Intel Corporation

Spansion Llc.

Richard Carmichael

Nathan Marushak

Richard C Beckett

Roger C Jeppsen

Devicharan Devidas

Systems and/or methods that facilitate that facilitate error correction of data stored in memory components, such as flash memory devices are presented. An optimized correction component can be used to break data into two or more data blocks. The optimized correction component can facilitated creating one or two redundancy blocks that can be associated with the data blocks, wherein data blocks and the redundancy blocks can be assembled into a data stripe that can be stored in three or more of the memory components. Upon retrieval of the data stripe, the optimal correction component, an error correction code (ECC) component or a combination thereof can correct data blocks that contain errors wherein the decision whether the optimized correction component or the error correction code (ECC) component corrects the errors can be based in part on a predetermined criteria.

System and method for optimized error correction in flash memory arrays

Systems and/or methods that facilitate reading data from a memory component associated with a network are presented. A pre-fetch generation component generates a pre-fetch request based in part on a received read command. To facilitate a reduction in latency associated with transmitting the read command via an interconnect network component to which the memory component is connected, the pre-fetch request is transmitted directly to the memory component bypassing a portion of the interconnect network component. The memory component specified in the pre-fetch request receives the pre-fetch request and reads the data stored therein, and can store the read data in a buffer and/or transmit the read data to the requester via the interconnect network component, even though the read command has not yet reached the memory component. The read data is verified by comparison with the read command at a convergence point.

Clock encoded pre-fetch to access memory data in clustering network environment

A method for error handling of corrupted repeating primitives during frame reception is disclosed. The method comprises identifying a portion of a received frame including a repeating primitive sequence, determining whether data in the repeating primitive sequence has one or more errors, and indicating a successful reception of the received frame with the one or more errors in the repeating primitive sequence if the number of errors is less than a determined threshold. Other embodiments are also disclosed.

Mechanism for error handling of corrupted repeating primitives during frame reception

One embodiment of a method may include, in response, at least in part, to one or more received frames, generating an interrupt and preventing transmission of one or more other frames. The one or more received frames may indicate, at least in part, an error condition or a commencement of a data transfer. The method of this embodiment also may include, in response, at least in part, to the interrupt, executing one or more instructions. The one or more instructions, when executed, may optionally result in deleting the one or more other frames, and if the one or more received frames indicate, at least in part, the error condition, commencing recovery from the error condition. If the one or more received frames indicate, at least in part, the commencement of the data transfer, the method of the embodiment may include storing data associated with the data transfer.

Growing community of inventors

Phoenix, AZ, United States of America

Richard Carmichael