程序執(zhí)行軌跡（Program executions traces，以下簡(jiǎn)稱traces）是程序執(zhí)行過(guò)程的指令流信息的記錄，trace完整地記錄了程序執(zhí)行過(guò)程中所執(zhí)行指令的內(nèi)容和順序。對(duì)于大多數(shù)程序，少數(shù)幾個(gè)較短的熱t(yī)race決定了系統(tǒng)的總體性能。本文提出了基于程序執(zhí)行軌跡提取加速模塊的軟硬件劃分方法。利用熱t(yī)race提取算法劃分系統(tǒng)中關(guān)鍵的trace到硬件，使用分支斷言構(gòu)造原子執(zhí)行單位，以較小的硬件代價(jià)獲得較高的加速比。在本文實(shí)驗(yàn)中，與采用模擬退火算法的指令級(jí)細(xì)粒度劃分相比，獲得的性能平均高9.6%，最終結(jié)果硬件面
積小29%。
關(guān)鍵詞: 程序執(zhí)行軌跡，硬件加速模塊，SoC，軟硬件劃分
A Method to Construct SoC Hardware Acceleration Module with Trace
Sang Sheng-tian1, Yu ming-yan1, Ye yi-zheng1
(1. Microelectronics Center of Harbin Institute of Technology, Harbin 150001, China)
Abstract: Traces are the records of the program executions, and for most programs a few of short hot traces affect
the overall performance. In this paper, we propose a hardware/software partitioning method to extract the
hardware acceleration modules based on traces. With the algorithm presented, hot traces are identified and
transformed to atomic execution units using branch assertions to get high speedup at low cost. Experiment result
shows that compared with annealing partitioning in granularity of instruction, the trace based method produces
higher performance by average 9.6% and the hardware area is reduced by 29%.
Key words: Trace, Hardware Acceleration Module, System on a Chip, Hardware/Software Partitioning