Abstract: Methods, systems, and devices for processing are described. A device may parse a set of layers of a deep neural network. The set of layers may be associated with a set of machine learning operations of the deep neural network. The device may determine one or more layer parameters based on the determined set of layers. In some aspects, the device may determine an execution time associated with executing a shader dispatch based on the one or more layer parameters. The device may batch the shader dispatch to a command buffer based on the execution time and process the command buffer based on the batching. The device may determine a target execution time based on an assembly time associated with the command buffer, a processing time associated with the command buffer, a frequency level associated with processing the command buffer, the one or more layer parameters, or some combination thereof.

Abstract: The present disclosure relates to methods and apparatus for machine learning processing. For example, disclosed techniques facilitate tile-based GPU machine learning acceleration. Aspects of the present disclosure can determine a tile size based on a memory size of a first memory and a job input size associated with executing a computational job. In some examples, the computational job may be one of a quantity of computational jobs configured to execute a machine learning primitive. Aspects of the present disclosure can also load, based on the tile size, input data associated with a batch of computational jobs from a second memory to the first memory. Further, aspects of the present disclosure can generate batch output data by executing the batch of computational jobs using the input data loaded to the first memory. Additionally, aspects of the present disclosure can store the generated batch output data to the second memory.

Abstract: A method for camera processing using a camera application programming interface (API) is described. A processor executing the camera API may be configured to receive instructions that specify a use case for a camera pipeline, the use case defining at least one or more processing engines of a plurality of processing engines for processing image data with the camera pipeline, wherein the plurality of processing engines includes one or more of fixed-function image signal processing nodes internal to a camera processor and one or more processing engines external to the camera processor. The processor may be further configured to route image data to the one or more processing engines specified by the instructions, and return the results of processing the image data with the one or more processing engines to the application.

Abstract: Methods, systems, and devices for processing are described. A device may parse a set of layers of a deep neural network. The set of layers may be associated with a set of machine learning operations of the deep neural network. The device may determine one or more layer parameters based on the determined set of layers. In some aspects, the device may determine an execution time associated with executing a shader dispatch based on the one or more layer parameters. The device may batch the shader dispatch to a command buffer based on the execution time and process the command buffer based on the batching. The device may determine a target execution time based on an assembly time associated with the command buffer, a processing time associated with the command buffer, a frequency level associated with processing the command buffer, the one or more layer parameters, or some combination thereof.

Abstract: A method for camera processing using a camera application programming interface (API) is described. A processor executing the camera API may be configured to receive instructions that specify a use case for a camera pipeline, the use case defining at least one or more processing engines of a plurality of processing engines for processing image data with the camera pipeline, wherein the plurality of processing engines includes one or more of fixed-function image signal processing nodes internal to a camera processor and one or more processing engines external to the camera processor. The processor may be further configured to route image data to the one or more processing engines specified by the instructions, and return the results of processing the image data with the one or more processing engines to the application.

Abstract: A method for camera processing using a camera application programming interface (API) is described. A processor executing the camera API may be configured to receive instructions that specify a use case for a camera pipeline, the use case defining at least one or more processing engines of a plurality of processing engines for processing image data with the camera pipeline, wherein the plurality of processing engines includes one or more of fixed-function image signal processing nodes internal to a camera processor and one or more processing engines external to the camera processor. The processor may be further configured to route image data to the one or more processing engines specified by the instructions, and return the results of processing the image data with the one or more processing engines to the application.

Abstract: A method for camera processing using a camera application programming interface (API) is described. A processor executing the camera API may be configured to receive instructions that specify a use case for a camera pipeline, the use case defining at least one or more processing engines of a plurality of processing engines for processing image data with the camera pipeline, wherein the plurality of processing engines includes one or more of fixed-function image signal processing nodes internal to a camera processor and one or more processing engines external to the camera processor. The processor may be further configured to route image data to the one or more processing engines specified by the instructions, and return the results of processing the image data with the one or more processing engines to the application.

Abstract: In an example, a method for rendering graphics data includes receiving a plurality of commands associated with a plurality of render targets, where the plurality of commands are received in an initial order. The method also includes determining an execution order for the plurality of commands including reordering one or more of the plurality of commands in a different order than the initial order based on data dependencies between commands. The method also includes executing the plurality of commands in the determined execution order.

Abstract: A method and apparatus for ray tracing may include a method, manufacture and apparatus for ray tracing that may include dividing a render target into a plurality of bins. Next, a visibility pass is performed using ray tracing to generate a visibility stream such that the visibility stream indicates, for each bin of the plurality of bins, which primitives are visible in the bin. Then, for at least one bin of the plurality of bins, each primitive in the bin that is indicated in the visibility stream as being visible in the bin is rendered.

Abstract: Systems, methods, and apparatus for performing queries in a graphics processing system are disclosed. These systems, methods, and apparatus may be configured to read a running counter at the start of the query to determine a start value, wherein the running counter counts discrete graphical entities, read the running counter at the end of the query to determine an end value, and subtract the start value from the end value to determine a result.

Abstract: A method and apparatus for ray tracing may include a method, manufacture and apparatus for ray tracing that may include dividing a render target into a plurality of bins. Next, a visibility pass is performed using ray tracing to generate a visibility stream such that the visibility stream indicates, for each bin of the plurality of bins, which primitives are visible in the bin. Then, for at least one bin of the plurality of bins, each primitive in the bin that is indicated in the visibility stream as being visible in the bin is rendered.

Abstract: This disclosure describes techniques for supporting intra-frame timestamps in a graphics system that performs tile-based rendering. The techniques for supporting intra-frame timestamps may involve generating a timestamp value that is indicative of a point in time based on a plurality of per-bin timestamp values that are generated by a graphics processing unit (GPU) while performing tile-based rendering for a graphics frame. The timestamp value may be a function of at least two of the plurality of per-bin timestamp values. The timestamp value may be generated by a central processing unit (CPU), the GPU, another processor, or any combination thereof. By using per-bin timestamp values to generate timestamp values for intra-frame timestamp requests, intra-frame timestamps may be supported by a graphics system that performs tile-based rendering.

Abstract: In an example, a method for rendering graphics data includes receiving a plurality of commands associated with a plurality of render targets, where the plurality of commands are received in an initial order. The method also includes determining an execution order for the plurality of commands including reordering one or more of the plurality of commands in a different order than the initial order based on data dependencies between commands. The method also includes executing the plurality of commands in the determined execution order.

Abstract: Systems, methods, and apparatus for performing queries in a graphics processing system are disclosed. These systems, methods, and apparatus may be configured to read a running counter at the start of the query to determine a start value, wherein the running counter counts discrete graphical entities, read the running counter at the end of the query to determine an end value, and subtract the start value from the end value to determine a result.

Abstract: This disclosure describes techniques for supporting intra-frame timestamps in a graphics system that performs tile-based rendering. The techniques for supporting intra-frame timestamps may involve generating a timestamp value that is indicative of a point in time based on a plurality of per-bin timestamp values that are generated by a graphics processing unit (GPU) while performing tile-based rendering for a graphics frame. The timestamp value may be a function of at least two of the plurality of per-bin timestamp values. The timestamp value may be generated by a central processing unit (CPU), the GPU, another processor, or any combination thereof. By using per-bin timestamp values to generate timestamp values for intra-frame timestamp requests, intra-frame timestamps may be supported by a graphics system that performs tile-based rendering.

Abstract: Methods, devices, and computer program products for virtual user interface controls in multi-display configurations are described herein. In one aspect, an electronic device includes a processor configured to generate a first image of the screen, the first image of the screen not containing a touch-sensitive user interface, generate a second image, the second image comprising a touch-sensitive user-interface configured to be overlayed onto the first image of the screen, transmit one or more of the first image of the screen and the second image to the first display device, and output the first image of the screen to a second display device.