Intel Corporation

Altera Corporation

Tahoe Research, Ltd.

Md Altaf Hossain

Ankireddy Nalamalpu

Dheeraj R Subbareddy

Mahesh K Kumashikar

Scott A Gilbert

Robert L Sankman

Chee Hak Teh

Lai Guan Tang

Atul Maheshwari

Ravindranath Vithal Mahajan

Mahesh A Iyer

Sandeep B Sane

Anshuman Thakur

Yuet Li

Gregg William Baeckler

Sean R Atsatt

Debendra Mallik

Rajat Agarwal

Ravi Prakash Gutala

Scott J Weber

Ram S Viswanath

Aravind Raghavendra Dasu

Sriram Srinivasan

Nagi Aboulenein

David W Browning

Chooi Pei Lim

Jin Zhao

Kevin Joseph Doran

Jeffrey Chromczak

Brian P Kelly

Cliff C Lee

Itai M Pines

Dinesh Somasekhar

Min Suet Lim

Eng Huat Goh

Rahul Pal

Dheemanth Nagaraj

Sreedhar Ravipalli

Daniel S Klowden

Kemal Aygün

John T Vu

Jayapratap Bharathan

Casey Thielen

Dheeraj Subareddy

Rajesh Vivekanandham

The present disclosure describes programmable logic that may be operated in a turbo processing mode to cause an ongoing operation to be completed faster than a scheduled completion time. With at least some of the remaining time to the scheduled completion time, power savings may be realized by operating the programmable logic into a deep sleep mode, where configuration memory associated with the programmable logic may be set to a suitable voltage level as to not cause data loss at lower or zero voltage levels but otherwise realize power savings relative to an amount of power consumed during average processing operations.

Workload-dependent integrated circuit operation based on power headroom

A circuit system includes an accelerator circuit and a compute circuit. The accelerator circuit generates a request in response to receiving packets of data. The accelerator circuit generates an indication of a low power state based on a reduced number of the packets of data being received. The compute circuit performs a processing operation for the accelerator circuit using the packets of data in response to receiving the request. The compute circuit comprises a power management circuit that decreases a supply voltage in the compute circuit and decreases a frequency of a clock signal in the compute circuit in response to the indication of the low power state from the accelerator circuit.

Techniques for power management in compute circuits

Systems or methods of the present disclosure may provide for determining a loadline for operation of a programmable logic fabric where the loadline is based at least in part on design configuration details for a design or a configuration rather for generic deployment of the programmable logic device. The loadline may be determined using software modeling for the design or configuration. Additionally or alternatively, the loadline may be determined using runtime testing and sensing of real-world parameters. This determination based on real-world parameters of a deployment of the configuration or design is based on a determination of a step load for the design or configuration.

Dynamic loadlines for programmable fabric devices

Systems or methods described herein may relate to latch-independent clock gating techniques to enable or disable an internal clock of an integrated circuit device. A programmable logic device includes a clock gating circuit that receives a clock signal and is latch independent. The clock gating circuit includes gating signal circuitry that generates a gating signal based on the clock signal and an enable signal. The clock gating circuit also includes a logic gate that generates a control signal based on the gating signal. The clock gating circuit also includes gated clock generation circuitry that generates a gated clock signal based on the clock signal and the control signal.

Self-gating flops for dynamic power reduction

Systems or methods of the present disclosure may provide efficient power consumption for programmable logic devices based on reducing guardband voltages. A programmable logic device may include circuit monitors to mimic critical paths of an implemented circuit design and generate timing information based on the critical paths. A controller on the programmable logic device may adjust the voltage guardband based on the timing information.

Systems and methods to reduce voltage guardband

A semiconductor device may include a programmable fabric and a processor. The processor may utilize one or more extension architectures. At least one of these extension architectures may be used to integrate and/or embed the programmable fabric into the processor as part of the processor. Systems and methods for transitioning data between the programmable fabric and the processor associated with different clock domains is described.

Flexible instruction set architecture supporting varying frequencies

Systems or methods of the present disclosure may provide for operating a programmable fabric including multiple programmable elements organized into a number of power domains that utilize a common voltage within the respective power domains. A current sensor senses a current of the programmable fabric. When the sensed current has crossed a threshold, the programmable fabric changes the number of power domains.

Switch based on load current

Embodiments disclosed herein include electronic packages. In an embodiment, the electronic package comprises, a package substrate, an interposer on the package substrate, a first die cube and a second die cube on the interposer, wherein the interposer includes conductive traces for electrically coupling the first die cube to the second die cube, a die on the package substrate, and an embedded multi-die interconnect bridge (EMIB) in the package substrate, wherein the EMIB electrically couples the interposer to the die.

Interconnect architecture with silicon interposer and EMIB

Systems or methods of the present disclosure may provide efficient electric power consumption of programmable logic devices based on providing different voltage levels to different portions (e.g., voltage islands) of the programmable logic device. For example, the programmable logic device may include circuitry to provide different voltage levels to different voltage islands. The programmable logic device may implement and operate logic configurations with different operating parameters using different operating voltages for efficient electric power consumption.

Power management using voltage islands on programmable logic devices

The present disclosure is directed to 3-D stacked architecture for Programmable Fabrics and Central Processing Units (CPUs). The 3-D stacked orientation enables reconfigurability of the fabric, and allows the fabric to function using coarse-grained and fine-grained acceleration for offloading CPU processing. Additionally, the programmable fabric may be able to function to interface with multiple other compute chiplet components in the 3-D stacked orientation. This enables multiple compute components to communicate without the need for offloading the data communications between the compute chiplets.

Three-dimensional stacked programmable logic fabric and processor design architecture

An integrated circuit includes a package substrate that includes first and second electrical traces. The integrated circuit includes first, second, third, and fourth configurable dies, which are mounted on the package substrate. The first and second configurable dies are arranged in a first row. The third and fourth configurable dies are arranged in a second row, which is approximately parallel to the first row. The first and third configurable dies are arranged in a first column. The second and fourth configurable dies are arranged in a second column, which is approximately parallel to the first column. The first electrical trace couples the first and third configurable dies, and the second electrical trace couples the second and third configurable dies. The second electrical trace is oblique with respect to the first electrical trace. The oblique trace improves the latency of signals transmitted between dies and thereby increases the circuit operating speed.

Innovative interconnect design for package architecture to improve latency

A programmable device may have logic circuitry formed in a top die and memory and specialized processing blocks formed in a bottom die, where the top die is stacked directly on top of the bottom die in a face-to-face configuration. The logic circuitry may include logic sectors, logic array blocks, logic elements, and other types of logic regions. The memory blocks may include large banks of multiport memory for storing data. The specialized processing blocks may include multipliers, adders, and other arithmetic components. The logic circuitry may access the memory and specialized processing blocks via an address encoded scheme. Configured in this way, the maximum operating frequency of the programmable device can be optimized such that critical paths will no longer need to traverse any unused memory and specialized processing blocks.

Programmable logic device with fine-grained disaggregation

Systems or methods of the present disclosure may improve scalability (e.g., component scalability, product variation scalability) of integrated circuit systems by disaggregating periphery intellectual property (IP) circuitry into modular periphery IP tiles that can be installed as modules. Such an integrated circuit system may include a first die that includes programmable fabric circuitry and a second die that includes a periphery IP tile. The periphery IP tile may be disaggregated from the programmable fabric die and may be communicatively coupled to the first die via a modular interface.

Modular periphery tile for integrated circuit device

A processor package module comprises a substrate, one or more compute die mounted to the substrate, and one or more photonic die mounted to the substrate. The photonic die have N optical I/O links to transmit and receive optical I/O signals using a plurality of virtual optical channels, the N optical I/O links corresponding to different types of I/O interfaces excluding power and ground I/O. The substrate is mounted into a socket that support the power and ground I/O and electrical connections between the one or more compute die and the one or more photonic die.

Processor package with universal optical input/output

An integrated circuit die includes input buffer circuits that are enabled during an input mode of operation in response to first control signals to transmit input signals into the integrated circuit die from conductive bumps. Each of the input buffer circuits is coupled to one of the conductive bumps. The integrated circuit die also includes output buffer circuits that are each coupled to one of the conductive bumps. The output buffer circuits are enabled during an output mode of operation in response to second control signals to transmit output signals from the integrated circuit die to the conductive bumps. The input buffer circuits are disabled from transmitting signals during the output mode of operation in response to the first control signals. The output buffer circuits are disabled from transmitting signals during the input mode of operation in response to the second control signals.

Techniques for clock signal transmission in integrated circuits and interposers

Growing community of inventors

Portland, OR, United States of America

Md Altaf Hossain