Abstract: A neural network system includes an input layer, one or more hidden layers, and an output layer. The input layer receives a training set including a sequence of batches and provides to its following layer output activations associated with the sequence of batches respectively. A first hidden layer receives, from its preceding layer, a first input activation associated with a first batch, receive a first input gradient associated with a second batch preceding the first batch, and provide, to its following layer a first output activation associated with the first batch based on the first input activation and first input gradient. The first and second batches have a delay factor associated with at least two batches. The output layer receives, from its preceding layer, a second input activation, and provide, to its preceding layer, a first output gradient based on the second input activation and the first training set.

Abstract: A circuit design is partitioned into a plurality of partitions during a first synthesis by a computer processor. After modification of the circuit design, the computer processor determines changed partitions and unchanged partitions of the circuit design. The computer processor then determines dependent partitions of the changed partitions. The changed partitions and the dependent partitions are re-synthesized by the computer processor into respective re-synthesized partitions, and the computer processor then combines the respective re-synthesized partitions and the unchanged partitions into a complete synthesized circuit design in a memory.

Abstract: Low-density parity check (LDPC) decoder circuitry is configured to decode an input codeword using a plurality of circulant matrices of a parity check matrix for an LDPC code. Multiple memory banks are configured to store elements of the input codeword. A memory circuit is configured for storage of an instruction sequence. Each instruction describes for one of the circulant matrices, a corresponding layer and column of the parity check matrix and a rotation. Each instruction includes packing factor bits having a value indicative of a number of instructions of the instruction sequence to be assembled in a bundle of instructions. A bundler circuit is configured to assemble the number of instructions from the memory circuit in a bundle. The bundler circuit specifies one or more no-operation codes (NOPs) in the bundle in response to the value of the packing factor bits and provides the bundle to the decoder circuitry.

Abstract: A circuit for implementing a polar decoder is described. The circuit includes a log-likelihood ratio processing circuit. A path metric update circuit is coupled to receive log-likelihood values for decoded bits from the log-likelihood ratio processing circuit, wherein the path metric circuit generates path metric values for the decoded bits. A partial sum calculation circuit is coupled to receive the path metrics; and a sort and cull circuit is coupled to receive a list of child path, wherein the sort and cull circuit eliminates invalid paths from the list of child paths. A method of implementing a polar decoder is also described.

Abstract: A receiver implemented in an integrated circuit device is described. The receiver circuit comprises a first receiver circuit configured to receive first data, wherein the first receiver circuit comprises a first memory element configured to receive the first data in response to a first clock signal; a latency mirror circuit configured to receive second data, wherein the latency mirror circuit comprises a second memory element configured to receive the second data in response to a second clock signal; and a latency control circuit configured to detect a latency in the second data, wherein the latency control circuit adjusts a phase of the first clock signal used to receive the first data in the first receiver circuit.

Abstract: Apparatus and associated methods relating to data packet deparsing include an editing circuit configured to perform one or more predetermined editing operations on headers of an incoming data packet step by step without extracting all headers from the incoming data packet. In an illustrative example, an editor circuit may include an updating circuit configured to receive the data packet and update a header in the data packet. The editor circuit may also include a removal circuit configured to remove a header from the data packet. The editor circuit may also include an insertion circuit configured to insert one or more consecutive headers to the data packet. A state machine may be configured to enable or disable the updating circuit, the removal circuit, and/or the insertion circuit based on the predetermined editing operations. By using the editing circuit, packet deparsing may be performed with less hardware resources and low latency.

Abstract: Examples herein describe techniques for testing a receiver interface on a die. In one embodiment, the die includes tester circuitry which includes a digital to analog convertor (DAC) which outputs an analog test signal to a selector circuit (e.g., a multiplexer) which forwards the analog test signal to a receiver. By varying the analog test signal, the tester circuitry can identify one or more trip points corresponding to the receiver. That is, by monitoring the output of the receiver, a testing application can determine when the output of the receiver switches states thereby indicating that the analog test signal at the input of the receiver corresponds to the trip point of the receiver. In this manner, internal circuitry (e.g., the tester circuitry) can be used to test a receiver interface that may otherwise be inaccessible.

Abstract: Examples herein describe a self-test process where an integrated circuit includes a test controller responsible for testing a plurality of frames in the memory of an integrated circuit. The test controller can receive a test pattern which the controller duplicates and stores in each of the plurality of frames. However, frames may be non-uniform meaning the frames have varying sizes. As such, some of the frames may only store parts of the test pattern rather than all of it. In any case, the test controller reads out the stored data and generates a checksum which can then be compared to a baseline checksum generated from simulating the integrated circuit using design code to determine whether there is a manufacturing defect in the frames.

Abstract: Method, apparatus and computer-readable medium for providing a partial reconfiguration of a reconfigurable module are described. In one example, a method reads a netlist for a design of a circuit comprising a reconfigurable module and sets the reconfigurable module to a first region. The method then generates a second region that encompasses the first region and places the design with the first region. The method routes the design with the second region and generates a partial bitstream for the reconfigurable module.

Abstract: A device may include a processor system and an array of data processing engines (DPEs) communicatively coupled to the processor system. Each of the DPEs includes a core and a DPE interconnect. The processor system is configured to transmit configuration data to the array of DPEs, and each of the DPEs is independently configurable based on the configuration data received at the respective DPE via the DPE interconnect of the respective DPE. The array of DPEs enable, without modifying operation of a first kernel of a first subset of the DPEs of the array of DPEs, reconfiguration of a second subset of the DPEs of the array of DPEs.

Abstract: An integrated circuit is described. The integrated circuit comprises an analog-to-digital converter circuit configured to receive an input signal at an input and generate an output signal at an output; and a monitor circuit coupled to the output of the analog-to-digital converter circuit, the monitor circuit configured to receive the output signal and to generate integration coefficients for the analog-to-digital converter circuit; wherein the integration coefficients are dynamically generated based upon signal characteristics of the output signal generated by the analog-to-digital converter circuit. A method of receiving data in an integrated circuit is also described.

Abstract: Examples described herein relate to dynamically structured single instruction, multiple data (SIMD) instructions, and systems and circuits implementing such dynamically structured SIMD instructions. An example is a method for processing data. A first SIMD structure is determined by a processor. A characteristic of the first SIMD structure is altered by the processor to obtain a second SIMD structure. An indication of the second SIMD structure is communicated from the processor to a numerical engine. Data is packed by the numerical engine into an SIMD instruction according to the second SIMD structure. The SIMD instruction is transmitted from the numerical engine.

Abstract: Data is received at a buffer used by a protocol processing stack which protocol processes the received data. The received data is made available to, for example, an application, before the protocol processing of the data is complete. If the protocol processing is successful the data made available to the application is committed.

Abstract: Examples of the present disclosure provide power gating for stacked die structures. In some examples, a stacked die structure comprises a first die and a second die bonded to the first die. In some examples, a power gated power path is from a bonding interface between the dies through TSVs in the second die, a power gating device in the second die, and routing of metallization layers in the second die to the circuit region in the second die. In some examples, a power gated power path comprises a power gating device in a power gating region of the first die and is configured to interrupt a flow of current through the power gated power path to a circuit region in the second die.

Abstract: The present disclosure relates to a processor for implementing artificial neural networks, for example, convolutional neural networks. The processor includes a memory controller group, an on-chip bus and a processor core, wherein the processor core further includes a register map, an instruction module, a data transferring controller, a data writing scheduling unit, a buffer pool, a data reading scheduling unit and a computation module. The processor of the present disclosure may be used for implementing various neural networks with increased computation efficiency.

Abstract: A test system for testing a wafer for integrated circuit devices is described. The test system comprises a first plurality of test probes adapted to make electrical contacts to first corresponding contacts of a wafer tested by the test system; a second plurality of test probes adapted to make electrical contacts to second corresponding contacts on a perimeter region of a portion of the wafer tested by the test system; and a control circuit coupled to the first plurality of test probes and the second plurality of test probes; wherein the control circuit determines whether the second plurality of test probes has a proper contact with the wafer based upon signals received by the second plurality of test probes. A method of testing a wafer for an integrated circuit is also described.

Abstract: An example multi-master system in a system-on-chip (SoC) includes a plurality of master circuits, an error-correcting code (ECC) proxy bridge comprising hardened circuitry in the SoC, a local interconnect configured to couple the plurality of master circuits to the ECC proxy bridge, a memory not having ECC support, and a system interconnect configured to couple the ECC proxy bridge to the memory. The ECC proxy bridge is configured to establish an ECC proxy region in the memory and, for each write transaction from the plurality of master circuits that targets the ECC proxy region, calculate and insert ECC bytes into the respective write transaction.

Abstract: Some examples described herein provide for a multi-chip structure including one or more memory dies stacked on a die having a programmable integrated circuit (IC). In an example, a multi-chip structure includes a package substrate, a first die, and a second die. The first die includes a programmable IC, and the programmable IC includes a memory controller. The first die is on and attached to the package substrate. The second die includes memory. The second die is stacked on the first die. The memory is communicatively coupled to the memory controller.

Abstract: Disclosed herein are integrated circuit devices and and methods for fabricating the same that include at least one non-I/O die having ESD protection circuitry. The ESD protection circuitry disclosed herein may also be utilized in I/O dies. In one example, an integrated circuit device includes a die having a first body. First and second contact pads are exposed to a surface of the first body. The first contact pad is configured to connect to a first supply voltage. The second contact pad is configured to connect to a second supply voltage or ground. A first charge-sensitive circuitry formed in the first body is coupled between the first and second contact pads. A first RC clamp formed in the first body is coupled between the first and second contact pads. The first RC clamp includes at least two BigFETs coupled between the first and second contact pads, and a trigger circuitry coupled in parallel to gate terminals of the at least two BigFETs.

Abstract: A method comprises receiving at a compiler a bit file description and a program, said bit file description comprising a description of routing of a part of a circuit. The method comprises compiling the program using said bit file description to output a bit file for said program.