Abstract:
Candidate layout patterns can be generated using a generative model trained based on known data, such as historical hot spot data, features extraction, and geometrical primitives. The generative model can be sampled to obtain candidate layouts that can be ranked and repaired using error optimization, design rule checking, optical proximity checking, and other methods to ensure that resulting candidates are manufacturable.
Abstract:
Candidate layout patterns can be generated using a generative model trained based on known data, such as historical hot spot data, features extraction, and geometrical primitives. The generative model can be sampled to obtain candidate layouts that can be ranked and repaired using error optimization, design rule checking, optical proximity checking, and other methods to ensure that resulting candidates are manufacturable.
Abstract:
Apparatus, method and computer program product for reducing overlay errors during a semiconductor photolithographic mask design process flow. The method obtains data representing density characteristics of a photo mask layout design; predicts stress induced displacements based on said obtained density characteristics data; and corrects the mask layout design data by specifying shift movement of individual photo mask design shapes to pre-compensate for predicted displacements. To obtain data representing density characteristics, the method merges pieces of data that are combined to make a photo mask to obtain a full reticle field data set. The merge includes a merge of data representing density characteristic driven stress effects. The density characteristics data for the merged reticle data are then computed. To predict stress-induced displacements, the method inputs said density characteristics data into a programmed model that predicts displacements as a function of density, and outputs the predicted shift data.
Abstract:
Various embodiments include computer-implemented methods, computer program products and systems for verifying an integrated circuit (IC) layout. In some cases, approaches include a computer-implemented method of verifying an IC layout, the method including: obtaining data about a process variation band for at least one physical feature in the IC layout; determining voltage-based process variation band thresholds for the at least one physical feature in the IC layout; determining whether the process variation band for the at least one physical feature in the IC layout meets design specifications for the IC layout based upon the voltage-based process variation band thresholds for the at least one physical feature in the IC layout; and modifying the IC layout in response to a determination that the process variation band for the at least one physical feature does not meet the design specifications.