Abstract: Methods and systems are disclosed for clock delay compensation in a multiple chiplet system. Techniques disclosed include distributing, by a clock generator, a clock signal across distribution trees of respective chiplets; measuring phases, by phase detectors, where each phase measurement is associated with a chiplet of the chiplets and is indicative of a propagation speed of the clock signal through the distribution tree of the chiplet. Then, for each chiplet, techniques are further disclosed that determine, by a microcontroller, based on the phase measurements associated with the chiplet, a delay offset, and that delay, based on the delay offset, the propagation of the clock signal through the distribution tree of the chiplet using a delay unit associated with the chiplet.

Abstract: A dynamically configurable overprovisioned microprocessor optimally supports a variety of different compute application workloads and with the capability to tradeoff among compute performance, energy consumption, and clock frequency on a per-compute application basis, using general-purpose microprocessor designs. In some embodiments, the overprovisioned microprocessor comprises a physical compute resource and a dynamic configuration logic configured to: detect an activation-warranting operating condition; undarken the physical compute resource responsive to detecting the activation-warranting operating condition; detect a configuration-warranting operating condition; and configure the overprovisioned microprocessor to use the undarkened physical compute resource responsive to detecting the configuration-warranting operating condition.

Abstract: Systems, apparatuses, and methods for implementing a centralized interrupt controller to aggregate interrupts generated across multiple semiconductor dies are disclosed. A system includes multiple interrupt sources on multiple semiconductor dies. A centralized interrupt controller on one of the semiconductor dies receives and aggregates interrupts from the multiple interrupt sources on the multiple semiconductor dies. This facilitates a single transmission point for forwarding the interrupts to the processor and operating system responsible for handling interrupts. Each interrupt source embeds an ID when conveying an interrupt to the interrupt controller. This allows the interrupt controller to differentiate between the interrupt sources and to identify which source generated a given interrupt. The interrupt controller conveys an indication of the source of the interrupt to the processor when forwarding the interrupt to the processor.

Abstract: Systems, apparatuses, and methods for converting pixel data to a custom swizzle mode are disclosed. A graphics engine receives data in a pre-defined swizzle mode. The graphics engine determines a custom swizzle mode for the data that has directionality aligned to the data itself to further optimize deltas that are used for compressing the data. The graphics engine groups incoming data into group of two neighboring pixels in both the horizontal and vertical directions. The graphics engine scores horizontal and vertical groupings against each other to make a first swizzle mode bit selection. Then the graphics engine increases the grouping of pixels to include additional pixels and scores the increased groupings against each other to make subsequent swizzle mode bit selections. The data is reswizzled into the custom swizzle mode and provided to a compressor to be compressed.

Abstract: A method and apparatus for normalizing an image in an image capturing device includes receiving a processed image by the image device. The processed image is brightness normalized to create a brightness normalized image. The brightness normalized image is provided to an artificial intelligence engine for processing.

Abstract: Systems, apparatuses, and methods for implementing an auto generation and tuning tool for convolution kernels are disclosed. A processor executes multiple tuning runs of a given layer of a neural network while using a different set of operating parameter values for each tuning run. The operating parameters can include one or more of input dataset fetch group size, output channel group size, and other parameters. The processor captures performance data for each tuning run and then after all tuning runs have finished, the processor determines which set of operating parameter values resulted in a better performance for the given neural network layer. The processor uses these operating parameter values for subsequent iterations of the given layer. The processor also performs the same techniques for other layers to determine which set of operating parameter values to use for each layer so as to maximize performance of the neural network.

Abstract: The disclosed method may include detecting, by a control circuit coupled to a first read only memory (ROM) device and a second ROM device, a failure of a first output signal from the first ROM device to a common output. The first ROM device is connected to the common output and the second ROM device is disconnected from the common output. The method also includes switching, by the control circuit in response to detecting the failure, the common output from the first ROM device to the second ROM device. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Systems, apparatuses, and methods for efficient parallel execution of multiple work units in a processor by reducing a number of memory accesses are disclosed. A computing system includes a processor core with a parallel data architecture. One or more of a software application and firmware implement matrix operations and support the broadcast of shared data to multiple compute units of the processor core. The application creates thread groups by matching compute kernels of the application with data items, and grouping the resulting work units into thread groups. The application assigns the thread groups to compute units based on detecting shared data among the compute units. Rather than send multiple read access to a memory subsystem for the shared data, a single access request is generated. The single access request includes information to identify the multiple compute units for receiving the shared data when broadcasted.

Abstract: A device for disabling faulty cores using proxy virtual machines includes a processor, a faulty core, and a physical memory. The processor is responsible for executing a hypervisor that is configured to assign a proxy virtual machine to the faulty core. The assigned proxy virtual machine also includes a minimal workload. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A processing unit such as a graphics processing unit (GPU) includes a set of queues that stores command buffers prior to execution in a corresponding plurality of pipelines. The processing unit also implements a kernel mode driver that allocates a first subset of the set of queues to a first application in response to receiving registration requests from the first application. The processing unit further includes a scheduler that schedules command buffers in the first subset of the set of queues for concurrent execution on a first subset of the set of pipelines. In some cases, an interrupt is generated in response to execution of a first command in a first command buffer in the first queue or the second queue. The interrupt includes an address indicating a location of a routine to be executed by a second subset of the plurality of pipelines.

Abstract: Profile switching for memory overclocking is described. In accordance with the described techniques, a memory is operated according to a first memory profile. During operation of the memory according to the first memory profile, a request is received to operate the memory according to a second memory profile. Responsive to the request, operation of the memory is switched to operate according to the second memory profile without rebooting. In one or more implementations, at least one of the first memory profile or the second memory profile comprises an overclocking memory profile that configures the memory to operate in an overclocking mode. In one or more implementations, the memory is trained to operate according to the overclocking memory profile prior to operating the memory according to the first memory profile.

Abstract: An approach allows concurrent execution of near-memory processing commands, referred to herein as “PIM commands,” and host memory commands. A memory controller determines and issues a plurality of register-only PIM commands that do not reference memory with host memory commands to allow concurrent execution of the register-only PIM commands and the host memory commands. The approach allows concurrent execution of register-only PIM commands and host memory commands without interference, even when the register-only PIM commands and the host memory commands are interleaved, and even for the same memory module, which improves resource utilization and performance. Further improvement of resource utilization and performance is achieved by extending a register-only phase by reordering register-only PIM commands before non-register-only PIM commands, subject to dependency constraints, and using shadow row buffers to provide local working copies of data from memory to near-memory compute elements.

Abstract: Stateful microbranch instructions, including: generating, based on an instruction, a first one or more microinstructions including a stateful microbranch instruction, wherein the stateful microbranch instruction includes: an address of a next instruction after the instruction; a branch target address; one or more microcode attributes; and executing the first one or more microinstructions.

Abstract: A method and apparatus for processing color data includes storing fragment pointer and color data together in a color buffer. A delta color compression (DCC) key indicating the color data to fetch for processing is stored, and the fragment pointer and color data is fetched based upon the read DCC key for decompression.

Abstract: A technique for performing ray tracing operations is provided. The technique includes determining a set of keys and a set of values corresponding to dimensions of a bounding box for a scene; sorting the set of keys and the set of values to generate a sorted set of values; and based on the sorted set of values, generating a Morton code for a triangle of the scene.

Abstract: A compilation technique is provided. The technique includes including a first instruction into a first executable for a first auxiliary processor, wherein the first instruction specifies execution by the first auxiliary processor; and including a second instruction into the first executable, wherein the second instruction targets resources that have affinity with the first auxiliary processor.

Abstract: Stability testing for memory overclocking is described. In accordance with the described techniques, operation of a memory with overclocked memory settings is testing during a boot up process of a computing device. Test results based on the testing are exposed via a user interface. The test results predict a stability of the memory over a subsequent time period if the memory is configured to operate with the overclocked memory settings.

Abstract: Devices and methods are provided for generating an accelerated data structure for ray tracing which include generating a first splitting plane at a first location of a space comprising objects represented by geometry, constructing a first level of an accelerated data structure based on portions of the geometry, straddling the first splitting plane, which are classified as located on opposing sides of the first splitting plane, after constructing the first level of the accelerated data structure, generating a second splitting plane at a second location, different from the first location, of the space and constructing a second level of the accelerated data structure based on portions of the geometry, straddling the second splitting plane, which are classified as located on opposing sides of the second splitting plane.

Abstract: The disclosed integrated circuit for offset cross field effect transistors can include a first transistor include a first channel oriented in a first direction; an oxide layer adjacent to the first transistor; and a second transistor adjacent to the oxide layer. The second transistor can include a second channel that is oriented in a direction orthogonal to the first direction, and the first channel and the second channel can be laterally offset such that the second channel does not cross over the first channel. Various other apparatuses, systems, and methods are also disclosed.

Abstract: Sparse matrix operations using processing-in-memory is described. In accordance with the described techniques, a processing-in-memory component of a memory module receives a request for a vector element stored at a first location in memory of the memory module. The processing-in-memory component identifies an index value for a non-zero element in a sparse matrix using a representation of the sparse matrix stored at a second location in the memory. The processing-in-memory component then outputs a result that includes the vector element by retrieving the vector element from the first location in memory using the index value.