| CPC G06F 30/392 (2020.01) [G06F 30/394 (2020.01); G06N 20/00 (2019.01)] | 14 Claims |
|
1. A method comprising:
identifying, by a computing system, net lines corresponding to connections between pins of a source layout design describing a first electronic device and pins of a target layout design describing a second electronic device;
determining, by the computing system implementing a first stage of a machine learning algorithm, a scan order for the net lines based, at least in part, on an orientation of the net lines between pins of the source layout design and the pins of the target layout design, wherein the first stage of the machine learning algorithm is configured with scan order training data generated by receiving an orientation of net lines between multiple training layout designs, iteratively selecting different scan orders for the net lines of the multiple training layout designs, and performing an unravel process on the net lines between the multiple training layout designs, and wherein the scan order training data is generated from the orientation of the net lines, the selected scan orders, and the performance of the unravel process on the net lines;
scanning, by the computing system, the net lines in the scan order for the net lines;
identifying, by the computing system implementing a second stage of the machine learning algorithm, a plurality of the scanned net lines cross each other; and
unraveling, by the computing system implementing the second stage of the machine learning algorithm, the crossed net lines by swapping pin assignments of the crossed net lines.
|