Thermal-aware management techniques for cyber-physical systems

Plant-Cyber Tradeoffs in Frame-Based Cyber-Physical Systems

“…where α(t) is the iTAAF at time t, E a the activation energy, k the Boltzmann constant, τ (t) the absolute temperature of the device at time t and τ ref is the reference temperature, with respect to which the acceleration factor is calculated [10]. TAAF is the factor by which the expected device lifetime is reduced due to thermal stress over a given interval, i.e., the effective aging of a processor over some interval, I is I α(t)dt.…”

“…TAAF is the factor by which the expected device lifetime is reduced due to thermal stress over a given interval, i.e., the effective aging of a processor over some interval, I is I α(t)dt. Just a small increase in temperature can significantly shorten expected device lifetime [10].…”

Plant-Cyber Tradeoffs in Frame-Based Cyber-Physical Systems

“…where α(t) is the iTAAF at time t, E a the activation energy, k the Boltzmann constant, τ (t) the absolute temperature of the device at time t and τ ref is the reference temperature, with respect to which the acceleration factor is calculated [10]. TAAF is the factor by which the expected device lifetime is reduced due to thermal stress over a given interval, i.e., the effective aging of a processor over some interval, I is I α(t)dt.…”

“…TAAF is the factor by which the expected device lifetime is reduced due to thermal stress over a given interval, i.e., the effective aging of a processor over some interval, I is I α(t)dt. Just a small increase in temperature can significantly shorten expected device lifetime [10].…”

The Thermal-Constrained Real-Time Systems Design on Multi-Core Platforms -- An Analytical Approach

“…The new liquid-based cooling method attracts researchers' and industrials' attention; however, the different thermal characteristics and controllability of liquid coolant raise new challenge in 3D processor design, e.g. the coolant temperature/heat removal capacity is quite different near the microchannel inlet and outlet, which exaggerates the thermal/performance imbalance across the chip [58,75,147,129,8,108,105,159,77,120,142,3,154,67,110,34,1] One of my research interests is to build a more aggressive and finer granularity cooling infrastructure, that can be incorporated into the task allocation strategy, such that the cooling itself is a dynamic and a smart self-adjustable mechanism. The research can be conducted from two directions: (1) hardware innovation: designing a non-uniformed microchannel with different pipe widths/densities/topologies to mitigate the 3D thermal gradient in nature; (2) hardware and software co-schedule:…”