A DDR memory access address mapping method and access address mapping unit

Abstract

The invention discloses a DDR memory access address mapping method and a memory access address mapping unit. The steps of the memory access address mapping method include: 1) setting corresponding address mapping rules according to the access modes of different hosts to the DDR memory; 2) DDR memory After accessing the access request of the host, select the corresponding address mapping rule according to the access mode required by the host and output: 3) access the address to be mapped and perform mapping according to the address mapping rule output in step 2); The address mapping unit includes a mapping rule setting module, a mapping mode selection module and a memory access address mapping module. The present invention can perform memory access address mapping according to the memory access characteristics of different hosts, so that hosts with different memory access characteristics can realize efficient memory access through a suitable address mapping method, and has simple structure, flexible and efficient memory access address mapping advantage.

Description

technical field [0001] The invention relates to the technical field of DDR (Double Data Rate) memory access, in particular to a DDR memory access address mapping method and a memory access address mapping unit. Background technique [0002] DDR memory is a high-speed memory that works on both edges of the clock. The nominal rate of DDR has developed from DDR's 133Mbps to today's DDR3-2133Mbps, and the higher rate DDR4 has also begun to enter industrial applications. DDR memory is different from the CMOS process used by SRAM (Static Random Access Memory, Static Random Access Memory) memory. DDR memory adopts DRAM dynamic circuit technology, and mostly uses capacitors to store value. Therefore, before reading and writing DDR memory, the data line must be precharged first, and then the read operation is performed; and because the read is a destructive read, it must be written back after reading; at the same time, the existence of leakage current makes the DDR must be kept perio...

AI Technical Summary

Problems solved by technology

In this case, a single fixed address mapping method can no longer achieve "reducing Page switching, and switching Banks when switching Pages", so it is impossible to optimize DDR memory access methods and improve DDR access efficiency

[0007] like figure 2 The address mapping method using high-bit crossover is shown in the figure, which inevitably encounters the problem of page failure; for example image 3 The address mapping method using low-order crossover is shown in the figure, which switches Banks at the same time as switching Pages, which can hide the overhead of closing and activating pages between different Banks. However, this mapping method can only provide a single address mapping data. Storage method, when there are access requirements of diverse applications, it often encounters failure

Images