Xilinx, Inc.

National Semiconductor Corporation

Atmel Corporation

Concurrent Logic, Inc.

Saratoga Semiconductor Corporation

Rafael C Camarota

Frederick Furtek

Tim Garverick

Shao-Pin Chen

Martin L Voogel

Scott K Pickett

David Warren Hawley

Sanjay Popli

Walford W Ho

Edward H Browder

Shankar Moni

Trevor J Bauer

David P Schultz

Sagheer Ahmad

Benjamin S Ting

Ian A Swarbrick

Ephrem C Wu

Weiguang Lu

Hongbin Ji

Henri Fraisse

Benjamin S Devlin

Tien-Yu Lee

Geoffrey A Richmond

David L Ferguson

Martin Newman

Kin Yip Sit

Ui Sun Han

Jim Sutherland

H William Wang

Arthur Smith, Jr

Venkata Alturi

Shin-Mann Day

Frederick C Furtek

Examples generally relate a programmable device having a unified programmable computational memory (PCM) and configuration network. In an example, a programmable device includes a die that includes a PCM integrated circuit having a PCM tile. The PCM tile includes a configuration memory (CM) and combinational logic (CL). The CM is capable of storing configuration data received via a node in the PCM tile. The CL is configured to receive internal control signal(s) and first and second input signals and to output a result signal. The CL is capable of outputting the result signal resulting from a logic function that is responsive to the internal control signal(s) and a signal of a group of signals including the first and second input signals. The CL is configured to receive the first input signal via the node in the PCM tile.

Unified programmable computational memory and configuration network

Certain aspects of the present disclosure provide techniques for relate to electronic devices that are configured to implement multi-rank high bandwidth memory (HBM) memory. In one aspect, an electronic device includes a chip that includes an interface circuit. The interface circuit is connected to first exterior pads. The first exterior pads have a first number of first data input/output exterior pads and a second number of clock enable output exterior pads. The first number is a first integer multiple of a number of data signals per channel of high bandwidth memory (HBM), and the second number is a second integer multiple of a number of clock enable signals per channel of the HBM. The second integer multiple is greater than the first integer multiple.

Multi-rank high bandwidth memory (HBM) memory

Examples described herein provide for an integrated circuit (IC) having a programmable logic region that is capable of being configured via a programmable network. In an example, an IC includes a programmable logic region, a controller, and a programmable network. The programmable network is connected between the controller and the programmable logic region. The controller is programmed to configure the programmable logic region via the programmable network. In some examples, the programmable logic region can be configured faster, among other benefits.

Configuring programmable logic region via programmable network

Programmable devices and methods of operation are disclosed. In some embodiments, a programmable device may include programmable logic selectively coupled to a plurality of input/output (I/O) interface circuits by a programmable interconnect fabric and a network-on-chip (NoC) interconnect system. The programmable logic may include configurable logic elements, programmable interconnects, and dedicated circuitry. The programmable interconnects may form part of the programmable interconnect fabric. In some embodiments, the programmable interconnect fabric selectively routes non-packetized data between the programmable logic and a first group of I/O interface circuits, and the NoC interconnect system selectively routes packetized data between the programmable logic and a second group of I/O interface circuits. The NoC interconnect system may operate according to a data packet protocol, and the second group of I/O interface circuits may include memory controllers compatible with the data packet protocol.

Integrating rows of input/output blocks with memory controllers in a columnar programmable fabric archeture

Examples described herein generally relate to a layered boundary interconnect in an integrated circuit (IC) and methods for operating such IC. In an example, an IC includes a programmable logic region, a plurality of input/output circuits, a plurality of hard block circuits, and a programmable native transmission network. The programmable native transmission network is connected to and between the plurality of input/output circuits and the plurality of hard block circuits. The plurality of hard block circuits is connected to and between the programmable native transmission network and the programmable logic region.

Layered boundary interconnect

Methods and apparatus are described for adding one or more features (e.g., high bandwidth memory (HBM)) to an existing qualified stacked silicon interconnect (SSI) technology programmable IC die (e.g., a super logic region (SLR)) without changing the programmable IC die (e.g., adding or removing blocks). One example integrated circuit (IC) package generally includes a package substrate; at least one interposer disposed above the package substrate and comprising a plurality of interconnection lines; a programmable IC die disposed above the interposer; a fixed feature die disposed above the interposer; and an interface die disposed above the interposer and configured to couple the programmable IC die to the fixed feature die using a first set of interconnection lines routed through the interposer between the programmable IC die and the interface die and a second set of interconnection lines routed through the interposer between the interface die and the fixed feature die.

Standalone interface for stacked silicon interconnect (SSI) technology integration

Examples described herein provide for a boundary logic interface (BLI) to a programmable logic region in an integrated circuit (IC), and methods for operating such IC. An example IC includes a programmable logic region and boundary logic interfaces. The programmable logic region includes columns of interconnect elements disposed between columns of logic elements. The boundary logic interfaces are at respective ends of and communicatively connected to the columns of interconnect elements. The boundary logic interfaces are outside of a boundary of the programmable logic region. A first boundary logic interface (BLI) of the boundary logic interfaces is configured to be communicatively connected to an exterior circuit. The first BLI includes an interface configured to communicate a signal between the exterior circuit and the programmable logic region.

Boundary logic interface

Embodiments herein describe techniques for assigning address ranges to ports in switches forming a packet protocol switch network in an integrated circuit. Instead of relying on a designer to provide the addresses, the integrated circuit can include an address bus which is incremented as addresses are assigned to the ports. In one embodiment, the port addresses are assigned from a root device and defines the address range of each branch port and the address of each endpoint in the network. As the address bus reaches an endpoint, an adder in the endpoint increments the value of the address bus (e.g., the current address). The address bus may use serial or parallel data communication to assign the addresses. In another embodiment, instead of using a separate address bus, a data bus typically used for packet communication assigns the addresses to the ports in the network.

Auto address generation for switch network

A network on a chip (NoC) testing interface (NTI) includes a plurality of switches whose ports are coupled to respective endpoints. In one embodiment, the ports and endpoints are coupled to a shared bus that starts and terminates at a root device. The endpoints are assigned unique address which the NTI uses to select one of the endpoints so that test data is forwarded to a device under test (DUT) coupled to the endpoint. In one embodiment, the endpoints include selection logic for determining whether the endpoint has been selected, and if so, forwarding test data to the DUT. For example, if the endpoint receives a data vector on the bus which has an address that matches the unique address of the endpoint, the selection logic forwards the test data contained in subsequently received data vectors to the DUT until a different address is received.

Test network for a network on a chip and a configuration network

Methods and apparatus are described for strategically arranging conductive elements (e.g., solder balls) of an integrated circuit (IC) package (and the corresponding conductive pads of a circuit board for electrical connection with the IC package) using a plurality of different pitches. One example integrated circuit (IC) package generally includes an integrated circuit die and an arrangement of electrically conductive elements coupled to the integrated circuit die. In at least one region of the arrangement, the conductive elements are disposed with a first pitch in a first dimension of the arrangement and with a second pitch in a second dimension of the arrangement, and the second pitch is different from the first pitch. The pitch of a given region may be based on mechanical, PCB routing, and/or signal integrity considerations.

Heterogeneous ball pattern package

A circuit for implementing a scan chain in programmable resources of an integrated circuit is described. The circuit comprises a programmable element configured to receive an input signal and generate an output signal based upon the input signal; a selection circuit configured to receive the output signal generated by the programmable element at a first input and to receive a scan chain input signal at a second input, wherein the selection circuit generates a selected output signal in response to a selection circuit control signal; and a register configured to receive the selected output signal of the selection circuit.

Circuit for and method of implementing a scan chain in programmable resources of an integrated circuit

Methods and apparatus are described for adding one or more features (e.g., HBM) to a qualified SSI technology programmable IC region by providing an interface (e.g., an HBM buffer region with a switch network) between the added feature device and the programmable IC region. One example IC package generally includes a package substrate; at least one interposer disposed above the package substrate; a programmable IC region disposed above the interposer; at least one fixed feature die disposed above the interposer; and an interface region disposed above the interposer and configured to couple the programmable IC region to the fixed feature die via a first set of interconnection lines routed through the interposer between a first plurality of ports of the interface region and the fixed feature die and a second set of interconnection lines routed between a second plurality of ports of the interface region and the programmable IC region.

High bandwidth memory (HBM) bandwidth aggregation switch

An integrated circuit (IC) die and integrated circuit (IC) chip packages having such dies are described that leverage the symmetry of the arrangement of micro-bumps to advantageously reduce interposer cost and size requirements. In one example, an integrated circuit (IC) die is provided. The IC die includes a die body, a plurality of programmable tiles disposed in the die body, and a plurality of micro-bumps disposed in the die body. The die body includes a front face connecting a bottom exterior surface and a top exterior surface. A centerline of the die body is perpendicular to the front face and bifurcates the top exterior surface. At least two of the programmable tiles are of a common type. The micro-bumps adjacent the front face and coupled to the common type of programmable tiles have a substantially symmetrical orientation relative to a symmetry axis. The symmetry axis being one of (a) collinear with the centerline of the die body, or (b) parallel to the centerline of the die body.

Rotated integrated circuit die and chip packages having the same

An apparatus for a monolithic integrated circuit die is disclosed. In this apparatus, the monolithic integrated circuit die has a plurality of modular die regions. The modular die regions respectively have a plurality of power distribution networks for independently powering each of the modular die regions. Each adjacent pair of the modular die regions is stitched together with a respective plurality of metal lines.

Growing community of inventors

San Jose, CA, United States of America

Rafael C Camarota