Backward Compression Layout for Continuous SoC Trace Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data compression methods for system-on-chip (SoC) debugging fail to ensure continuous data recovery when memory space is limited, as they require complete data to be lost for decompression, leading to discontinuous traced data and inefficient memory utilization.
Innovation Solution
A backward compression system comprising a registering device, an encoding device, and an output selecting device that identifies relationships between input and registered data, encodes data based on these relationships, and selectively outputs encoded or registered data, prioritizing encoded data for overwrite in circular memory to preserve original data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional compression methods store complete data first and then store only differences, then memory space is reduced, but data continuity is lost when complete data is overwritten
Solution Approach 1:
The patent inverts the conventional compression approach by storing difference data first and complete data second in memory, rather than the traditional complete-then-difference order. This inversion ensures that when memory is full and data must be overwritten, the complete data is preserved while only the difference data is overwritten, thereby maintaining data continuity and reliability while still achieving compression goals.
2Quantity of substance
If circular memory is used to maximize space utilization, then memory efficiency improves, but traced data becomes discontinuous
Solution Approach 1:
The patent applies the inversion principle to the storage order in circular memory, storing difference data at lower addresses and complete data at higher addresses. This ensures that during circular overwrite operations, the complete data is preserved longer, maintaining the stability and continuity of the traced data stream while still achieving high memory efficiency through circular buffer utilization.
Solution Approach 2:
The patent segments the stored data into two distinct parts: difference data and complete data, storing them in separate memory regions with difference data at lower addresses and complete data at higher addresses. This segmentation allows independent management and overwrite of each segment, ensuring that complete data is preserved while difference data can be efficiently overwritten, thus maintaining both memory efficiency and data continuity.
3Measurement precision
If complete data is preserved in memory, then data recovery accuracy improves, but memory consumption increases
Solution Approach 1:
The patent segments memory into two functional regions: one for storing difference data and another for storing complete data. This segmentation allows the system to maintain complete data for accurate recovery when needed, while using the difference data region for compression, thereby achieving high data recovery accuracy without proportionally increasing total memory consumption.
Solution Approach 2:
The patent changes the storage parameter from storing only complete data or only difference data to storing both in a specific sequence and ratio. By controlling the amount and placement of complete data versus difference data in memory, the system optimizes the balance between data recovery accuracy and memory consumption, ensuring sufficient complete data is available for accurate recovery while maintaining efficient memory utilization.
Data Source
AI summary
A backward compression system and a backward compression method are provided. By using the system and the method, under limited memory space, a relationship between input data and previous data is identified in real time, the input data or previous data is encoded according to the relationship, and the order of data output is such that the encoded data is output first and the complete data is output second. When the output data is stored in circular memory and the memory gets full, the oldest encoded data is overwritten first, and the complete data is not overwritten until all the related encoded data is overwritten. Therefore, more original data can be preserved after decompression. Thus, system signals can be compressed and utilization of limited memory space can be improved.