Semi-Sorted Cache Tag Compression for Predictable Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing compression algorithms for cache lines yield inconsistent compression ratios, complicating hardware cache design and verification, and often exceed latency thresholds, making them impractical for implementation in cache hierarchies.
Innovation Solution
The semi-sorting compression method uses encoding and decoding tables to create a code word representing compressed prefix values, allowing for consistent compression and decompression of data items with high entropy, while operating within latency constraints by sorting data items based on prefixes and storing suffixes separately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional compression algorithms are used on cache tags, then storage space is reduced, but compression ratio variability increases and hardware design complexity increases
Solution Approach 1:
The patent applies preliminary sorting of cache tags before compression. By sorting the tags in advance based on their prefix values, the compression algorithm can operate on already-ordered data, which enables consistent compression ratios and simplifies the hardware design by eliminating the need to handle variable-length compressed tags.
Solution Approach 2:
The patent segments the cache tag into a prefix portion and a suffix portion. The prefix is compressed using the semi-sorting compression algorithm while the suffix is handled separately. This segmentation allows the compressed prefix to have consistent length, simplifying the overall tag structure and hardware design.
2Quantity of substance
If conventional compression algorithms are used on cache tags, then storage space is reduced, but latency increases beyond acceptable thresholds
Solution Approach 1:
The sorting operation is performed in advance before the compression step. This preliminary action reorganizes the data so that the actual compression process becomes faster and more predictable, reducing the latency impact on the cache operation.
Solution Approach 2:
The patent changes the approach from using complex general-purpose compression algorithms to a specialized semi-sorting compression algorithm that operates on sorted data with specific parameters (prefix/suffix division). This parameter change enables the compression to meet latency requirements while achieving consistent compression ratios.
3Quantity of substance
If compression algorithms with high compression ratios are used, then storage efficiency improves, but implementation simplicity decreases
Solution Approach 1:
By performing the sorting operation beforehand, the patent simplifies the compression implementation. The compression algorithm only needs to process already-sorted data with a straightforward prefix-suffix division, making the hardware implementation simpler while maintaining good compression efficiency.
Solution Approach 2:
The patent changes the compression approach to work with sorted data and fixed prefix lengths. This parameter change transforms a complex variable-length compression problem into a simpler fixed-structure compression task that is easier to implement in hardware while achieving consistent compression ratios.
Data Source
AI summary
A data storage system performs partial compression and decompression of a set of memory items. The memory items each include a data block and a tag with a prefix making up at least part of the tag. The memory items are ordered based on the prefixes. A code word is created containing compressed information representing values of the prefixes for the set of memory items. The code word and block data for each of the memory items are stored in a memory. The code word is decompressed to recover the prefixes.