000169993 001__ 169993
000169993 005__ 20190509132414.0
000169993 0247_ $$2doi$$a10.5075/epfl-thesis-5253
000169993 02470 $$2urn$$aurn:nbn:ch:bel-epfl-thesis5253-2
000169993 02471 $$2nebis$$a6672472
000169993 037__ $$aTHESIS
000169993 041__ $$aeng
000169993 088__ $$a5253
000169993 245__ $$aDesign of Thermal Management Control Policies for Multiprocessors Systems on Chip
000169993 260__ $$bEPFL$$c2011$$aLausanne
000169993 269__ $$a2011
000169993 300__ $$a173
000169993 336__ $$aTheses
000169993 520__ $$aThe contribution of this thesis is a thorough study of  thermal aware policy design for MPSoCs. The study includes  the modelling of their thermal behavior as well as the  improvement and the definition of new thermal management and  balancing policies. The work is structured on three main specific disciplines.  The areas of contributions are: modeling, algorithms and  system design. This thesis extends the field of modeling by proposing new  techniques to represent the thermal behavior of MPSoCs using  a mathematical formalization. Heat transfer and modelling of  physical properties of MPSoCs have been extensively studied.  Special emphasis is given to the way the system cools down  (i.e. micro-cooling, natural heat dissipation etc.) and the  heat propagates inside the MPSoC. The second contribution of this work is related to  policies, which manage MPSoC working frequencies and  micro-cooling pumps to satisfy user requirements in the most  effective possible way, while consuming the lowest possible  amount of resources. Several families of thermal policies  algorithms have been studied and analyzed in this work for  both 2D and 3D MPSoCs including liquid cooling  technologies. The discipline of system design has also been extended  during the development of this thesis. Thermal management  policies have been implemented in real emulation platforms  and contributions in this area are related to the design and  implementation of proposed innovations in real MPSoC  platforms.
000169993 6531_ $$aThermal
000169993 6531_ $$aManagement
000169993 6531_ $$aModel
000169993 6531_ $$aPredictive control
000169993 6531_ $$aTemperature
000169993 6531_ $$aAlgorithm
000169993 6531_ $$a3D-MPSoC
000169993 6531_ $$aOptimization
000169993 6531_ $$aAccuracy
000169993 700__ $$0242412$$g178989$$aZanini, Francesco
000169993 720_2 $$aDe Micheli, Giovanni$$edir.$$g167918$$0240269
000169993 8564_ $$uhttps://infoscience.epfl.ch/record/169993/files/EPFL_TH5253.pdf$$zTexte intégral / Full text$$s3803360$$yTexte intégral / Full text
000169993 909C0 $$xU11140$$0252283$$pLSI1
000169993 909CO $$pthesis$$pthesis-bn2018$$pDOI$$pIC$$ooai:infoscience.tind.io:169993$$qDOI2$$qGLOBAL_SET$$pSTI
000169993 918__ $$dEDIC2005-2015$$cISIM$$aIC
000169993 919__ $$aLSI
000169993 920__ $$b2011
000169993 970__ $$a5253/THESES
000169993 973__ $$sPUBLISHED$$aEPFL
000169993 980__ $$aTHESIS