Look-Ahead Compression Pipeline for Repeating Substring Hash Conflicts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hardware compression accelerators face reduced data throughput and efficiency due to hash conflicts when processing consecutive repeating characters, as they are forced to process these characters in a serial manner, leading to suboptimal performance.
Innovation Solution
The implementation of an accelerated compression method and apparatus that utilizes a look-ahead buffer and string matching processing pipeline, where substrings with consecutive repeating characters are identified and only a subset of instances are sent to the string matching processing pipeline for matching operations, avoiding unnecessary hash requests until the substrings become distinct, thereby mitigating hash conflicts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all P instances with consecutive repeating characters are sent to the string matching processing pipeline, then complete matching coverage is achieved, but hash conflicts occur and data throughput is reduced
Solution Approach 1:
Multiple identical instances are merged into a single processing path. When P substrings are identified as identical through comparison logic, only one instance is sent to the hash matcher while others are skipped, eliminating redundant hash conflicts and preserving data throughput while maintaining matching coverage through the representative instance.
Solution Approach 2:
Identical instances are extracted from the parallel processing flow and handled separately. The comparison logic detects identical substrings and extracts only the necessary representative instance for hash matching, removing redundant processing paths that cause hash conflicts and throughput reduction.
2Speed
If parallel processing of P instances is maintained, then processing speed is improved, but hash conflicts from identical substrings force serial processing
Solution Approach 1:
The processing mode dynamically adapts based on instance identity. The system maintains parallel processing capability but introduces dynamic conditional logic that switches between parallel and serial handling depending on whether instances are identical. This dynamic adjustment preserves speed while ensuring effective throughput by avoiding redundant serial processing of identical substrings.
3Reliability
If redundant matching operations are performed on identical substrings, then matching accuracy is maintained, but compression efficiency is reduced
Solution Approach 1:
Instead of processing multiple identical instances, the system creates a logical copy approach where one representative instance processes the matching operation and results are effectively copied to all identical instances. This maintains matching accuracy through thorough processing of the representative instance while dramatically improving compression efficiency by eliminating redundant operations.
Data Source
AI summary
An accelerated compression method and apparatus are provided. The accelerated compression apparatus includes a look-ahead buffer, a string matching processing pipeline and a control circuit. A string to be compressed extracted from the data register is stored to the look-ahead buffer. P instances are issued in parallel from the look-ahead buffer. When P substrings corresponding to the P instances issued in a first issue cycle are identical, the control circuit sends a first instance and a second instance of the P instances to the string matching processing pipeline for a matching operation and does not send the remaining instances of these P instances to the string matching processing pipeline. In consecutive issue cycles after the first issue cycle, the control circuit does not send any of the P instances to the string matching processing pipeline until the P substrings corresponding to the P instances are not identical.