Intel Corporation

Arijit Biswas

Michael D Powell

Guy M Therien

Guy G Sotomayor

Shubhendu Sekhar Mukherjee

Venkatesh Ramani

Eric J DeHaemer

Steven E Raasch

Joel S Emer

Ian M Steiner

Reid James Riedlinger

Paul B Racunas

Glenn J Hinton

Raanan Sade

Stanislav Shwartsman

Deborah T Marr

Ronak Singhal

Larisa Novakovsky

Adrian C Moga

John G Holm

Sebastien Hily

Niranjan Soundararajan

Sudhanva Gurumurthi

Apparatuses, methods and storage medium for computing including determination of work placement on processor cores are disclosed herein. In embodiments, an apparatus may include one or more processors, devices, and/or circuitry to identify a favored core of the processor cores. The one or more processors, devices, and/or circuitry may be configured to determine whether to migrate a thread to or from the favored core. In some embodiments, the determination may be by a process executed by a driver and/or by an algorithm executed by a power control unit of the processor.

Systems, methods and devices for determining work placement on processor cores

A multicore processor may include multiple processing cores that were previously designated as active cores and at least one processing core that was previously designated as a functional spare. The processor may include an interface to receive, during operation of the processor in an end-user environment, a request to change the designation of at least one of the processing cores. The processor may be to store, into a desired cores configuration data structure in response to the request, data representing a bitmask that reflects the requested change, and to execute a reset sequence. During the reset sequence, the processor may activate, dependent on the bitmask, a processing core previously designated as a functional spare, or may deactivate, dependent on the bitmask, a processing core previously designated as an active core. The processor may include a predetermined maximum number of active cores and a predetermined minimum number of functional spares.

Systems and methods for in-field core failover

Work can be migrated between processor cores. For example, a thread causing a heavy load on a first core can be detected. A power control unit can determine to migrate the thread from the first less-efficient core to the second more-efficient core. The power control unit can request that the first core and the second core transition to a low-power state (e.g., a sleep state, a C6 power state, etc.). The first core can transfer its software context to a first core software context storage, halt and quiesce. The second core can halt and quiesce. The software context from the first core software context storage can be transferred to a second core software context storage of the second core. A processing core identifier of the first core can be assigned to the second core. The power control unit can then request the second core to transition to an active state (such as a C0 state).

Systems, methods and devices for work placement on processor cores

A disclosed method includes accessing characterization data indicating first and second sets of performance characteristics for first and second processing cores of a processor; determining, based on a performance objective and the characterization data, a first power state for the first processing core and a second power state for the second processing core; and applying the first power performance objective to the first processing core and the second power performance objective to the second processing core.

Exploiting process variation in a multicore processor

In one embodiment, the present invention includes a method for providing a cache block in an exclusive state to a first cache and providing the same cache block in the exclusive state to a second cache when cores accessing the two caches are executing redundant threads. Other embodiments are described and claimed.

Obtaining data for redundant multithreading (RMT) execution

A system, processor and method to monitor specific cache events and behavior based on established principles of quantized architectural vulnerability factor (AVF) through the use of a dynamic cache write policy controller. The output of the controller is then used to set the write back or write through mode policy for any given cache. This method can be used to change cache modes dynamically and does not require the system to be rebooted. The dynamic nature of the controller provides the capability of intelligently switching from reliability to performance mode and back as needed. This method eliminates the residency time of dirty lines in a cache, which increases soft errors (SER) resiliency of protected caches in the system and reduces detectable unrecoverable errors (DUE), while keeping implementation cost of hardware at a minimum.

Dynamic cache write policy

A processor includes a core to execute instructions and a cache memory coupled to the core and having a plurality of entries. Each entry of the cache memory may include a data storage including a plurality of data storage portions, each data storage portion to store a corresponding data portion. Each entry may also include a metadata storage to store a plurality of portion modification indicators, each portion modification indicator corresponding to one of the data storage portions. Each portion modification indicator is to indicate whether the data portion stored in the corresponding data storage portion has been modified, independently of cache coherency state information of the entry. Other embodiments are described as claimed.

Method and apparatus for error correction in a cache

An apparatus and method is described herein for replacing faulty core components. General purpose hardware is provided to replace core pipeline components, such as execution units. In the embodiment of execution unit replacement, a proxy unit is provided, such that mapping logic is able to map instruction/operations, which correspond to faulty execution units, to the proxy unit. As a result, the proxy unit is able to receive the operations, send them to general purpose hardware for execution, and subsequently write-back the execution results to a register file; it essentially replaces the defective execution unit allowing a processor with defective units to be sold or continue operation.

General purpose hardware to replace faulty core components that may also provide additional processor functionality

Embodiments of an invention for synchronizing redundant processors using state history are disclosed. In one embodiment, an apparatus includes two processors, state storage for each processor, and control logic. Each processor is to execute the same instructions. The state storage is to store compressed processor state information for each instruction executed by the processors. The control logic is to synchronize the two processors based on entries from the state storage.

Growing community of inventors

Holden, MA, United States of America

Arijit Biswas