以圖形應用為主的用戶端多核心嵌入式系統---子計畫三：用戶端多核心嵌入式系統程式發展工具與系統模擬 (I)

Abstract

未來的行動裝置上, 決勝點將在於是否有高品質的三維圖型輸出, 高效能, 低耗電, 和適應執行環境的能力. 而未來高階的行動裝置系統架構上將有數個同質多核心的CPU, 搭配GPU, 同時CPU本身會支援一些SIMD指令集. 所以在三維圖型的處理上可以有很多不同的做法. 以混合CPU/GPU來進行進階全域照明的繪圖法(結合光線追蹤與傳統rendering的優點),不但提高圖形品質而且較容易依系統執行時的效能, 網頻, 耗電, 供電情況作調整. 在這方法上,以光線追蹤為基礎的計算可利用同質多核心的MIMD特性, 而需要預先計算rasterization則交由GPU處理, 或是為省電計, 甚至可以交由CPU利用SIMD指令再配上MIMD來處理. 為了支援這一類的三維繪圖法的架構, 本計畫提供系統軟體開發工具和系統模擬，以充分發揮這一新系統架構所帶來的助益和效能. 平行化編譯器(包括同質多核心和異質多核心),執行時的動態管理優化系統, 和系統模擬器是本計劃中需要研究開發最主要的工具, 這些工具擔負起銜接圖型應用程式, 效能與耗能評估與管理的工作

For a market successful mobile device, the determining factor would be if the product has high quality 3-D graphics, crisp response, great power efficiency and adaptable. Future high-end mobile devices are likely to have multiple CPU cores, with a GPU, and the CPU also supports a SIMD instruction set. Therefore, 3-D graphics processing will be very versatile. We believe a hybrid CPU/GPU rendering method which combines the advantage of ray tracing and traditional rendering, can improve the quality of graphics and adapt to the actual conditions of the device in terms of power consumption, power supply, available network bandwidth, and the resolution of the display. In the hybrid method, the ray tracing part can take advantage of the homogeneous multi-cores, and the pre-computation of rasterization can be dispatched to GPU. Or for saving power, the pre-computation can be carried out by SIMD instructions. To support this graphics model, this project will provide the essential system software development tools including the parallelizing compilers (both homogeneous multi-core and heterogeneous multi-cores), a dynamic monitoring and optimization system, and a highly efficient system emulation environment. These tools connect the applications developed based on the hybrid graphics model and the power/performance monitoring mechanisms proposed in the target architecture.