Abstract: A method of performing automatic tuning on a deep learning model includes: utilizing an instruction-based learned cost model to estimate a first type of operational performance metrics based on a tuned configuration of layer fusion and tensor tiling; utilizing statistical data gathered during a compilation process of the deep learning model to determine a second type of operational performance metrics based on the tuned configuration of layer fusion and tensor tiling; performing an auto-tuning process to obtain a plurality of optimal configurations based on the first type of operational performance metrics and the second type of operational performance metrics; and configure the deep learning model according to one of the plurality of optimal configurations.

Abstract: The present disclosure describes a method for forming metallization layers that include a ruthenium metal liner and a cobalt metal fill. The method includes depositing a first dielectric on a substrate having a gate structure and source/drain (S/D) structures, forming an opening in the first dielectric to expose the S/D structures, and depositing a ruthenium metal on bottom and sidewall surfaces of the opening. The method further includes depositing a cobalt metal on the ruthenium metal to fill the opening, reflowing the cobalt metal, and planarizing the cobalt and ruthenium metals to form S/D conductive structures with a top surface coplanar with a top surface of the first dielectric.

Abstract: The present disclosure describes a method for forming metallization layers that include a ruthenium metal liner and a cobalt metal fill. The method includes depositing a first dielectric on a substrate having a gate structure and source/drain (S/D) structures, forming an opening in the first dielectric to expose the S/D structures, and depositing a ruthenium metal on bottom and sidewall surfaces of the opening. The method further includes depositing a cobalt metal on the ruthenium metal to fill the opening, reflowing the cobalt metal, and planarizing the cobalt and ruthenium metals to form S/D conductive structures with a top surface coplanar with a top surface of the first dielectric.

Abstract: A winding method for pole changeable stator is provided by determining a plurality of classification conditions according to plural types of pole, a plural phases, and current characteristic of the winding coupled to each slot of the stator before and after the pole change, coupling the winding of each slot meeting each classification condition thereby obtaining a plurality of winding groups respectively corresponding to the plurality of classification conditions, and, finally, coupling the plurality of winding groups by a plurality of switching elements so as to form a pole changeable stator.