Method for dynamically allocating on-chip heterogeneous memory resources by utilizing virtual memory mechanism

Abstract

The invention discloses a method for dynamically allocating on-chip heterogeneous memory resources by utilizing a virtual memory mechanism, comprising the following steps of: analyzing the time and space distribution of data sections (comprising global data, stack data, pile data and constant pool data) with data Cache conflicts caused by high frequency by adopting a time slot analyzing method, then obtaining the time slot conflict diagram of the data Cache by the distribution and carrying out mathematical abstraction for the time slot conflict diagram. According to the target function of energy consumption and the target function of performance, a program part to be optimized is selected by using an integer nonlinear programming method and then is dynamically remapped in an on-chip data SPM (Scratch-pad Memory) by utilizing clock interrupt, and thereby, the extra memory access caused by the data Cache conflicts can be avoided. The method can reduce the energy consumption of a system and enhance the performance of the system due to the full utilization of on-chip heterogeneous memory.

Description

technical field [0001] The invention relates to the field of embedded on-chip memory, in particular to a method for dynamically allocating on-chip heterogeneous storage resources (data parts of Cache and SPM) by using a virtual memory mechanism. Background technique [0002] With the development of microelectronic technology, the embedded computing platform based on SoC (System-on-a-Chip) is becoming more and more mature. However, due to the increasing gap between processor speed and external memory speed, the SoC memory subsystem has become the bottleneck of system performance, power consumption and cost. Therefore, how to optimize the architecture and management strategy of the storage subsystem has always been a hot spot in embedded research. [0003] As a traditional on-chip memory, Cache is managed by hardware, transparent to software in most cases, and can automatically load frequently accessed instructions and data into the on-chip memory. However, the shortcomings ...

AI Technical Summary

Problems solved by technology

However, the shortcomings of Cache, such as high power consumption, large footprint, and unpredictable program execution time, have always limited its wide application in embedded systems.

[0008] So far, there is no related research involving the unified analysis of program data, including global data, constant pool, heap and stack, and dynamically mapping them into SPM

Images