Memory Offloading via Data Compression for Bandwidth Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory systems face limitations in efficiently managing and processing data, particularly in scenarios with high demand for simultaneous data access and task execution, leading to bottlenecks and reduced system performance.
Innovation Solution
The memory system incorporates an offloading mechanism using a zSwap process, where data is compressed before being transferred to a peripheral memory device, reducing the size of data to be swapped and minimizing memory space usage and bandwidth consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is transferred uncompressed between memory devices, then data access speed is improved, but memory space usage and bandwidth consumption increase
Solution Approach 1:
The patent applies compression algorithms to change the physical state of data from uncompressed to compressed form, reducing the volume of data stored in external memory while maintaining the ability to restore original data when needed
Solution Approach 2:
Data is pre-compressed before being transferred to external memory, so that when data needs to be accessed, the compressed version is already in place and can be quickly decompressed and transferred back, reducing both storage requirements and transfer time
2Quantity of substance
If data is compressed before transfer to external memory, then memory space usage is reduced, but processing time and system complexity increase
Solution Approach 1:
The compression and decompression functions are extracted from the main host system and implemented as separate memory processing units that can operate autonomously on the memory devices themselves, reducing the processing burden on the host CPU
Solution Approach 2:
Memory processing units act as intermediaries between the host and external memory, handling compression, decompression, and data transfer operations, thereby simplifying the host's role and reducing overall system complexity
3Quantity of substance
If more data is kept in external memory, then main memory capacity is effectively increased, but bandwidth consumption and power usage increase
Solution Approach 1:
Data is compressed to reduce its volume, allowing more data to be stored in external memory with the same physical bandwidth, thereby reducing the frequency and volume of data transfers required and lowering power consumption
Solution Approach 2:
Only the portions of data that need to be accessed are decompressed and transferred back to main memory, rather than continuously managing large amounts of uncompressed data, reducing unnecessary bandwidth consumption and power usage
Data Source
AI summary
A memory system includes an external memory device and a first memory processing unit of a main memory device. The external memory device and a host may be disposed on different boards, and the external memory device may transfer compressed data to the host in response to a swap-in request from the host. The first memory processing unit and the host may be disposed on the same board. The first memory processing unit may load the compressed data, decompress the compressed data to obtain decompressed data, and store the decompressed data in a main memory of the main memory device.