Device hang detection and recovery

摘要

The techniques described in the disclosure are generally related to gradual, iterative hang recovery for a graphics processing unit (GPU). The techniques described in the disclosure attempt to re-execute instructions of an application in response to a GPU hang, rather than stopping the execution of the application. If the re-execution causes the GPU to hang again, the techniques described in the disclosure cause the GPU to iteratively execute next set of instructions.

主权项

1. A method for hang recovery, the method comprising:
determining a first set of instructions of an application a graphics processing unit (GPU) was executing at a time of a GPU hang; resetting the GPU; instructing the GPU to re-execute the first set of instructions of the application that the GPU was executing at the time of the GPU hang; determining that the GPU hung during the re-execution of the first set of instructions of the application; and in response to the determination that the GPU hung during the re-execution of the first set of instructions:
resetting the GPU; andinstructing the GPU to execute a second set of instructions of the application.