一种内存高速读写的优化方法
By dividing the memory address space into multiple address units using existing technologies, and setting up a local uncommitted area and a stable mirror record in each unit, the problems of merging write requests and read consistency in the memory access path are solved, achieving efficient memory read and write operations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUIRONG ELECTRONIC SYST ENG CO LTD
- Filing Date
- 2026-04-30
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies fail to effectively address the issue of merging consecutive write requests for the same address unit before physical commit in memory access paths. This leads to an increase in the number of repeated physical writes, insufficient separation of read and write paths, inconsistent read results, and excessively long consistency control links.
By dividing the memory address space according to a preset address granularity, multiple address units are formed, and a local area to be committed is set in each unit to record write requests. This enables the merging of write requests and the priority response to read requests. Stable mirroring and version identification mechanisms are used to ensure the isolation and consistency of read and write operations.
It reduces the number of repeated physical writes, improves write throughput, ensures the consistency of read results and the stability of data state, and is suitable for high-concurrency, low-latency memory read and write scenarios.
Smart Images
Figure CN122131986B_ABST