High-Efficiency Transient Temperature Calculations for Applications in Dynamic Thermal Management of Electronic Devices

Abstract

The highly nonuniform transient power densities in modern semiconductor devices present difficult performance and reliability challenges for circuit components, multiple levels of interconnections and packaging, and adversely impact overall power efficiencies. Runtime temperature calculations would be beneficial to architectures with dynamic thermal management, which control hotspots by effectively optimizing regional power densities. Unfortunately, existing algorithms remain computationally prohibitive for integration within such systems. This work addresses these shortcomings by formulating an efficient method for fast calculations of temperature response in semiconductor devices under a time-dependent dissipation power. A device temperature is represented as output of an infinite-impulse response (IIR) multistage digital filter, processing a stream of sampled power data; this method effectively calculates temperatures by a fast numerical convolution of the sampled power with the modeled system's impulse response. Parameters such as a steady-state thermal resistance or its extension to a transient regime, a thermal transfer function, are typically used with the assumption of a linearity and time-invariance (LTI) to form a basis for device thermal characterization. These modeling tools and the time-discretized estimates of dissipated power make digital filtering a well-suited technique for a run-time temperature calculation. A recursive property of the proposed algorithm allows a highly efficient use of an available computational resource; also, the impact of all of the input power trace is retained when calculating a temperature trace. A network identification by deconvolution (NID) method is used to extract a time-constant spectrum of the device temperature response. We verify this network extraction procedure for a simple geometry with a closed-form solution. In the proposed technique, the amount of microprocessor clock cycles needed for each temperature evaluation remains fixed, which results in a linear relationship between the overall computation time and the number of temperature evaluations. This is in contrast to time-domain convolution, where the number of clock cycles needed for each evaluation increases as the time window expands. The linear dependence is similar to techniques based on FFT algorithms; in this work, however, use of z-transforms significantly decreases the amount of computations needed per temperature evaluation, in addition to much reduced memory requirements. Together, these two features result in vast improvements in computational throughput and allow implementations of sophisticated runtime dynamic thermal management algorithms for all high-power architectures and expand the application range to embedded platforms for use in a pervasive computing environment.

Evaluating Broader Impacts of Nanoscale Thermal Transport Research

“…Infrared thermography and thermoreflectance thermal imaging [164,165] provide noncontact, micrometer-to submicrometer-resolution thermal images of operating devices, and atomic force microscopy (AFM)-based thermal mapping [7,166] yields even higher resolution thermal maps. Compact thermal simulation methods [167][168][169][170][171][172][173] allow rapid design optimization and provide a framework for implementing dynamic thermal management strategies. In modern transistors, subcontinuum effects and complex electron-phonon interactions impact the temperature profiles.…”

Limits to Fourier theory in high thermal conductivity single crystals

“…Currently, designers typically use Fourier theory to evaluate the benefits of high thermal conductivity heat spreaders. 8,11 However, the accuracy of Fourier theory for describing thermal transport in SiC or diamond when heat sources have submicron dimensions is rarely considered.…”

Compact thermal modeling methodology for predicting skin temperature of passively cooled devices

“…The main contribution of Touzelbaev et al [8] and Yang et al [9e11], was a time discretization formulation of the above defined using recursive Infinite Impulse Response (IIR) multistage digital filter, processing a stream of sampled power data; this method effectively calculates temperatures by a fast numerical convolution of the sampled power with the modeled system's impulse response. Network identification by de-convolution (NID) method was used to extract a time constant spectrum of the device temperature response.…”

“…The challenge then becomes dealing with efficiency and accuracy of the power management algorithm which, in effect, is directly dependent on the precision of the thermal model used for predicting temperature. Touzelbaev et al [8] and Yang et al [9e11] used Foster thermal RC-networks for compact modeling of transient heat conduction in packages. Their goal was to answer fundamental reliability questions in package production operations.…”