Abstract: The present disclosure relates to methods and apparatus for compute processing. For example, disclosed techniques facilitate improving performance of matrix multiplication in streaming processor. Aspects of the present disclosure can execute, with a load control unit, a first load instruction to load a set of input data of an input matrix from a first memory to a second memory. Aspects of the present disclosure can also execute, with the load control unit, a second load instruction to load a set of weight data of a weight matrix from the first memory to the second memory. Additionally, aspects of the present disclosure can perform, with an ALU component, a matrix multiplication operation using the set of input data and the set of weight data to generate an output matrix. Further, aspects of the present disclosure can store the output matrix at a general purpose register accessible to the ALU component.

Abstract: The curved fork-like grating is provided where curves formed by all points in the same relative positions in respective periods are distributed to be curved and fork-like, so that incident light irradiated on the curved fork-like grating is converted into Bessel-Gaussian light. Such a curved fork-like structure can be achieved by various gratings, such as a metal grating, a dielectric grating, or a metal and dielectric hybrid grating, without affecting their intrinsic characteristics, such as a diffraction efficiency, a use band, a diffraction angle, and a polarization characteristic except that a diffraction light field carries a high-order Bessel phase. Compared with a method for obtaining Bessel-Gaussian light in the prior art, the curved fork-like grating provided by the present invention is simpler in light path and appropriate for a wide band.

Abstract: This disclosure relates to polishing compositions that include (1) at least one abrasive; (2) at least one organic acid or a salt thereof; (3) at least one first amine compound, the at least one first amine compound including an alkylamine having a 6-24 carbon alkyl chain; (4) at least one second amine compound containing at least two nitrogen atoms, the second amine compound being different from the first amine compound; and (5) an aqueous solvent.

Abstract: Embodiments of this application provide a method for annotating video data performed by a computer device. The method includes: displaying target video data in a video application; displaying annotated media data in the video application in response to a trigger operation performed on a progress time point annotation function in the video application, the annotated media data being media data associated with a current progress time point, the current progress time point being a progress time point corresponding to the trigger operation to which the target video data is played; and sharing the annotated media data with another user in response to a confirmation operation performed on the annotated media data, the annotated media data enabling the second user to perform a progress jump on the target video data based on the current progress time point. Through this application, efficiency of video progress positioning can be improved.

Abstract: A method for catalytically synthesizing furaneol, which uses a specific peptide to function as a catalyst, uses rhamnose to function as a raw material, and uses an organic solvent and a phosphate buffer to function as a reaction solvent to be co-heated to prepare furaneol.

Abstract: A video-based interaction method includes playing a video on a video playback interface to which a target object is logged in, acquiring a local viewing marking record of the video and scene identifiers of markable scene clips in the video, comparing the local viewing marking record with the scene identifiers to determine an unmarked target scene clip in the video, displaying, in response to playing to the target scene clip and on the video playback interface, a target scene clip marking entry that indicates performing viewing marking on the target scene clip, and displaying a viewing marking credential for the target clip in response to a viewing marking trigger operation on the target scene clip marking entry. The viewing marking credential covers a part of the video playback interface, is an independent operable object, and includes description information of the video and identity information of the target object.

Abstract: This disclosure relates polishing compositions that include (1) at least one abrasive; (2) at least one organic acid or a salt thereof; (3) at least one first amine compound, the at least one first amine compound including an alkylamine having a 6-24 carbon alkyl chain; (4) at least one second amine compound containing at least two nitrogen atoms, the second amine compound being different from the first amine compound; and (5) an aqueous solvent.

Abstract: The present disclosure relates to methods and apparatus for compute processing. For example, disclosed techniques facilitate improving performance of matrix multiplication in streaming processor. Aspects of the present disclosure can execute, with a load control unit, a first load instruction to load a set of input data of an input matrix from a first memory to a second memory. Aspects of the present disclosure can also execute, with the load control unit, a second load instruction to load a set of weight data of a weight matrix from the first memory to the second memory. Additionally, aspects of the present disclosure can perform, with an ALU component, a matrix multiplication operation using the set of input data and the set of weight data to generate an output matrix. Further, aspects of the present disclosure can store the output matrix at a general purpose register accessible to the ALU component.

Abstract: The present disclosure relates to methods and apparatus for graphics processing, e.g., a GPU. The apparatus may receive an image including a plurality of pixels associated with one or more workgroups and one or more pixel tiles, each of the workgroups and the pixel tiles including one or more pixels of the plurality of pixels. The apparatus may determine whether the one or more workgroups are misaligned with the one or more pixel tiles. The apparatus may determine a conversion order of the one or more workgroups when the one or more workgroups are misaligned with the one or more pixel tiles, the conversion order corresponding to a common multiple of one of the one or more workgroups and one of the one or more pixel tiles. The apparatus may convert each of the one or more workgroups based on the conversion order of the one or more workgroups.

Abstract: The present disclosure relates to methods and apparatus for compute processing. For example, disclosed techniques facilitate improving performance of matrix multiplication in streaming processor. Aspects of the present disclosure can execute, with a load control unit, a first load instruction to load a set of input data of an input matrix from a first memory to a second memory. Aspects of the present disclosure can also execute, with the load control unit, a second load instruction to load a set of weight data of a weight matrix from the first memory to the second memory. Additionally, aspects of the present disclosure can perform, with an ALU component, a matrix multiplication operation using the set of input data and the set of weight data to generate an output matrix. Further, aspects of the present disclosure can store the output matrix at a general purpose register accessible to the ALU component.

Abstract: A polishing composition includes an anionic abrasive, a pH adjuster a low-k removal rate inhibitor, a ruthenium removal rate enhancer, and water. A method of polishing a substrate includes the steps of: applying the polishing composition described herein to a surface of a substrate, wherein the surface comprises ruthenium or a hard mask material; and bringing a pad into contact with the surface of the substrate and moving the pad in relation to the substrate.

Abstract: Provided are dynamic region division and region passage identification methods and a cleaning robot. The dynamic region division method includes: acquiring environment information collected by a robot when working in a first region; determining whether the robot has completed a work task in the first region, when a presence of a passage entering a second region is determined based on the environment information; and complementing a boundary at the passage to block the passage, when the work task is not completed. According to the technical solution provided by the embodiment of the present application, the occurrence probability of repeated sweeping and miss sweeping is reduced, and the cleaning efficiency is high. In addition, the technical solution provided by the embodiment of the present application relies on the environment information collected during the work, rather than relying on historical map data, so that the environmental adaptability is high.

Abstract: This disclosure relates polishing compositions that include (1) at least one abrasive; (2) at least one organic acid or a salt thereof; (3) at least one first amine compound, the at least one first amine compound including an alkylamine having a 6-24 carbon alkyl chain; (4) at least one second amine compound containing at least two nitrogen atoms, the second amine compound being different from the first amine compound; and (5) an aqueous solvent.

Abstract: A method for catalytically synthesizing furaneol, which uses a specific peptide to function as a catalyst, uses rhamnose to function as a raw material, and uses an organic solvent and a phosphate buffer to function as a reaction solvent to be co-heated to prepare furaneol.

Abstract: The present disclosure relates to methods and apparatus for compute processing. For example, disclosed techniques facilitate improving performance of matrix multiplication in streaming processor. Aspects of the present disclosure can execute, with a load control unit, a first load instruction to load a set of input data of an input matrix from a first memory to a second memory. Aspects of the present disclosure can also execute, with the load control unit, a second load instruction to load a set of weight data of a weight matrix from the first memory to the second memory. Additionally, aspects of the present disclosure can perform, with an ALU component, a matrix multiplication operation using the set of input data and the set of weight data to generate an output matrix. Further, aspects of the present disclosure can store the output matrix at a general purpose register accessible to the ALU component.

Abstract: Provided are dynamic region division and region passage identification methods and a cleaning robot. The dynamic region division method includes: acquiring environment information collected by a robot when working in a first region; determining whether the robot has completed a work task in the first region, when a presence of a passage entering a second region is determined based on the environment information; and complementing a boundary at the passage to block the passage, when the work task is not completed. According to the technical solution provided by the embodiment of the present application, the occurrence probability of repeated sweeping and miss sweeping is reduced, and the cleaning efficiency is high. In addition, the technical solution provided by the embodiment of the present application relies on the environment information collected during the work, rather than relying on historical map data, so that the environmental adaptability is high.

Abstract: The present invention discloses a wafer-level semiconductor device and a manufacturing method thereof. The wafer-level semiconductor device comprises a wafer-level substrate; a plurality of serial groups formed on a surface of the substrate and are disposed in parallel, each serial group comprising a plurality of parallel groups disposed in series, each parallel groups comprising a plurality of unit cells disposed in parallel, wherein each unit cell is an independent functional unit which is formed by processing a semiconductor layer directly grown on a surface of the substrate; and a lead, which is at least electrically connected between two selected parallel groups in each serial group to make ON-voltages of all the serial groups substantially consistent. The device of the present invention, with a simple structure, a simple and convenient manufacturing process, and a high efficiency to produce qualified products, can be put into large-scale production and application.

Abstract: In one implementation, the number of half open session initiation protocol (SIP) sessions per-destination (e.g., SIP device) or globally is limited by SIP application layer gateway (ALG) as a SIP DoS/DDoS countermeasure. Compared with traditional SIP DoS/DDoS countermeasures, the proposed solution is simple to implement and, thus, less likely to degrade SIP ALG performance. Moreover, this solution automatically adapts to DoS/DDoS attack arrival rate, while at the same time not degrading legal SIP traffic even if throttling is enforced for the SIP device.

Abstract: The present invention discloses a wafer-level semiconductor device and a manufacturing method thereof. The wafer-level semiconductor device comprises a wafer-level substrate; a plurality of serial groups formed on a surface of the substrate and are disposed in parallel, each serial group comprising a plurality of parallel groups disposed in series, each parallel groups comprising a plurality of unit cells disposed in parallel, wherein each unit cell is an independent functional unit which is formed by processing a semiconductor layer directly grown on a surface of the substrate; and a lead, which is at least electrically connected between two selected parallel groups in each serial group to make ON-voltages of all the serial groups substantially consistent. The device of the present invention, with a simple structure, a simple and convenient manufacturing process, and a high efficiency to produce qualified products, can be put into large-scale production and application.

Abstract: In one implementation, the number of half open session initiation protocol (SIP) sessions per-destination (e.g., SIP device) or globally is limited by SIP application layer gateway (ALG) as a SIP DoS/DDoS countermeasure. Compared with traditional SIP DoS/DDoS countermeasures, the proposed solution is simple to implement and, thus, less likely to degrade SIP ALG performance. Moreover, this solution automatically adapts to DoS/DDoS attack arrival rate, while at the same time not degrading legal SIP traffic even if throttling is enforced for the SIP device.