Data caching method, cache and computer system

Abstract

Embodiments of the present invention provide a data cache method, cache and computer system. In the embodiment of the present invention, when the Cache needs to determine the Cache line to be replaced when the access request misses, not only the historical access frequency of the Cache line, but also the memory type corresponding to the Cache line should be considered, so that the replacement corresponding to the DRAM memory type can be prioritized The Cache line reduces the cache amount of data stored in DRAM by Cache, so that Cache can increase the cache amount of data stored in NVM, so that the access requests for data stored in NVM can be in Cache as much as possible Find the corresponding data, thereby reducing the situation of reading data from NVM, reducing the delay of reading data from NVM, and effectively improving access efficiency.

Description

technical field [0001] Embodiments of the present invention relate to storage technology, and in particular, to a data cache method, cache and computer system. Background technique [0002] At present, more and more applications are centered on data, such as Internet applications and big data applications. These applications require powerful storage support. [0003] In the prior art, a dynamic random-access memory (Dynamic Random-Access Memory, hereinafter referred to as DRAM) is generally used as a computer system. However, limited by the process, DRAM has a small capacity and high energy consumption, and it is difficult to meet the application requirements for large capacity and low energy consumption. In recent years, non-volatile memory (Non-Volatile Memory, hereinafter referred to as: NVM) has been widely used. It has the advantages of large storage capacity and low energy consumption. Using NVM to replace DRAM as a computer system can meet the needs of applications ...

AI Technical Summary

Problems solved by technology

But compared with DRAM, NVM has a longer read and write latency

Images