Full metadata record
DC FieldValueLanguage
dc.contributor.author曹孝櫟en_US
dc.contributor.authorTsao Shiao-Lien_US
dc.date.accessioned2014-12-13T10:43:01Z-
dc.date.available2014-12-13T10:43:01Z-
dc.date.issued2011en_US
dc.identifier.govdocNSC100-2220-E009-038zh_TW
dc.identifier.urihttp://hdl.handle.net/11536/99543-
dc.identifier.urihttps://www.grb.gov.tw/search/planDetail?id=2313566&docId=361782en_US
dc.description.abstract本計畫第一年在於建構處理器高效率熱模型與熱評估工具之設計與開發,先進微處理器可以透過熱能管理技術大幅改進其效能與可靠性,而獲得微處理器內部各個位置上精確的溫度資訊則是熱能管理中重要的第一步。但是,在使用軟體模型獲取詳細的溫度資訊上,需要花費相當的運算時間以及記憶體空間才能獲得,以至於不是所有求得溫度的方法都可以適用於動態溫度管理。在本計畫中,我們提出一個新的溫度評估模型,有效地獲得微處理器詳細的溫度分佈。其實驗結果與知名溫度模擬器HotSpot只有0.3至1.5%的溫度差異,但有著34至47倍的運算速度提升,且僅消耗其0.45%的記憶體使用率。zh_TW
dc.description.abstractTo obtain detail thermal distribution of a microprocessor is one of critical tasks for thermal management which improves the reliability, performance, etc. of modern microprocessors significantly. However, thermal simulator requires considerable computation time and memory space to provide fine-grained temperature information, and hence may not be suitable for dynamic thermal management. In this project, we propose a novel model to efficiently derive detailed temperature information of a microprocessor. The thermal simulation results produced by the proposed model have merely 0.3% to 1.4% differences between that generated by HotSpot [28] with 34 to 47 times speedup in computation and only 0.45% of the memory usage.en_US
dc.description.sponsorship行政院國家科學委員會zh_TW
dc.language.isozh_TWen_US
dc.subject熱評估zh_TW
dc.subject動態散熱管理zh_TW
dc.subject熱模型zh_TW
dc.subjectThermal estimationen_US
dc.subjectdynamic thermal management (DTM)en_US
dc.subjectthermal model.en_US
dc.titleGreenArmy:綠色微雲伺服系統晶片平台技術-子計畫四:基於3D堆疊架構綠色微雲伺服系統晶片之電源與散熱管理研究( I )zh_TW
dc.titlePower and Thermal Management for 3d-Stacked Cloudlet Server on Chipen_US
dc.typePlanen_US
dc.contributor.department國立交通大學資訊工程學系(所)zh_TW
Appears in Collections:Research Plans