Abstract: Matrix multiplication systolic array feed methods and related processing element (PE) microarchitectures for efficiently implementing systolic array generic matrix multiplier (SGEMM) in integrated circuits is provided. A systolic array architecture may include a processing element array, a column feeder array, and a row feeder array. A bandwidth of external memory may be reduced by a factor of reduction based on interleaving of the matrix data via a feeding pattern of the column feeder array and the row feeder array.

Abstract: An apparatus to facilitate processing of a sparse matrix for arbitrary graph data is disclosed. The apparatus includes a graphics processing unit having a data management unit (DMU) that includes a scheduler for scheduling matrix operations, an active logic for tracking active input operands, and a skip logic for tracking unimportant input operands to be skipped by the scheduler. Processing circuitry is coupled to the DMU. The processing circuitry comprises a plurality of processing elements including logic to read operands and a multiplication unit to multiply two or more operands for the arbitrary graph data and customizable circuitry to provide custom functions.

Abstract: An apparatus to facilitate processing of a sparse matrix for arbitrary graph data is disclosed. The apparatus includes a graphics processing unit having a data management unit (DMU) that includes a scheduler for scheduling matrix operations, an active logic for tracking active input operands, and a skip logic for tracking unimportant input operands to be skipped by the scheduler. Processing circuitry is coupled to the DMU. The processing circuitry comprises a plurality of processing elements including logic to read operands and a multiplication unit to multiply two or more operands for the arbitrary graph data.

Abstract: Matrix multiplication systolic array feed methods and related processing element (PE) microarchitectures for efficiently implementing systolic array generic matrix multiplier (SGEMM) in integrated circuits is provided. A systolic array architecture may include a processing element array, a column feeder array, and a row feeder array. A bandwidth of external memory may be reduced by a factor of reduction based on interleaving of the matrix data via a feeding pattern of the column feeder array and the row feeder array.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises at least one processor to perform operations to implement a neural network and accelerator logic to perform communicatively coupled to the processor to perform compute operations for the neural network.

Abstract: An apparatus and method for compressing floating-point values.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises at least one processor to perform operations to implement a neural network and accelerator logic to perform communicatively coupled to the processor to perform compute operations for the neural network.

Abstract: Matrix multiplication systolic array feed methods and related processing element (PE) microarchitectures for efficiently implementing systolic array generic matrix multiplier (SGEMM) in integrated circuits is provided. A systolic array architecture may include a processing element array, a column feeder array, and a row feeder array. A bandwidth of external memory may be reduced by a factor of reduction based on interleaving of the matrix data via a feeding pattern of the column feeder array and the row feeder array.

Abstract: An embodiment of an integrated circuit may comprise a front end unit, and circuitry coupled to the front end unit, the circuitry to provide a high confidence, multiple branch offset predictor. For example, the circuitry may be configured to identify an entry in a multiple-taken-branch prediction table that corresponds to a conditional branch instruction, determine if a confidence level of the entry exceeds a threshold confidence level, and, if so determined, provide multiple taken branch predictions that stem from the conditional branch instruction from the entry in the multiple-taken-branch prediction table. Other embodiments are disclosed and claimed.

Abstract: Hardware accelerator templates and design frameworks for implementing recurrent neural networks (RNNs) and variants thereof are described. A design framework module obtains a flow graph for an RNN algorithm. The flow graph identifies operations to be performed to implement the RNN algorithm and further identifies data dependencies between ones of the operations. The operations include matrix operations and vector operations. The design framework module maps the operations of the flow graph to an accelerator hardware template, yielding an accelerator instance comprising register transfer language code that describes how one or more matrix processing units and one or more vector processing units are to be arranged to perform the RNN algorithm. At least one of the one or more MPUs, as part of implementing the RNN algorithm, is to directly provide or directly receive a value from one of the one or more VPUs.

Abstract: Hardware accelerator templates and design frameworks for implementing recurrent neural networks (RNNs) and variants thereof are described. A design framework module obtains a flow graph for an RNN algorithm. The flow graph identifies operations to be performed to implement the RNN algorithm and further identifies data dependencies between ones of the operations. The operations include matrix operations and vector operations. The design framework module maps the operations of the flow graph to an accelerator hardware template, yielding an accelerator instance comprising register transfer language code that describes how one or more matrix processing units and one or more vector processing units are to be arranged to perform the RNN algorithm. At least one of the one or more MPUs, as part of implementing the RNN algorithm, is to directly provide or directly receive a value from one of the one or more VPUs.

Abstract: An apparatus and method for offloading iterative, parallel work to a data parallel cluster. For example, one embodiment of a processor comprises: a host processor to execute a primary thread; a data parallel cluster coupled to the host processor over a high speed interconnect, the data parallel cluster comprising a plurality of execution lanes to perform parallel execution of one or more secondary threads related to the primary thread; and a data parallel cluster controller integral to the host processor to offload processing of the one or more secondary threads to the data parallel cluster in response to one of the cores executing a parallel processing call instruction from the primary thread.

Abstract: An apparatus and method for data parallel single program multiple data (SPMD) execution. For example, one embodiment of a processor comprises: instruction fetch circuitry to fetch instructions of one or more primary threads; a decoder to decode the instructions to generate uops; a data parallel cluster (DPC) to execute microthreads comprising a subset of the uops, the DPC further comprising: a plurality of execution lanes to perform parallel execution of the microthreads; an instruction decode queue (IDQ) to store the uops prior to execution; and a scheduler to evaluate the microthreads based on associated variables including instruction pointer (IP) values, the scheduler to gang microthreads into fragments for parallel execution on the execution lanes based on the evaluation.

Abstract: An apparatus and method for performing efficient, adaptable tensor operations.

Abstract: An apparatus and method for compressing floating-point values.

Abstract: An apparatus and method for offloading iterative, parallel work to a data parallel cluster. For example, one embodiment of a processor comprises: a host processor to execute a primary thread; a data parallel cluster coupled to the host processor over a high speed interconnect, the data parallel cluster comprising a plurality of execution lanes to perform parallel execution of one or more secondary threads related to the primary thread; and a data parallel cluster controller integral to the host processor to offload processing of the one or more secondary threads to the data parallel cluster in response to one of the cores executing a parallel processing call instruction from the primary thread.

Abstract: An apparatus and method for data parallel single program multiple data (SPMD) execution. For example, one embodiment of a processor comprises: instruction fetch circuitry to fetch instructions of one or more primary threads; a decoder to decode the instructions to generate uops; a data parallel cluster (DPC) to execute microthreads comprising a subset of the uops, the DPC further comprising: a plurality of execution lanes to perform parallel execution of the microthreads; an instruction decode queue (IDQ) to store the uops prior to execution; and a scheduler to evaluate the microthreads based on associated variables including instruction pointer (IP) values, the scheduler to gang microthreads into fragments for parallel execution on the execution lanes based on the evaluation.

Abstract: An apparatus and method for performing efficient, adaptable tensor operations.

Abstract: Techniques for enabling enhanced parallelism for sparse linear algebra operations having write-to-read dependencies are disclosed. A hardware processor includes a plurality of processing elements, a memory that is heavily-banked into a plurality of banks, and an arbiter. The arbiter is to receive requests from threads executing at the plurality of processing elements seeking to perform operations involving the memory, and to maintain a plurality of lock buffers corresponding to the plurality of banks. Each of the lock buffers is able to track up to a plurality of memory addresses within the corresponding bank that are to be treated as locked in that the values stored at those memory addresses cannot be updated by those of the threads that did not cause the memory addresses to be locked until those memory addresses have been removed from being tracked by the plurality of lock buffers.

Abstract: Techniques involving a compute engine architecture to support data-parallel loops with reduction operations are described. In some embodiments, a hardware processor includes a memory unit and a plurality of processing elements (PEs). Each of the PEs is directly coupled via one or more neighbor-to-neighbor links with one or more neighboring PEs so that each PE can receive a value from a neighboring PE, provide a value to a neighboring PE, or both receive a value from one neighboring PE and also provide a value to another neighboring PE. The hardware processor also includes a control engine coupled with the plurality of PEs that is to cause the plurality of PEs to collectively perform a task to generate one or more output values by each performing one or more iterations of a same subtask of the task.