots being the aerial image intensity values for the plurality of sites on each of the bridging-type hotspot candidates;
storing information of the pinching-type hotspots and the bridging-type hotspots; and
finalizing the layout design to repair the pinching-type hotspots and the bridging-type hotspots, wherein a circuit is thereafter manufactured according to the finalized layout design using a photolithographic process.
12. The system recited in claim 11 , wherein the one or more machine learning models for pinching-type hotspots comprise a machine learning model for horizontal pinching-type hotspots and a machine learning model for vertical pinching-type hotspots, the one or more machine learning models for bridging-type hotspots comprise a machine learning model for horizontal bridging-type hotspots and a machine learning model for vertical bridging-type hotspots, the pinching-type hotspot candidates are separated into a group of horizontal pinching-type hotspot candidates and a group of vertical pinching-type hotspot candidates, and the bridging-type hotspot candidates are separated into a group of horizontal bridging-type hotspot candidates and a group of vertical bridging-type hotspot candidates.
13. The system recited in claim 11 , wherein the predetermined criteria may be set according to design checking rules.
14. The system recited in claim 11 , wherein the one or more machine learning models for pinching-type hotspots and the one or more machine learning models for bridging-type hotspots are C-type Support Vector Machine models.
15. The system recited in claim 11 , wherein a number of the plurality of sites is greater than 8 and less than 20.