Abstract: A method resumes an accelerated processing device (APD) wavefront in which a subset of elements have faulted. A restore command for a job including a wavefront is received. A list of context states for the wavefront is read from a memory associated with a APD. An empty shell wavefront is created for restoring the list of context states. A portion of not acknowledged data is masked over a portion of acknowledged data within the restored wavefronts.

Abstract: A method and apparatus includes a multi-processor apparatus including a plurality of integrated circuit processors having a shared thermal platform. Each processor has at least one subsystem operable at a plurality of different power settings, at least one internal thermal parameter detector providing power data related to the processor, and a power management unit. The method and apparatus illustratively shares power data from the at least one internal thermal parameter detector of each processor between the power management units of the plurality of processors; compares the shared power data from the plurality of processors to a thermal design power limit for the shared thermal platform; and controls a power setting of the at least one subsystem of the plurality of processors within the shared thermal platform based on the comparison of the shared power data to the thermal design power limit for the shared thermal platform.

Abstract: Provided herein is a method and system for providing and analyzing unified data signaling that includes setting, or analyzing a state of a single indicator signal, generating or analyzing a data pattern of a plurality of data bits, and signal, or determine, based on the state of the single indicator signal and the pattern of the plurality of data bits, that data bus inversion has been applied to the plurality of data bits or that the plurality of data bits is poisoned.

Abstract: Sophisticated gate stacks including a high-k dielectric material and a metal-containing electrode material may be covered by a protection liner, such as a silicon nitride liner, which may be maintained throughout the entire manufacturing sequence at the bottom of the gate stacks. For this purpose, a mask material may be applied prior to removing cap materials and spacer layers that may be used for encapsulating the gate stacks during the selective epitaxial growth of a strain-inducing semiconductor alloy. Consequently, enhanced integrity may be maintained throughout the entire manufacturing sequence, while at the same time one or more lithography processes may be avoided.

Abstract: A method and apparatus determines an activity history context for each of a plurality of virtual machines sharing use of a graphics processing core. Each activity history context provides information related to a power setting of at least one engine of the graphics processing core during at least one prior use of the graphics processing core by the corresponding virtual machine. The method and apparatus controls a power setting of the at least one engine of the graphics processing core based on the activity history context corresponding to an active virtual machine using the graphics processing core.

Abstract: A method and apparatus controls power management of a graphics processing core when multiple virtual machines are allocated to the graphics processing core on a much finer-grain level than conventional systems. In one example, the method and apparatus processes a plurality of virtual machine power control setting requests to determine a power control request for a power management unit of a graphics processing core. The method and apparatus then controls power levels of the graphics processing core with the power management unit based on the determined power control request.

Abstract: A system and method embodiments for optimally allocating compute kernels to different types of processors, such as CPUs and GPUs, in a heterogeneous computer system are disclosed. These include comparing a kernel profile of a compute kernel to respective processor profiles of a plurality of processors in a heterogeneous computer system, selecting at least one processor from the plurality of processors based upon the comparing, and scheduling the compute kernel for execution in the selected at least one processor.

Abstract: A system on a chip including a bus, a bootup module coupled to the bus and configured to cause the system on a chip to bootup in accordance with a selected security mode, an input module coupled to the bus and configured to receive an input signal and to provide the input signal to the bus, a processor coupled to the bus and configured to process the input signal to provide an intermediate signal, in accordance with a type of content protection associated with the input signal, an encryption module coupled to the bus and configured to cause at least a portion of the intermediate signal to be encrypted to produce an encrypted signal, in accordance with the type of the content protection, and an output module coupled to the bus and configured to output the encrypted signal.

Abstract: In a transistor, a strain-inducing semiconductor alloy, such as silicon/germanium, silicon/carbon and the like, may be positioned very close to the channel region by providing gradually shaped cavities which may then be filled with the strain-inducing semiconductor alloy. For this purpose, two or more “disposable” spacer elements of different etch behavior may be used in order to define different lateral offsets at different depths of the corresponding cavities. Consequently, enhanced uniformity and, thus, reduced transistor variability may be accomplished, even for sophisticated semiconductor devices.

Abstract: Integrated circuits with memory built-in self test (MBIST) circuitry and methods are disclosed that employ enhanced features. In one aspect of the invention, MBST circuitry is used set memory elements of arrays to a first state and then to an inverse state during a burn-in operation to maintain each of the two opposing states for a desired time in order to either force a failure of the integrated circuit component or produce a pre-stressed component beyond an infancy stage. Preferably, an integrated circuit is provided having MIBST circuitry configured to serially test multiple arrays of memory elements within a component of the integrated circuit and to also conduct parallel initialization of the serially tested arrays.

Abstract: A semiconductor device is presented here. The semiconductor device includes an integrated inductor formed on a semiconductor substrate, a transistor arrangement formed on the semiconductor substrate to modulate loop current induced by the integrated inductor, dielectric material to insulate the integrated inductor from the transistor arrangement, and a controller coupled to the transistor arrangement. The controller is used to select conductive and nonconductive operating states of the transistor arrangement. A conductive operating state of the transistor arrangement allows formation of induced loop current in the transistor arrangement, and a nonconductive operating state of the transistor arrangement inhibits formation of induced loop current in the transistor arrangement.

Abstract: Systems and methods for multi-precision computation are disclosed. One embodiment of the present invention includes a plurality of multiply-add units (MADDs) configured to perform one or more single precision operations and an arrangement generator to generate one or more mantissa arrangements using a plurality of double precision numbers. Each MADD is configured to receive and load said mantissa arrangements from the arrangement generator. The MADDs compute a result of a multi-precision computation using the mantissa arrangements. In an embodiment, the MADDs are configured to simultaneously perform operations that include, single precision operations, double-precision additions and double-precision multiply and additions.

Abstract: Methods and test receiver apparatus are provided for loopback testing of a unidirectional physical layer device. The disclosed methods and test receiver apparatus allow for the phase of a sampling clock implemented at the test receiver apparatus to be aligned with the phase of a test data signal.

Abstract: Systems and methods for synchronizing thread wavefronts and associated events are disclosed. According to an embodiment, a method for synchronizing one or more thread wavefronts and associated events includes inserting a first event associated with a first data output from a first thread wavefront into an event synchronizer. The event synchronizer is configured to release the first event before releasing events inserted subsequent to the first event. The method further includes releasing the first event from the event synchronizer after the first data is stored in the memory. Corresponding system and computer readable medium embodiments are also disclosed.

Abstract: Methods and systems relating to providing constants are provided. In an embodiment, a method of providing constants in a processing device includes copying a constant of a first constant buffer to a second constant buffer, the first and second constant buffers being included in a ring of constant buffers and a size of the ring being one greater than a number of processes that the processing device can process concurrently, updating a value of the constant in the second buffer, and binding a command to be executed on the processing device to the second constant buffer.

Abstract: A method and apparatus are described for processing data during an execution pipeline cycle of a processor. Valid bits of the data are generated according to a designated data size. Each of the valid bits is inserted into at least one of a plurality of bit positions. The valid bits are rotated in a predetermined direction (i.e., left or right rotation) by a designated number of bit positions. Valid bits are removed from a portion of the plurality of bit positions after being rotated. Zeros or most significant bits (MSBs) of the data may be inserted in the bit positions from which the valid bits were removed. The number of bit positions to rotate the valid bits by may be designated by a first bit subset and a second bit subset. The first bit subset may indicate a number of bytes, and the second bit subset may indicate a number of bits.

Abstract: In an embodiment, a method of processing memory requests in a first processing device is provided. The method includes generating a memory request associated with a memory address located in an unpinned memory space managed by an operating system running on a second processing device; and responsive to a determination that the memory address is not resident in a physical memory, transmitting a message to the second processing device. In response to the message, the operating system controls the second processing device to bring the memory address into the physical memory.

Abstract: The method and accompanying apparatus provides secure register access. In one example, as part of a secure boot process, data is written into a managed secure register (MSR) register and access policy data is written into programmable MSR policy registers. During run-time, the MSR register securely stores data in compliance with the programmable register access policy data. Access policy is enforced during run-time based on the programmable register access policy data.

Abstract: Techniques are disclosed relating to reducing wander created by AC couplers. In one embodiment, an integrated circuit is disclosed that includes an AC coupler and a DC-level shifter. The AC coupler is configured to receive a differential input signal at first and second nodes, and to shift a common-mode voltage of the differential input signal. The DC-level shifter is coupled to the first and second nodes, and configured to reduce wander of the AC coupler. In various embodiments, the DC-level shifter is configured to supply a differential reference signal to the AC coupler, and to create the differential reference signal from the differential input signal at the first and second nodes by changing a common-mode voltage of the differential input signal.

Abstract: A multiplexer (MUX) circuit with balanced select line loading is provided. The MUX circuit includes a plurality of 2:1 MUX units coupled together in a multistage cascading arrangement, along with a selection module coupled to the MUX units. The MUX units are arranged in an initial MUX stage, at least one intermediate MUX stage coupled to and following the initial MUX stage, and a final MUX stage coupled to and following the at least one intermediate MUX stage. Each MUX unit is controlled with a respective select bit input value provided by the selection module. The selection module controls the operation of the MUX units in the initial MUX stage with a first plurality of different select bits, controls the operation of the MUX units in the at least one intermediate MUX stage with a second plurality of different select bits, and controls the operation of the final MUX stage with a devoted select bit.