Via Technologies, Inc.

Dell USA L.p.

Ip-first, LLC

Via Technology, Inc.

Darius D Gaskins

G Glenn Henry

James R Lundberg

Terry J Parks

Stephan Gaskins

Rodney E Hooker

Charles P Zeller

Charles John Holthaus

Vanessa Canac

Keith D Matteson

Matthew Russell Nixon

Colin Eddy

Gerard M Col

Brent Bean

Michael L Longwell

Raymond A Bertram

Jason Chen

Mark J Brazell

Craig S Jones

Thomas J Kocis

Albert J Loper

Thomas H Holman, Jr

A microprocessor includes a plurality of processing cores and a configuration register configured to indicate whether each of the plurality of processing cores is enabled or disabled. Each enabled one of the plurality of processing cores is configured to read the configuration register in a first instance to determine which of the plurality of processing cores is enabled or disabled and generate a respective configuration-related value based on the read of the configuration register in the first instance. The configuration register is updated to indicate that a previously enabled one of the plurality of processing cores is disabled. Each enabled one of the plurality of processing cores is configured to read the configuration register in a second instance to determine which of the plurality of processing cores is enabled or disabled and generate the respective configuration-related value based on the read of the configuration register in the second instance.

Dynamic reconfiguration of multi-core processor

A multi-core microprocessor is organized into a plurality of resource-associated domains including core domains, group domains, and a global domain. Each domain relates to either local resources, group resources, or global resources that are respectively used by a single core, a group of cores, or all the cores. Each core has its own independently settable target operating state selected from a plurality of possible target operating states that designate configurations for the local resources, group resources, and global resources. Each core is provided with coordination logic configured to implement or request implementation of the core's target operating state, but only to the extent that implementation of the target operating state would not reduce performance of any other core below its own target operating state.

Domain-differentiated power state coordination system

A method is provided for managing power consumption within a multi-core microprocessor. An operating system issues an operating system instruction to transition a recipient core to a targeted power and/or performance state that is one of many possible states into which a microprocessor can place a core. Each core of the microprocessor has its own target state, and different cores may have different target states. After receiving the instruction, the recipient core implements any settings associated with its target core state that wouldn't affect resources shared with other cores. The recipient core also initiates an inter-core discovery process to determine a target multi-core state of all the cores sharing the resource. The target multi-core state reflects one or more settings that match the settings of the recipient core's target core state as much as possible without lowering a performance of any resource-sharing core below that core's own target core state.

Method of managing power consumption within a multi-core microprocessor utilizing an inter-core state discovery process to identify a least power-conserving target core state of all of the cores that share the resource

A multi-die package for a microprocessor provides a power management synchronization system. The package has a plurality of dies. Each die has a plurality of cores, including a single master core. A plurality of sideband non-system-bus inter-die communication wires communicatively couple the dies to each other for a purpose of synchronizing power management. The master core of each die is configured to use one and only one of the inter-die communication wires to transmit power management synchronization messages to each of the other master cores. The master core of each die is also configured to receive power management synchronization messages from each of the other master cores via one or more inter-die communication wires. The cores use this system of inter-die communication wires to synchronize management of resources that affect both the performance and power consumption of the cores.

Power management synchronization messaging system

Microprocessors with multi-core dies that include bypass buses are provided. Each microprocessor comprises a plurality of physical pins for coupling the microprocessor to a processor bus coupled to a chipset. The multi-core die has at least two complementary sets of one or more processing cores, each providing a bus interface coupling respective core inputs and outputs to corresponding processor bus lines. A bypass bus on the die enables cores of the complementary sets to bypass the processor bus and communicate directly with each other. The bypass bus does not carry signals off the die, drive signals on the processor bus to the chipset, or receive chipset-drive signals from the processor bus. Moreover, the microprocessor is operable to detect whether the chipset or a complementary core is driving the processor bus, and if the latter, to select the higher quality bypass bus signals over the corresponding processor bus signals.

Multi-core microprocessor internal bypass bus

Microprocessors are provided with decentralized logic and associated methods for indicating power related operating states, such as desired voltages and frequency ratios, to shared microprocessor power resources such as a voltage regulator module (VRM) and phase locked loops (PLLs). Each core is configured to generate a value to indicate a desired operating state of the core. Each core is also configured to receive a corresponding value from each other core sharing the applicable resource, and to calculate a composite value compatible with the minimal needs of each core sharing the applicable resource. Each core is further configured to conditionally drive the composite value off core to the applicable resource based on whether the core is designated as a master core for purposes of controlling or coordinating the applicable resource. The composite value is supplied to the applicable shared resource without using any active logic outside the plurality of cores.

Distributed management of a shared clock source to a multi-core microprocessor

A first reticle set designed for manufacturing dies with a limited number of cores is modified into a second reticle set suitable for manufacturing at least some dies with at least twice as many cores. The first reticle set defines scribe lines to separate the originally defined dies. At least one scribe line is removed from pairs of adjacent but originally distinctly defined dies. Inter-core communication wires are defined to connect the adjacent cores, which are configured to enable the adjacent cores to communicate during operation without connecting to any physical input/output landing pads of the resulting more numerously cored die, which will not carry signals through the inter-core communication wires off the P-core die. The inter-core communication wires may be used for power management coordination purposes or to bypass the external processor bus.

Multi-core dies produced by reticle set modification

A multi-core processor includes microcode distributed in each core enabling each core to participate in a de-centralized inter-core state discovery process. In a related microcode-implemented method, states of a multi-core processor are discovered by at least two cores participating in a de-centralized inter-core state discovery process. The inter-core state discovery process is carried out through a combination of microcode executing on each participating core and signals exchanged between the cores through sideband non-system-bus communication wires. The discovery process is unmediated by any centralized non-core logic. Applicable discoverable states include target and composite power states, whether and how many cores are enabled, the availability and distribution of various resources, and hierarchical structures and coordination systems for the cores. The inter-core state discovery process may be carried out in accordance with various hierarchical coordination systems involving chained inter-core communications.

Power state synchronization in a multi-core processor

A microprocessor includes an input that receives an indication of the amount of instantaneous power being supplied to the microprocessor by an external power source. The microprocessor includes a plurality of processing cores that each receive the indication from the input and responsively determine an amount of energy consumed by the microprocessor during a preceding period. The period is a predetermined length of time. Each processing core operates at a frequency above a predetermined frequency in response to determining that the amount of energy consumed by the microprocessor during the preceding period is less than a predetermined amount of energy. The predetermined frequency may be: a frequency at which all the cores can operate over the predetermined length of time without the microprocessor consuming more than the predetermined amount of energy, or alternatively the maximum frequency at which system software may request the two or more processing cores to operate.

Microprocessor with multicore processor power credit management feature

A core configuration discovery method and corresponding microprocessor are provided that does not rely on off-core logic or queries by system BIOS. Reset microcode is provided in the microprocessor's cores. Upon reset, the microcode queries and/or receives from other cores configuration-revealing information and collects the configuration-revealing information to determine a composite core configuration for the microprocessor. The composite core configuration may reveal the number of enabled cores, identify the enabled cores, reveal a hierarchical coordination system of the multi-core processor, such as a nodal map of the cores for certain inter-core communication processes or restricted activities, identify various domains and domain masters within such a system, and/or identify resources, such as voltage sources, clock sources, and caches, shared by various domains of the microprocessor. The composite core configuration may be used for power state management, reconfiguration, and other purposes.

Master core discovering enabled cores in microprocessor comprising plural multi-core dies

A microprocessor includes two or more processing cores each configured to compute a first value in response to detecting a power event. The first value represents an amount of energy the core consumed during a time interval leading up to the event. The length of the time interval is predetermined. Each core reads from the memory one or more second values that represent an amount of energy the other cores consume during approximately the time interval. The second values were previously computed and written to the memory by the other cores. Each core adjusts its operating frequency based on the first and second values. The predetermined frequency may be: a frequency at which all the cores can operate over the predetermined length of time without the microprocessor consuming more than the predetermined amount of energy, or alternatively the maximum frequency at which system software may request the cores to operate.

Multicore processor power credit management in which multiple processing cores use shared memory to communicate individual energy consumption

A clock system receiving a reference clock signal via an alignment location and developing a functional clock signal provided to a functional circuit via a clock path. The clock system includes a low bandwidth PLL, a high bandwidth PLL, and a delay path. The low bandwidth PLL receives the reference clock signal and a feedback clock signal and provides a filtered clock signal. The high bandwidth PLL receives the filtered clock signal and provides the functional clock signal, and has a feedback input coupled to its output via a local feedback path. The delay path is coupled between the output of the low bandwidth PLL and the alignment location to provide the feedback clock signal to the low bandwidth PLL. The delay and clock paths are substantially matched. The bandwidths of the low and high bandwidth PLLs may be individually configured to reduce both input jitter and internal jitter, respectively.

Apparatus and method for generating a clock signal with reduced jitter

An apparatus including a JTAG interface, synchronous bus optimizer, core clocks generator, synchronous strobe driver, and a DLL. The JTAG interface receives control information indicating a first amount to advance a synchronous data strobe associated with a first data group and a second amount to delay a data bit signal associated with a second data group. The synchronous bus optimizer receives the control information, and develops a first value on a first ratio bus that indicates the first amount and a second value on a second ratio bus that indicates the second amount. The core clocks generator advances a data strobe clock by the first amount. The synchronous strobe driver employs the data strobe clock to generate the synchronous data strobe, where the synchronous data strobe, when enabled, is advanced also by the first amount. The DLL generates a delayed data bit signal, delayed by the second amount.

Growing community of inventors

Austin, TX, United States of America

Darius D Gaskins