17899171. RECALL PENDING CACHE LINE EVICTION simplified abstract (Micron Technology, Inc.)
Contents
RECALL PENDING CACHE LINE EVICTION
Organization Name
Inventor(s)
Dean E. Walker of Allen TX (US)
Tony M. Brewer of Plano TX (US)
RECALL PENDING CACHE LINE EVICTION - A simplified explanation of the abstract
This abstract first appeared for US patent application 17899171 titled 'RECALL PENDING CACHE LINE EVICTION
Simplified Explanation
- System and techniques for recall pending cache line eviction are described - Queue with deferred memory request for cache line - Metadata stored in cache line tag - Metadata written to first recall storage when recall is needed - Memory request ID written to second recall storage referenced by message ID of recall request - Queue for cache line can be restored using message ID in response to lookup memory request ID from second recall storage, then lookup metadata from first recall storage, and write metadata into tag for cache line
Potential Applications
- Improving cache management in computer systems - Enhancing memory performance in high-demand applications
Problems Solved
- Efficiently managing cache line eviction - Ensuring data integrity during cache recalls
Benefits
- Improved system performance - Reduced latency in memory access - Enhanced reliability in cache management
Original Abstract Submitted
System and techniques for recall pending cache line eviction are described herein. A queue that includes a deferred memory request is kept for a cache line. Metadata for the queue is stored in a cache line tag. When a recall is needed, the metadata is written from the tag to a first recall storage, referenced by a memory request ID. After the recall request is transmitted, the memory request ID is written to a second recall storage referenced by the message ID of the recall request. Upon receipt of a response to the recall request, the queue for the cache line can be restored by using the message ID in the response to lookup the memory request ID from the second recall storage, then using the memory request ID to lookup the metadata from the first recall storage, and then writing the metadata into the tag for the cache line.
