| T. Simunic et al., "Event-driven power management," IEEE Tran. Computer-Aided Design of Integrated Circuits and Systems, vol. 20, pp. 840-857, July 2001. |
No context found.
T. Simunic, L. Benini, P. Glynn, G. De Micheli, "Event-driven Power Management," IEEE Transactions on CAD, July 2001.
No context found.
T. Simunic, L. Benini and G. De Micheli, "Event-driven power management", International Symposium on System Synthesis, pp. 18--23, 1999.
No context found.
T. Simunic, L. Benini, and G. De Micheli, "Event-driven power management, " in Proc. Int. Symp. System Synthesis, Apr. 1999, pp. 18--23.
....All state transitions are assumed to follow stationary geometric distribution. The decision evaluation is repeated periodically, even when the system is idle, thus wasting power. Extensions to the DTMDP model are event driven continuous time (CTMDP) and semi Markov (SMDP) decision process models [3, 4]. Both CTMDP and SMDP models assume exponential service request arrival times, and thus can have high energy costs and a large performance penalty. The power manager makes one decision as soon as the system is idle. If the decision is to stay awake, the system will wait until another arrival ....
T. Simunic, L. Benini, and G. D. Micheli. Event-driven power management. In International Symposium on System Synthesis, pages 18--23, 1999. 1
....approaches based on stochastic models can guarantee optimal results. Stochastic models to date have been formulated with open loop control model, where statistics of the system are collected and characterized ahead of time, and the control is derived based on those with no adaptation at run time[14, 16,17]. An exception is the adaptive approach presented in [15] that uses only memoryless distributions to describe the history dependent system behavior. In addition to transitioning components into low power states during idle times, power manager can also adjust processing frequency and voltage in ....
....of NOCs can be used as the basis for optimization of power consumption under QoS constraints. Figure 1 shows a sample NOC. The NOC can be modeled as a queuing network with a number of service points representing cores. Each core can be modeled using Renewal model much like the one presented in [17] for portable devices. Renewal theory studies stochastic systems that have a state called renewal state, in which the process statistically begins anew. The time between successive visits to renewal state is called renewal time, and one cycle from renewal state, through other states and then back ....
[Article contains additional citation context not shown here]
T. Simunic, L. Benini, P. Glynn, G. De Micheli, "Event-driven Power Management," IEEE Transactions on CAD, pp.840-857, July 2001.
....one decision as soon as the system goes idle. If the decision is to stay awake, the system will wait until another arrival before revising the decision, possibly missing large idle times, such as a lunch break. CTMDP model was further generalized with semi Markov decision process model (SMDP) [12]. SMDP model can treat a general distribution occuring at the same time with an exponential distribution. In [12] the hard disk transition to and from the sleep state was modeled using uniform distribution, while the exponential distribution still represented the user request arrivals. Some ....
....arrival before revising the decision, possibly missing large idle times, such as a lunch break. CTMDP model was further generalized with semi Markov decision process model (SMDP) 12] SMDP model can treat a general distribution occuring at the same time with an exponential distribution. In [12] the hard disk transition to and from the sleep state was modeled using uniform distribution, while the exponential distribution still represented the user request arrivals. Some measurements constrasting CTMDP and SMDP are shown in [13] Although exponential distribution model can be used to ....
T. Simunic, L. Benini and G. De Micheli, "Event-driven power management", Proceedings of ISSS, pp. 18-23, 1999.
....between off and active states are best fit with uniform distribution. Our model has two non exponential transitions occurring at the same time when the card transitions from doze mode into off state. Thus we could not apply policy optimization algorithms based on exponential models, such as [9, 10, 13]. Large errors result if exponential distribution is used for all transitions, as was shown in [11] Another approach to handle non exponential transitions is to use adaptive method, such as in [12] This method requres policy interpolation at very short time increments, regardless of the device ....
T. Simunic, L. Benini and G. De Micheli, "Event-driven power management", Proceedings of ISSS, 1999.
No context found.
T. Simunic et al., "Event-driven power management," IEEE Tran. Computer-Aided Design of Integrated Circuits and Systems, vol. 20, pp. 840-857, July 2001.
No context found.
Simunic T, Benini L, Glynn P, De Micheli G. "Event-driven power management," Computer-Aided Design of Integrated Circuits and Systems, IEEE Transactions on, Vol. 20, pp. 840857, Jul. 2001.
No context found.
T. Simunic et al., "Event-driven power management," IEEE Tran. Computer-Aided Design of Integrated Circuits and Systems, vol. 20, pp. 840-857, July 2001.
No context found.
T. Simunic et al, "Event-driven power management," IEEE Tran. Computer-Aided Design of Integrated Circuits and Systems, vol. 20, pp. 840-857, July 2001.
No context found.
T. Simunic et al, "Event-driven power management", IEEE Trans. CAD, vol. 20, pp. 840-857, July 2001.
No context found.
T. Simunic, L. Benini, P. Glynn, and G. De Micheli, "Event driven power management," IEEE Trans. Computer-Aided Design, vol. 20, pp. 840--857, July 2001.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC