Abstract: A graphics processing unit (GPU) and a method is disclosed that performs a convolution operation recast as a matrix multiplication operation. The GPU includes a register file, a processor and a state machine. The register file stores data of an input feature map and data of a filter weight kernel. The processor performs a convolution operation on data of the input feature map and data of the filter weight kernel as a matrix multiplication operation. The state machine facilitates performance of the convolution operation by unrolling the data of the input feature map and the data of the filter weight kernel in the register file. The state machine includes control registers that determine movement of data through the register file to perform the matrix multiplication operation on the data in the register file in an unrolled manner.

Abstract: A method for performing an atomic memory operation may include receiving an atomic input, receiving an address for an atomic memory location, and performing an atomic operation on the atomic memory location based on the atomic input, wherein performing the atomic operation may include performing a first operation on a first portion of the atomic input, and performing a second operation, which may be different from the first operation, on a second portion of the atomic input. The method may further include storing a result of the first operation in a first portion of the atomic memory location, and storing a result of the second operation in a second portion of the atomic memory location. The method may further include returning an original content of the first portion of the atomic memory location concatenated with an original content of the second portion of the atomic memory location.

Abstract: A system and a method are disclosed improving forward progress of preempted workloads. A graphics pipeline processes tiles of a first low-priority job. A controller stops the first job by resetting the GPU and preempting the first job with a second job having a higher priority, determine whether the first job has been previously preempted one or more times, and adjust a batch-binning parameter reducing a likelihood that the first job will again be preempted in the current frame. In one embodiment, the controller is configured to stop the first job at a preemption boundary during a draw call or by resetting the GPU. A batch-binning parameter may include postponing sorting primitives into tiles during a binning process, increasing a number of tiles for backend rendering, reducing a quality of anti-aliasing, decreasing a shading rate quality, and/or decreasing input resolution and increasing upscaling of the first job.

Abstract: A system and method is disclosed that allows multiple casting devices to work together to populate a large display screen according to the subject matter disclosed herein. The system includes a receiving device that includes two or more screen-cast receivers and a controller. Each screen-cast receiver receives from a corresponding casting device at least a portion of a frame of original content of the corresponding casting device generated in a native resolution of the corresponding casting device. The controller synchronizes each received portion of the frame of the original content of the corresponding casting device to form a video output signal that comprises a combination of each received portion, in addition to any internally generated content derived by the receiving display. A casting device may be a smartphone, a tablet, or a computing device, such as a laptop computer.

Abstract: A graphics processing unit (GPU) and a method is disclosed that performs a convolution operation recast as a matrix multiplication operation. The GPU includes a register file, a processor and a state machine. The register file stores data of an input feature map and data of a filter weight kernel. The processor performs a convolution operation on data of the input feature map and data of the filter weight kernel as a matrix multiplication operation. The state machine facilitates performance of the convolution operation by unrolling the data of the input feature map and the data of the filter weight kernel in the register file. The state machine includes control registers that determine movement of data through the register file to perform the matrix multiplication operation on the data in the register file in an unrolled manner.

Abstract: A system and method is disclosed that allows multiple casting devices to work together to populate a large display screen according to the subject matter disclosed herein. The system includes a receiving device that includes two or more screen-cast receivers and a controller. Each screen-cast receiver receives from a corresponding casting device at least a portion of a frame of original content of the corresponding casting device generated in a native resolution of the corresponding casting device. The controller synchronizes each received portion of the frame of the original content of the corresponding casting device to form a video output signal that comprises a combination of each received portion, in addition to any internally generated content derived by the receiving display. A casting device may be a smartphone, a tablet, or a computing device, such as a laptop computer.

Abstract: A method for performing an atomic memory operation may include receiving an atomic input, receiving an address for an atomic memory location, and performing an atomic operation on the atomic memory location based on the atomic input, wherein performing the atomic operation may include performing a first operation on a first portion of the atomic input, and performing a second operation, which may be different from the first operation, on a second portion of the atomic input. The method may further include storing a result of the first operation in a first portion of the atomic memory location, and storing a result of the second operation in a second portion of the atomic memory location. The method may further include returning an original content of the first portion of the atomic memory location concatenated with an original content of the second portion of the atomic memory location.

Abstract: A GPU is disclosed, which may include a VRS interface to provide spatial information and/or primitive-specific information. The GPU may include one or more shader cores including a control logic section to determine a shading precision value based on the spatial information and/or the primitive-specific information. The control logic section may modulate a shading precision according to the shading precision value. A method for controlling shading precision by a GPU may include providing, by a VRS interface, the spatial information and/or primitive-specific information. The method may include determining, by a control logic section, a shading precision value based on the spatial information and/or the primitive-specific information. The method may include modulating a shading precision according to the shading precision value.