Abstract: An integrated chip package assembly test system and method for testing a chip package assembly are described herein. In one example, an integrated circuit chip package test system includes a socket and a workpress. The socket is configured to receive a chip package assembly for testing in the test system. The workpress includes a plurality of pusher pins. The plurality of pusher pins have tips extending from a bottom surface of the workpress. Each of the plurality of pusher pins is configured to apply an independent and discrete force to the chip package assembly disposed in the socket.

Abstract: An integrated circuit interconnects are described herein that are suitable for forming integrated circuit chip packages. In one example, an integrated circuit interconnect is provided that includes a first substrate containing first circuitry, a first contact pad, a first pillar, a first pillar protection layer, a second substrate containing second circuitry, and a solder ball disposed on the first pillar and electrically and mechanically coupling the first substrate to the second substrate. The first contact pad is disposed on the first substrate and coupled to the first circuitry. The first pillar electrically disposed over the first contact pad. The first pillar protection layer is hydrophobic to solder and is disposed on a side surface of the first pillar.

Abstract: An electrically insulative pusher pin is disclosed. In one example, an electrically insulative pusher pin includes a first plunger member, a second plunger member, and a spring. The first plunger member has a first end and an exposed second end. The second plunger member has a first end and an exposed second end. The second plunger member is movable relative to the first plunger member, where the exposed second ends of the first and second plunger members defining a length of the pusher pin. The spring disposed between the first ends of the first and second plunger members and biases the exposed second end of the first plunger member away from the exposed second end of the second plunger member. An electrically insulative path is defined between the exposed second end of the first plunger member and the exposed second end of the second plunger member through the pusher pin.

Abstract: An interposer block, a chip package assembly test system and method for testing a chip package assembly are described herein. In one example, an interposer block for an integrated circuit chip package test system is provided. The interposer block includes a main body, a retainer plate, and a cover plate. A plurality of spring pins are each disposed in a respective one of a plurality of spring pin receiving holes formed in the main body. The retainer plate is coupled to the main body and captures the spring pins within the plurality of spring pin receiving holes. The cover plate is movably coupled to the main body. The cover plate has a plurality of spring pin clearance holes form therethrough that align with the plurality of spring pin receiving holes formed in the main body.

Abstract: An integrated chip package assembly test system and method for testing a chip package assembly are described herein. In one example, an integrated circuit chip package test system includes a socket and a workpress. The socket is configured to receive a chip package assembly for testing in the test system. The workpress is positioned over the socket and has a bottom surface that is dynamically conformable to a multi-planar top surface topography of the chip package assembly.

Abstract: Methods and systems are disclosed for booting an integrated circuit (IC). In an example implementation, boot read only memory (ROM) code is loaded for execution by a processor circuit of the IC. Via execution of the boot ROM code on the processor circuit, a first boot image is retrieved. A memory address is communicated from a host device to the processor circuit of the IC via an external data bus coupled to a bus interface circuit in the IC. The bus interface circuit is configured by execution of the first boot image to map a first block of addresses on the internal data bus to a second block of addresses on the host device starting at the memory address. When bus mastering is enabled, the processor retrieves a second boot image from the host device by issuing read requests to the first block of addresses.

Abstract: A quadrature clock correction (QCC) circuit includes: a first pair of clock correction circuits that output in-phase and anti-in-phase clock signals, respectively, of a four-phase clock signal; a second pair of clock correction circuits that output quadrature-phase and anti-quadrature-phase clock signals, respectively, of the four-phase clock signal; a detector circuit configured to detect duty cycle error and in-phase/quadrature-phase (IQ) phase mismatch in the four-phase clock signal; and a calibration circuit configured to supply a first pair of control signals to each the first pair of clock correction circuits, and a second pair of control signals to each of the second pair of clock correction circuits, to correct both the duty cycle error and the IQ phase mismatch based output of the detector circuit.

Abstract: Embodiments described herein include techniques for providing information regarding a hardware part using a scannable code so that a customer can make an informed decision when placing the hardware part in a larger computing system. A customer may purchase hardware parts that are categorized into a certain bin which has guaranteed range of power consumption or performance. The customer may over design the computing system to accommodate the worst parameter in the range (e.g., the minimum performance or the maximum power consumption) to ensure the timing or power specifications are not violated. Instead, the embodiments herein provide a scannable code on the hardware part which the customer can use to access a database which stores more granular information about the part. The customer can use the performance parameters to make better informed decisions to determine where to place the part in the computing system.

Abstract: Methods and apparatus relate to a bidirectional differential interface having a voltage-mode transmit driver architecture formed of multiple selectively enabled slices for coarse output resistance impedance matching. In an illustrative example, the transmit driver may include a programmable resistance for fine-tuning to impedance match the output resistance for transmit operation. During receive operation, protective voltage may be proactively applied to gates of drive transistors, for example, to minimize voltage stresses applied by external signal sources. Some implementations may automatically float the sources of the drive transistors, for example, to prevent back-feeding externally driven signal currents during receive mode operations. The transmit driver may have programmable voltage swing on, for example, the upper and/or lower bounds to enhance compatibility. A programmable common mode voltage node may be selectively applied, for example, through common mode resistors for receive mode operations.

Abstract: In disclosed approaches of neural network processing, a host computer system copies an input data matrix from host memory to a shared memory for performing neural network operations of a first layer of a neural network by a neural network accelerator. The host instructs the neural network accelerator to perform neural network operations of each layer of the neural network beginning with the input data matrix. The neural network accelerator performs neural network operations of each layer in response to the instruction from the host. The host waits until the neural network accelerator signals completion of performing neural network operations of layer i before instructing the neural network accelerator to commence performing neural network operations of layer i+1, for i?1. The host instructs the neural network accelerator to use a results data matrix in the shared memory from layer i as an input data matrix for layer i+1 for i?1.

Abstract: Embodiments herein describe techniques for static scheduling a neural network implemented in a massively parallel hardware system. The neural network may be scheduled using three different scheduling levels referred to herein as an upper level, an intermediate level, and a lower level. In one embodiment, the upper level includes a hardware or software model of the layers in the neural network that establishes a sequential order of functions that operate concurrently in the hardware system. In the intermediate level, identical processes in the functions defined in the upper level are connected to form a systolic array or mesh and balanced data flow channels are used to minimize latency. In the lower level, a compiler can assign the operations performed by the processing elements in the systolic array to different portions of the hardware system to provide a static schedule for the neural network.

Abstract: Embodiments herein describe techniques for interfacing a neural network application with a neural network accelerator using a library. The neural network application may execute on a host computing system while the neural network accelerator executes on a massively parallel hardware system, e.g., a FPGA. The library operates a pipeline for submitting the tasks received from the neural network application to the neural network accelerator. In one embodiment, the pipeline includes a pre-processing stage, an FPGA execution stage, and a post-processing stage which each correspond to different threads. When receiving a task from the neural network application, the library generates a packet that includes the information required for the different stages in the pipeline to perform the tasks. Because the stages correspond to different threads, the library can process multiple packets in parallel which can increase the utilization of the neural network accelerator on the hardware system.

Abstract: At least one neural network accelerator performs operations of a first subset of layers of a neural network on an input data set, generates an intermediate data set, and stores the intermediate data set in a shared memory queue in a shared memory. A first processor element of a host computer system provides input data to the neural network accelerator and signals the neural network accelerator to perform the operations of the first subset of layers of the neural network on the input data set. A second processor element of the host computer system reads the intermediate data set from the shared memory queue, performs operations of a second subset of layers of the neural network on the intermediate data set, and generates an output data set while the neural network accelerator is performing the operations of the first subset of layers of the neural network on another input data set.

Abstract: In the disclosed methods and systems for processing in a neural network system, a host computer system writes a plurality of weight matrices associated with a plurality of layers of a neural network to a memory shared with a neural network accelerator. The host computer system further assembles a plurality of per-layer instructions into an instruction package. Each per-layer instruction specifies processing of a respective layer of the plurality of layers of the neural network, and respective offsets of weight matrices in a shared memory. The host computer system writes input data and the instruction package to the shared memory. The neural network accelerator reads the instruction package from the shared memory and processes the plurality of per-layer instructions of the instruction package.

Abstract: A disclosed neural network processing system includes a host computer system, a RAMs coupled to the host computer system, and neural network accelerators coupled to the RAMs, respectively. The host computer system is configured with software that when executed causes the host computer system to write input data and work requests to the RAMS. Each work request specifies a subset of neural network operations to perform and memory locations in a RAM of the input data and parameters. A graph of dependencies among neural network operations is built and additional dependencies added. The operations are partitioned into coarse grain tasks and fine grain subtasks for optimal scheduling for parallel execution. The subtasks are scheduled to accelerator kernels of matching capabilities. Each neural network accelerator is configured to read a work request from the respective RAM and perform the subset of neural network operations on the input data using the parameters.

Abstract: An example digital-to-time converter (DTC) includes: a delay chain circuit having a plurality of delay cells coupled in sequence, the delay chain circuit including a first input to receive a first clock signal and a second input to receive a second clock signal; and a DEM controller coupled to the delay chain circuit to provide a plurality of control signals to the plurality of delay cells, respectively.

Abstract: An example preprocessor circuit for formatting image data into a plurality of streams of image samples includes: a first buffer configured to store a plurality of rows of the image data and output a row of the plurality of rows; a second buffer, coupled to the first buffer, including a plurality of storage locations to store a respective plurality of image samples of the row output by the first buffer; a plurality of shift registers; an interconnect network including a plurality of connections, each connection coupling a respective one of the plurality of shift registers to more than one of the plurality of storage locations, one or more of the plurality of storage locations being coupled to more than one of the plurality of connections; and a control circuit configured to load the plurality of shift registers with the plurality of image samples based on the plurality of connections and shift the plurality of shift registers to output the plurality of streams of image samples.

Abstract: A circuit for testing bond connections between a first die and a second die is described. The circuit comprises a defect monitoring circuit implemented on the first die, which is configured as a test die; and a plurality of bond connections between the first die and the second die; wherein the defect monitoring circuit is configured to detect a defect in a bond connection of the plurality of bond connections between the first die and the second die. A method of testing bond connections between a first die and a second die is also described.

Abstract: A circuit for storing data in an integrated circuit is described. The circuit comprises an inverter comprising a first transistor having a first gate configured to receive input data and a first output configured to generate a first inverted data output and a second transistor having a second gate configured to receive the input data and a second output configured to generate a second inverted data output; a first pass gate coupled to the first output of the inverter; a second pass gate coupled to the second output of the inverter; and a storage element having an input coupled to receive an output of the first pass gate and an output of the second pass gate. A method of storing data in an integrated circuit is also described.

Abstract: Chip packages and electronic devices are provided that include a thermal capacitance element that improves the operation of IC dies at elevated temperatures. In one example, a chip package is provided that includes an integrated circuit (IC) die, a lid thermally connected to the IC die, and a thermal capacitance element thermally connected to the lid. The thermal capacitance element includes a container and a capacitance material sealingly disposed in the container. The capacitance material has a phase transition temperature that is between 80 and 100 percent of a maximum designed operating temperature in degrees Celsius of the IC die.