Improved Clustered Voltage Scaling for Low Power Cell-Based Design
碩士 === 國立交通大學 === 電子工程系所 === 93 === As the semiconductor technologies make progress by scaling-down the feature size,integrated circuits can operate at higher frequencies and achieve higher performance. However, increasing operating frequencies means deteriorating power dissipation problems; moreove...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2005
|
Online Access: | http://ndltd.ncl.edu.tw/handle/81648341779259645339 |
id |
ndltd-TW-093NCTU5428098 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-093NCTU54280982016-06-06T04:10:45Z http://ndltd.ncl.edu.tw/handle/81648341779259645339 Improved Clustered Voltage Scaling for Low Power Cell-Based Design 給低功率元件式設計使用的改良型叢聚式電壓調降法 Shwang-Yi Tan 譚雙議 碩士 國立交通大學 電子工程系所 93 As the semiconductor technologies make progress by scaling-down the feature size,integrated circuits can operate at higher frequencies and achieve higher performance. However, increasing operating frequencies means deteriorating power dissipation problems; moreover, scaling-down causes larger leakage current. Power consumption problems increase the design difficulty for battery powered applications, and also affect ordinary designs in terms of time to market, cost, and reliability. CVS (Clustered Voltage Scaling) is an effective way to reduce IC power consumption. CVS utilizes the excess time slacks inside circuits and trade them for power reduction. Methods based on CVS for saving power have been studied for years. We propose an improved CVS method, Bilateral CVS (BCVS). BCVS is a general Clustered Voltage Scaling method which subsumes both CVS and ECVS. In this thesis, we also discuss why our improvements work by experimental results. Hung-Ming Chen 陳宏明 2005 學位論文 ; thesis 27 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立交通大學 === 電子工程系所 === 93 === As the semiconductor technologies make progress by scaling-down the feature size,integrated circuits can operate at higher frequencies and achieve higher performance. However, increasing operating frequencies means deteriorating power dissipation problems; moreover, scaling-down causes larger leakage current. Power consumption problems increase the design difficulty for battery powered applications, and also affect ordinary designs in terms of time to market, cost, and reliability.
CVS (Clustered Voltage Scaling) is an effective way to reduce IC power consumption. CVS utilizes the excess time slacks inside circuits and trade them for power reduction. Methods based on CVS for saving power have been studied for years. We propose an improved CVS method, Bilateral CVS (BCVS). BCVS is a general Clustered Voltage Scaling method which subsumes both CVS and ECVS. In this thesis, we also discuss why our improvements work by experimental results.
|
author2 |
Hung-Ming Chen |
author_facet |
Hung-Ming Chen Shwang-Yi Tan 譚雙議 |
author |
Shwang-Yi Tan 譚雙議 |
spellingShingle |
Shwang-Yi Tan 譚雙議 Improved Clustered Voltage Scaling for Low Power Cell-Based Design |
author_sort |
Shwang-Yi Tan |
title |
Improved Clustered Voltage Scaling for Low Power Cell-Based Design |
title_short |
Improved Clustered Voltage Scaling for Low Power Cell-Based Design |
title_full |
Improved Clustered Voltage Scaling for Low Power Cell-Based Design |
title_fullStr |
Improved Clustered Voltage Scaling for Low Power Cell-Based Design |
title_full_unstemmed |
Improved Clustered Voltage Scaling for Low Power Cell-Based Design |
title_sort |
improved clustered voltage scaling for low power cell-based design |
publishDate |
2005 |
url |
http://ndltd.ncl.edu.tw/handle/81648341779259645339 |
work_keys_str_mv |
AT shwangyitan improvedclusteredvoltagescalingforlowpowercellbaseddesign AT tánshuāngyì improvedclusteredvoltagescalingforlowpowercellbaseddesign AT shwangyitan gěidīgōnglǜyuánjiànshìshèjìshǐyòngdegǎiliángxíngcóngjùshìdiànyādiàojiàngfǎ AT tánshuāngyì gěidīgōnglǜyuánjiànshìshèjìshǐyòngdegǎiliángxíngcóngjùshìdiànyādiàojiàngfǎ |
_version_ |
1718294500547231744 |