天堂国产午夜亚洲专区-少妇人妻综合久久蜜臀-国产成人户外露出视频在线-国产91传媒一区二区三区

當(dāng)前位置:主頁 > 科技論文 > 計(jì)算機(jī)論文 >

嵌入式處理器取指單元關(guān)鍵部件低功耗技術(shù)研究

發(fā)布時(shí)間:2018-02-05 03:06

  本文關(guān)鍵詞: 取指單元 低功耗 指令高速緩存 分支目標(biāo)緩存 路預(yù)測 循環(huán)過濾 分支跟蹤 出處:《浙江大學(xué)》2012年碩士論文 論文類型:學(xué)位論文


【摘要】:低功耗是嵌入式處理器的主要研究方向,而取指單元包含大量存儲器訪問,其功耗已經(jīng)成為整個(gè)處理器功耗的重要組成部分。本文重點(diǎn)研究了嵌入式處理器取指單元關(guān)鍵部件的低功耗技術(shù)。這些創(chuàng)新技術(shù)的應(yīng)用能夠使嵌入式CPU在保持處理器性能的基礎(chǔ)上,有效降低取指單元動態(tài)功耗。本文的主要內(nèi)容及創(chuàng)新點(diǎn)包括: 1、基于指令歷史鏈接關(guān)系的路預(yù)測高速緩存低功耗技術(shù)。本文提出了一種基于分支目標(biāo)緩存(BTB)復(fù)用和跳轉(zhuǎn)歷史鏈接關(guān)系表的兩級跳轉(zhuǎn)路預(yù)測架構(gòu),以極低的硬件成本解決了分支目標(biāo)指令所在行的沖突問題,提高了跳轉(zhuǎn)預(yù)取路預(yù)測的準(zhǔn)確率,進(jìn)而有效降低了指令Cache的動態(tài)功耗。 2、基于循環(huán)體訪問過濾的分支目標(biāo)緩存低功耗技術(shù)。針對循環(huán)體內(nèi)指令對BTB訪問產(chǎn)生的冗余功耗問題,本文提出了一種循環(huán)體訪問過濾機(jī)制,消除循環(huán)體指令流中順序指令對BTB的無效訪問;進(jìn)一步提出了一種分支跟蹤方法補(bǔ)償由于循環(huán)過濾機(jī)制對循環(huán)體中非循環(huán)類分支指令的錯(cuò)誤過濾造成的性能損失。這兩種技術(shù)在保障處理器性能的基礎(chǔ)上降低了BTB的動態(tài)功耗。 本文提出的低功耗技術(shù)設(shè)計(jì)簡單、硬件成本低,對嵌入式處理器取指單元低功耗設(shè)計(jì)具有積極意義。
[Abstract]:Low power consumption is the main research direction of embedded processor, and the fetch cell contains a large number of memory access. The power consumption of embedded processor has become an important part of the whole processor power. This paper focuses on the low power technology of the key components of the embedded processor. The application of these innovative technologies can enable embedded CPU to be guaranteed in the embedded system. Based on processor performance. The main contents and innovations of this paper are as follows: 1. Low power technology of path prediction cache based on instruction history link relationship. In this paper, a two-level hopping path prediction architecture based on branch target buffer (BTB) multiplexing and hopping history link relation table is proposed. The collision problem of the line of branch target instruction is solved with very low hardware cost, and the accuracy of jump prefetching prediction is improved, and the dynamic power consumption of instruction Cache is reduced effectively. 2. Low power consumption technology of branch target buffer based on cyclic body access filtering. Aiming at the redundant power consumption of BTB access caused by loop body instruction, this paper proposes a cyclic volume access filtering mechanism. Eliminating invalid access to BTB by sequential instructions in the flow of circular body instructions; Furthermore, a branch tracking method is proposed to compensate for the performance loss caused by the error filtering of the loop filter mechanism on the non-cyclic branch instructions of the loop body. These two techniques can reduce the processor performance on the basis of ensuring the processor performance. The dynamic power consumption of BTB. The low power technology proposed in this paper has the advantages of simple design and low hardware cost, which is of great significance to the low power design of the embedded processor.
【學(xué)位授予單位】:浙江大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2012
【分類號】:TP332

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 喻明艷;張祥建;楊兵;;基于跳躍訪問控制的低功耗分支目標(biāo)緩沖器設(shè)計(jì)[J];計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào);2010年04期

2 徐懿;李麗;高明倫;黃壯雄;楊盛光;;納米集成電路靜態(tài)功耗機(jī)理及低功耗設(shè)計(jì)技術(shù)[J];微電子學(xué)與計(jì)算機(jī);2007年05期

3 楊名;于立新;;基于標(biāo)志壓縮的低功耗指令cache設(shè)計(jì)[J];微電子學(xué)與計(jì)算機(jī);2008年05期

,

本文編號:1491940

資料下載
論文發(fā)表

本文鏈接:http://sikaile.net/kejilunwen/jisuanjikexuelunwen/1491940.html


Copyright(c)文論論文網(wǎng)All Rights Reserved | 網(wǎng)站地圖 |

版權(quán)申明:資料由用戶df771***提供,本站僅收錄摘要或目錄,作者需要刪除請E-mail郵箱bigeng88@qq.com