| 摘要 |
A method and structure for reducing access latency and contention in a processing system is disclosed. The invention detects when the amount of available memory is outside a prescribed range, and responsively selects data blocks for compression (to add to the available memory,) or decompression (to use surplus available memory for uncompressed data,) until the amount of available memory is within the prescribed range. When data blocks are compressed, a DOC is determined and stored as an attribute in the directory entry associated with the data block. A most recently used list of recent data block addresses prevents those, as well those data blocks with poor DOC attributes, from being selected for recompression. All zero data blocks are detected to avoid standard compression/decompression overhead. |
