Abstract: A neural network training tool selects from a plurality of parallelizing techniques and selects from a plurality of forward-propagation computation techniques. The neural network training tool performs a forward-propagation phase to train a neural network using the selected parallelizing technique and the selected forward-propagation computation technique based on one or more inputs. Additionally, the neural network training tool selects from a plurality computation techniques and from a plurality of parallelizing techniques for a backward-propagation phase. The neural network training tool performs a backward-propagation phase of training the neural network using the selected backward-propagation parallelizing technique and the selected backward-propagation computation technique to generate error gradients and weight deltas and to update weights associated with one or more layers of the neural network.

Abstract: In some implementations, the operations of a software system may include the execution of untrusted device drivers. The execution of an untrusted device driver may be initiated when an untrusted device driver that is a set of executable instructions is obtained. A computing resource for the execution of the device driver may be further determined from a device-driver manifest of the untrusted device driver. Computing resources may include one or more of a hardware resource, a memory, an input/output port, an interrupt request line, and an inter-process communication channel. Trusted local access objects may be used by the untrusted device driver to access the computing resources.

Abstract: Embodiments of the present invention comprise methods and systems for augmenting machine virtualization by entirely bypassing resource emulation for performance-critical features, such as 3D graphics acceleration, through the use of high-performance interfaces between the guest OS and the host OS. Embodiments of the present invention also comprise methods and systems for ameliorating the performance penalties and functionality restrictions of conventional resource emulation.

Abstract: Embodiments of the present invention comprise methods and systems for augmenting machine virtualization by entirely bypassing resource emulation for performance-critical features, such as 3D graphics acceleration, through the use of high-performance interfaces between the guest OS and the host OS. Embodiments of the present invention also comprise methods and systems for ameliorating the performance penalties and functionality restrictions of conventional resource emulation.

Abstract: Described herein are one or more implementations for describing and/or addressing the configuration requirements of applications, extensions, device drivers, and other components of a software system.