Compression Dictionary Generation Using Accelerator Offload
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computational systems face challenges in efficiently managing cryptographic and compression tasks, leading to increased processor load and reduced efficiency.
Innovation Solution
Implementing accelerator devices to offload dictionary creation and data compression tasks, utilizing dictionaries to replace data patterns with shorter representations, thereby reducing processor burden and improving compression ratios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the processor performs cryptographic and compression tasks directly, then the computational accuracy is maintained, but the processor load increases and efficiency decreases
Solution Approach 1:
The patent extracts compression and cryptographic tasks from the main processor and relocates them to dedicated accelerator devices. The processor initiates these tasks by writing control blocks to memory, while the actual compression operations are performed by separate compression accelerators, thereby reducing processor load and improving overall system efficiency.
Solution Approach 2:
The patent introduces control blocks as intermediary data structures that mediate between the processor and compression accelerators. These control blocks contain all necessary parameters and state information, allowing the processor to offload tasks while maintaining control and coordination through memory-based communication.
2Productivity
If a generic dictionary is used for compression, then the device complexity is reduced, but the compression ratio deteriorates
Solution Approach 1:
The patent implements dynamic dictionary selection where the system automatically chooses between generic and custom dictionaries based on the data characteristics and compression requirements. Custom dictionaries can be generated on-demand for specific data sets, providing adaptive optimization without permanently increasing system complexity.
Solution Approach 2:
The patent changes the parameter of dictionary specificity dynamically. Instead of using a fixed generic dictionary, the system can generate custom dictionaries with parameters optimized for specific data patterns, thereby improving compression ratios when needed while falling back to simpler generic dictionaries when appropriate.
3Productivity
If custom dictionaries are generated for different data sets, then the compression ratio is improved, but the time and computational resources required increase
Solution Approach 1:
The patent performs preliminary dictionary generation by creating custom dictionaries in advance for known data sets or data types. These pre-generated dictionaries are stored and can be directly applied during compression operations, eliminating the need for real-time dictionary generation and reducing compression time.
Solution Approach 2:
The patent applies partial dictionary customization by generating custom dictionaries only for portions of data that benefit from it, rather than generating dictionaries for all data. This selective approach reduces the overall time and computational resources required while still achieving compression improvements where needed.
Data Source
AI summary
Examples described herein relate to an accelerator configured to: based on receipt of a request to generate a dictionary: generate a dictionary for data compression based on data associated with the request and store the dictionary in a memory device and based on receipt of a second request to generate a second dictionary and compress second data. In some examples, generating the second dictionary for compression of the second data includes loading the second dictionary into a history buffer, compressing the second data, and storing the compressed second data into the memory device.