Advanced Micro Devices Corporation

Globalfoundries Inc.

Ati Technologies Ulc

Paul C Miranda

Larry David Hewitt

Gerald Robert Talbot

Dale E Gulick

Scott A White

William Alexander Hughes

Derrick R Meyer

Emerson S Fang

Jonathan Mercer Owen

Scott E Swanstrom

Frank P Helms

Richard W Reeves

Ross Voigt La Fetra

Rohit Kumar

Paul Mackey

Philip Ng

Mark Douglas Hummel

Frederick Daniel Weber

Paul J Moyer

Brian D McMinn

William Louie Walker

R Stephen Polzin

Richard Martin Born

Douglas Benson Hunt

Eric Christopher Morton

David E Norris

Stephen C Ennis

Frank Barth

Shwetal Arvind Patel

Elizabeth Morrow Cooper

Kenneth Alan House

Joerg Winkler

Charles K Bachand

James Bunnell

Richard H Lee

A method for steering data for an I/O write operation includes, in response to receiving the I/O write operation, identifying, at an interconnect fabric, a cache of one of a plurality of compute complexes as a target cache for steering the data based on at least one of: a software-provided steering indicator, a steering configuration implemented at boot initialization, and coherency information for a cacheline associated with the data. The method further includes directing, via the interconnect fabric, the identified target cache to cache the data from the I/O write operation. The data is temporarily buffered at the interconnect fabric, and if the target cache attempts to fetch the data while the data is still buffered at the interconnect fabric, the interconnect fabric provides a copy of the buffered data in response to the fetch operation instead of initiating a memory access operation to obtain the data from memory.

I/O writes with cache steering

A technique for characterizing a communications interface includes determining a voltage margin and a timing margin of the interface based on data sampled by a sampling device of a receiver of the interface. In at least one embodiment of the invention, a method for determining margin associated with a receiver circuit of an integrated circuit includes periodically sampling a signal over a time period by a receiver sampling circuit of the receiver circuit to generate a sampled version of the signal. The method includes incrementally varying a value of the parameter associated with the signal. The varying of the parameter is through a range of values of the parameter over the time period. The method includes determining a margin value of the receiver circuit associated with the parameter based, at least in part, on the sampled version of the signal.

Transmitter voltage and receiver time margining

An integrated circuit device includes a first plurality of non-volatile memory locations such as fuses that supply programmed values corresponding to initially selected device features such as voltage, frequency, clock speed, and cache parameters. The device is programmed with a lock value in a second plurality of non-volatile memory locations. That lock value may be a randomly generated number that is unique for each device. After initial programming of the device, access to the device is prevented by appropriately programming access control. In order to unlock the device and modify device features, an unlock key value is supplied to the device. If the unlock key value correctly corresponds to the lock value, the device features can be modified. In that way device features can be modified, but security is maintained to prevent unauthorized modification to device features.

Method and apparatus to control access to device enable features

An improved interface technique for use in a southbridge or I/O hub or in similar devices is provided where non-posted read requests are received from at least one requestor, and upstream commands based on these requests are transmitted. Response data is received in reply to commands that were previously transmitted, and responses are transmitted to the at least one requester based on the response data. A buffer unit is provided for storing command identification data that identifies commands that were already transmitted or that are still to be transmitted, and response availability data that specifies response data that has been received by the receive engine. The improvement may enable multiple outstanding read requests.

Buffering non-posted read commands and responses

Various embodiments of systems and methods for dynamically reallocating buffers used in communicating packets in various communication channels are disclosed. In some embodiments, a method may involve transmitting packets in several communication channels dependent on availability of buffers allocated among the communication channels; tracking a history of packet traffic in each of the communication channels; and dynamically reallocating one or more of the buffers dependent on the history.

Method and apparatus for dynamically reallocating buffers for use in a packet transmission

A technique for reducing crosstalk between communications paths includes scrambling data using scrambling functions that reduce or substantially minimize a probability that worst-case data patterns occur on communications paths adjacent to a potential victim communications path. In at least one embodiment of the invention, a method includes scrambling a plurality of data bits based at least in part on respective ones of a plurality of distinct combinations of one or more taps of a linear feedback shift register (LFSR). The plurality of data bits are scrambled for transmission during a first bit-time on corresponding ones of a plurality of adjacent communications paths.

Method and apparatus to reduce the effect of crosstalk in a communications interface

An in-band configuration technique configures a data communications link for high-speed data communications between at least a first and second integrated circuit using in-band communications between the first and second integrated circuits. The technique configures at least one equalizer of the data communications link with predetermined equalizer settings selected from a plurality of predetermined equalizer settings based on a selected rate of data communications for the link.

In-band method to configure equalization levels

A technique for negotiating the width of a link between a first device and a second device includes detecting, during initialization, a respective signal on one or more control lines associated with at least a portion of an N-bit link. The N-bit link is configured as a single link having a width of N or multiple sublinks having a width less than N based on a respective value of the respective signal on the one or more control lines.

Technique for negotiating a width of a packet-based communication link

An integrated circuit communications interface operable consistent with multiple data transmission protocols includes error detection circuitry that implements a cyclic redundancy check (i.e., CRC) function. The error detection circuitry generates a checksum based, at least in part, on a selected one of the multiple data transmission protocols. The error detection circuitry includes at least one circuit that generates a digital code according to an operation including terms common to the multiple data transmission protocols. That digital code is combined with a selected digital code to generate the CRC. The selected digital code is generated by an individual circuit corresponding to a respective one of the multiple data transmission protocols. The individual circuit generates the selected digital code according to an operation including at least terms exclusive to the respective one of the multiple data transmission protocols.

Error detection in a communications link

A first portion of a communication link is operated in a power savings mode at the same time that a second portion of the communication link is operated in a normal operational mode. For the first portion, a refresh mode is entered from the power savings mode in which one or more training patterns are transmitted over the first portion, while the second portion remains in the normal operational mode. An indication when to activate and deactivate the refresh mode may be sent over the second portion of the communication link. The refresh mode may be periodically entered from the power savings mode based on an interval register specifying the amount of time the communication link should remain in the power savings mode before a refresh occurs. In addition, the amount of time spent in the refresh mode may be programmable.

In-band power management in a communication link

A computer system includes a first and a second integrated circuit coupled by a communication link. The communication link operates in a power savings mode in which data is not transmitted over the link. Periodically, the communication link enters a training phase in which training patterns are transmitted over the communication link for a predetermined time period. The communication link returns to the power savings mode after the predetermined time period has elapsed. At least one sideband signal, separate from the communication link, and coupled between the first and second integrated circuits, is used to indicate when to enter the training phase from the power savings mode and exit the training phase and return to the power savings mode.

Power management in a communication link

A system including a host coupled to a serially connected chain of memory modules. In one embodiment, each of the memory modules includes a memory control hub for controlling access to a plurality of memory chips on the memory module. The memory modules are coupled serially in a chain to the host via a plurality of memory links. Each memory link may include an uplink for conveying transactions toward the host and a downlink for conveying transactions originating at the host to a next memory module in the chain. The uplink and the downlink may convey transactions using packets that include control and configuration packets and memory access packets. The memory control hub may convey a transaction received on a first downlink of a first memory link on a second downlink of a second memory link independent of decoding the transaction.

System including a host connected to a plurality of memory modules via a serial memory interconnect

A message based power management approach is utilized to provide power management for a multi-processor system. A power management message is received at one processor of the multi-processor system over an input/output communication link that provides input/output access for the processors of the multi-processor system. The power management message includes a power management field encoding a power management state for processors of the multi-processor system. The processor that received the power management message over the input/output communication link sends a power management message to other processors in the multi-processor system over one or more inter-processor communication links encoding the power management state.

Message based power management in a multi-processor system

A message based power management system converts legacy signals used in power management, and other signals used to differentiate between power states, to messages sent over a communication link. A system message sent on a communication link includes a field encoding the level of power management for the device receiving the system message. Further, one or more additional signals, separate from the communication link, may be used to indicate when to take action after the power management message has been received.

Message based power management

A host is coupled to a serially connected chain of memory modules. In one embodiment, a method for initializing the host and each of memory modules includes the host transmitting a first synchronization pattern and a second synchronization pattern downstream in response to a reset condition. The method also includes each memory module in the serially connected chain of memory modules receiving and forwarding the first and the second synchronization pattern. Each memory module receives and forwards the first and the second synchronization pattern. Further, the method includes the host transmitting a plurality of NOP packets downstream in response to transmitting the second synchronization pattern. Lastly, the method includes a portion of the memory modules injecting and transmitting NOP packets upstream in response to receiving the second synchronization pattern from downstream.

Growing community of inventors

Austin, TX, United States of America

Paul C Miranda