Abstract: Embodiments of the present invention provide a method for video transcoding, an apparatus, an electronic device, a computer readable storage medium, and a computer program product.

Abstract: The present application relates to a data access processing method and apparatus for a GPU, and a storage medium. The method comprises: outputting a first access request, the first access request indicating a data address of a cross-partition SRS accessed this time; performing a hit test for the first access request; and when the result of the hit test is a first test result, reading data corresponding to the data address from a cache line and returning the same, wherein the first test result indicates that the data address accessed this time is hit and a data address accessed historically is hit. That is, a hardware circuit for reading a cache inside a GPU is provided, such that when a data address accessed this time is hit and a data address accessed historically is hit, data can be directly read from a cache line and returned.

Abstract: The method includes: performing a key point detection on a face image to be processed to obtain first specific key point data; determining a first face feature vector corresponding to the face image to be processed; generating initial digital face data based on the first face feature vector, and updating the initial digital face data with the first specific key point data to obtain three-dimensional digital face data; and processing the three-dimensional digital face data by digital human generation software to obtain a target three-dimensional digital human.

Abstract: A method for processing a plurality of tasks includes: obtaining task description information for a task, the task description information comprising a dependency information identifier and a dependent information identifier, the dependency information identifier indicating the number of a first type of tasks associated with the task, and the dependent information identifier at least indicating whether a second type of task associated with the task exists in the plurality of tasks; in response to the task execution ending, updating the dependency information identifier of the second type of task associated with the task, so as to determine from the second type of task associated with the task any tasks to be executed which no longer directly depend on any task; adding the tasks to be executed into a tasks to be executed queue; and executing the tasks to be executed located in the tasks to be executed queue.

Abstract: The present application provides a data caching method and a chip. The chip includes a plurality of video data sub-processors for caching different data types of video data respectively and a data prefetcher. The method includes: determining target video data according to identification information of a video data sub-processor of to-be-cached video data; determining a data storage address of the target video data in an external memory; and obtaining, by the data prefetcher, corresponding target video data from the external memory according to the data storage address. The present application may improve the data caching efficiency, reduce the data access, improve the computational efficiency of video data processing, and reduce the power consumption.

Abstract: The present application provides a data caching method and a chip. The chip includes a plurality of video data sub-processors for caching different data types of video data respectively and a data prefetcher. The method includes: determining target video data according to identification information of a video data sub-processor of to-be-cached video data; determining a data storage address of the target video data in an external memory; and obtaining, by the data prefetcher, corresponding target video data from the external memory according to the data storage address. The present application may improve the data caching efficiency, reduce the data access, improve the computational efficiency of video data processing, and reduce the power consumption.

Abstract: Systems and methods to dynamically adjust operating conditions of a graphics processing unit (GPU) are disclosed. A machine learning model is trained to determine operating voltages and frequencies to be provided to a GPU core of the GPU to execute a workload comprising a plurality of commands. The trained machine learning model is deployed to firmware of the GPU. A command in the workload to be executed by the GPU core is received. The trained machine learning model determines operating voltage and frequency for the GPU core to execute the command.