International Business Machines Corporation

David L Arlington

Hehching Harry Li

Chin-Long Chen

David John Krolak

Bruce G Hazelzet

Edward Kelley Evans

Bharat J Oza

Jerry K Radcliffe

A Frank Weaver

Jacqueline M Cole

Jacqueline K Morris

An apparatus and method for polling a plurality of components in a data processing system includes a sensor for determining when a request for power up has been made, a series of parallel circuits extending between a power sequence monitor/power control device and individual computer system components, a configuration bus connected between the system components and the power sequence monitor/power control device and comparison circuitry connected to the configuration bus for comparing a signal sent from the power sequence monitor/power control device, through the configuration bus and system components, and returned to the power sequence monitor/power control device, with a predetermined code for identifying and locating the system components.

Preliminary polling for identification and location of

A 'smart' memory card architecture and interface provides significantly increased performance, in part, by using fast access dynamic random access memory (DRAM) technologies which allows up to 8-byte data transfers from the memory card every 27ns after the initial access. The 27ns transfer rate includes the time required for error correction code (ECC), parity generation, and other reliability functions. Only two complementary metal oxide seminconductor (CMOS) integrated circuit (IC) logic chips or modules provide all the function required. The simplicity and flexibility afforded by the 'smart' memory card approach provides a means to allow one card interface to be used with a broad range of hardware technologies and in different systems. The architecture of the memory card provides a full range of direct and partial store operations in a manner transparent to the system.

Smart memory card architecture and interface

An extended error code particularly applicable to a code that can correct any number of errors in one sub-field but can only detect the existence of any number of errors in two sub-fields. If the initial pass of the data through the error correction code indicates an uncorrected error, the data is complemented and restored in the memory and then reread. The retrieved data is recomplemented and again passed through the error correction code. If an uncorrected error persists, then a bit-by-bit comparision is performed between the originally read data and the retrieved complemented data to isolate the hard fails in the memory. The bits in the sub-field associated with the hard fail are then sequentially changed and then the changed data word is passed through the error correction code. A wrong combination is detected by the error correction code. The sequential changing continues until the bits in the sub-field associated with the hard fail match the originally stored data, in which case the error correction code can correct the remaining errors in the remaining sub-fields.

Extended error correction for package error correction codes

Growing community of inventors

Eggertsville, NY, United States of America

David L Arlington