嵌入式異質多核心間溝通之效能實測與最佳化

Abstract

近年來，異質多核心被大量應用於嵌入式系統中以解決耗電問題並增進系統效能。然而異質核心間需要適當的溝通機制，才使得整個系統運作。先前的研究發現這樣的核心間溝通機制(Inter-processor communication)將占去相當多的系統資源，對於需要核心間頻繁溝通的應用程式來說，這樣的溝通程序將造成系統相當大的負擔。因此在本研究中，我們透過一連串的實驗，量測與評估影響異質多核心間溝通機制效能的各項參數。透過實驗的分析，我們得到一些建議設計與參數，透過實際在一個嵌入式系統的案例上實作，發現所提之建議可以大幅降低35%之處理器負擔，提升系統效能。

A heterogeneous multi-core processor composed of a general purpose processor (GPP) which handles the program flow and I/O and a digital signal processor (DSP) which processes mass data is widely used in an embedded system to improve the performance and energy efficiency. To exchange the data between a GPP and DSP, inter-processor communication (IPC) mechanism is required. In this paper, we evaluate the performance of the IPC for an embedded heterogeneous multi-core processor under different design strategies and parameters through a comprehensive experimental study. Based on the experimental results and findings, we further improve the IPC performance of a voice over IP (VoIP) phone as a case study. The experimental results demonstrate that the workload of the GPP significantly reduces 35% without scarifying the functionalities and voice quality of the VoIP system.