Compressed Memory Management for Lower Swap Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory management techniques, such as swap operations to auxiliary storage devices, incur significant performance degradation due to the slow data transfer and retrieval costs associated with moving data between main and auxiliary memory.
Innovation Solution
Implementing a memory management method that includes a near memory processing unit to compress and store data in a compressed memory area within the main memory, reducing the need for external data transfer by managing memory operations locally.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is moved to auxiliary storage device for memory management, then memory shortage is prevented, but data transfer cost and latency increase significantly
Solution Approach 1:
The patent divides the memory management system into two distinct areas within the same memory device: a normal memory area for uncompressed data and a compressed memory area for compressed data. This segmentation allows data to be stored in the same physical memory device, eliminating the need for transfers to auxiliary storage while maintaining memory expansion capabilities.
Solution Approach 2:
The patent introduces a near memory processing unit as an intermediary component that performs compression operations on data before storing it in the compressed memory area. This intermediary processing unit enables efficient data transformation and storage within the same memory device, avoiding external transfers to auxiliary storage.
2Adaptability or versatility
If data is moved between main memory and auxiliary storage, then memory allocation flexibility is improved, but application performance degrades
Solution Approach 1:
The patent merges the normal memory area and compressed memory area into a single memory device, allowing the system to allocate memory flexibly between uncompressed and compressed regions while keeping all data accessible within the same high-speed memory architecture. This maintains application performance by eliminating slow auxiliary storage transfers.
Solution Approach 2:
The patent changes the compression parameter of data stored in memory, creating a compressed memory area where data is stored in compressed form. This parameter change allows more data to be stored in the same physical space while maintaining fast access speeds, as the compressed data remains within the high-speed memory device rather than being moved to slow auxiliary storage.
3Quantity of substance
If compression is applied to data in memory, then memory efficiency is improved, but processing complexity increases
Solution Approach 1:
The patent extracts the compression function from the main application processing path and places it in a dedicated near memory processing unit. This extraction allows compression operations to be performed independently without adding complexity to the main application logic, while still achieving improved memory efficiency through compressed data storage.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A memory management method based on a compressed memory and an apparatus using the same are provided. A memory device includes a memory and a near memory processing unit. The memory includes a normal memory area for storing uncompressed data and a compressed memory area for storing compressed data. The near memory processing unit is configured to receive a first command to frontswap-store a huge page, which is stored in the normal memory area, in the compressed memory area; identify addresses of sub-pages of the huge page based on an address of the huge page; compress the sub-pages using the addresses of the sub-pages to generate compressed sub-pages; and store the compressed sub-pages in the compressed memory area. A size of the huge page is larger than a regular page.