CACHE BLOCK STORE INSTRUCTION OPERATIONS WHERE CACHE COHERENCY IS ACHIEVED WITHOUT WRITING ALL THE WAY BACK TO MAIN MEMORY

Patent №

US 5,909,698

Granted

1999-06-01

Filed 1997

Owner

INTERNATIONAL BUSINESS MACHINES CORPORATION

AI components

1

hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

08818845

In a processor employing separate instruction and data caches in at least one cache hierarchy level, a cache control instruction forces modified data within the separate data cache to a lower cache hierarchy level. An existing cache access attribute is employed to distinguish between occasions when the data must be written all the way to main memory and occasions when the data need only be written to a cache hierarchy level from which fetches are made to the separate instruction cache. Thus, the separate instruction and data caches may be made coherent without writing all the way to main memory, but the ability to write modified data to main memory whenever necessary is preserved. Utilization of the existing attribute avoids increasing processor complexity and/or resources.

AI hardwareG06F 12/0804G06F 12/0808

AI classification

AI hardware0.93
Natural language0.01
Planning0.01
Knowledge representation0.00
Vision0.00
Machine learning0.00
Speech0.00
Evolutionary computation0.00

Ownership

INTERNATIONAL BUSINESS MACHINES CORPORATION

assignment · 85900227

Assignors

ARIMILLI, RAVI K., DODSON, JOHN S., LEWIS, JERRY D.

On an employer assignment, the assignors are typically the inventors.

From the same owner

© 2026 NYSGPT2525 LLC