3-D integration is a promising technology to sustain transistor density scaling in the future, as well as facilitating new architectural designs that were not possible with traditional integration techniques. However, 3-D integration comes with some serious challenges, chief among them heat removal. A promising technology for thermal issues is microfluidic (MF) cooling. In this paper, we perform a design space analysis study on 3-D CPUs. We show that aggressive cooling solutions such as MF cooling are necessary to unlock the true potential of 3-D ICs. Without such cooling the thermal feasibility region of the design space is significantly reduced. We observe that interactions between thermal, electrical, and physical aspects of 3-D CPUs with MF cooling are substantial, and must be cooptimized during our analysis to correctly identify optimal design points. We simulate a spectrum of 3-D CPU architectures which offer vast improvements to performance, but are energy inefficient and thermally infeasible with air cooling. Furthermore, we show a 2.30× (1.59×) improvement in performance (energy efficiency) when MF cooling and floorplan cooptimization are added to our design space analysis simulation flow.

The inexorable rise in chip power dissipation and emergence of on-chip hot spots with heat fluxes approaching 1 =kW/cm/sup 2/ has turned renewed attention to direct cooling with dielectric liquids. Use of dielectric liquids in intimate contact with the heat dissipating surfaces eliminates the deleterious effects of solid-solid interface resistances and harnesses the highly efficient phase-change processes to the critical thermal management of advanced IC chips. In the interest of defining the state-of-the-art in direct liquid cooling, this paper begins with a discussion of the thermophysics of phase-change processes and a description of the available dielectric liquid cooling techniques and their history. It then describes the phenomenology of pool boiling, spray/jet impingement, gas-assisted evaporation, and synthetic jet impingement with dielectric liquids. Available correlations for predicting the heat transfer coefficients and limiting heat transfer rates, as well as documented empirical results for these promising techniques for on-chip hot spot cooling, are also provided and compared.

The simplicity and convenience of direct cooling with freely circulating air, as well as the ubiquity of such systems in the marketplace, make this the technique of choice for many thermal packaging tasks. While standard heat transfer correlations are generally unsuitable for determining the detailed thermal behavior of individual components, the maximum surface temperature of integrated circuit packages mounted on PCBs can often be shown to lie between two analytical bounds: an upper, "asymmetric isoflux," bound based on heat transfer from just one side of the PCB and a lower, "symmetric isoflux," bound based on identical thermal transport from both sides of each PCB. Following a brief review of the development and validation of the theoretical symmetric and asymmetric isoflux relations for smooth-plate channels, this study focuses on the application of the proposed bounding relations to widely and closely spaced component-carrying PCBs. Successful comparison with much of the available data demonstrates the acceptable accuracy of these analytical expressions, for all except the narrowest spacings, and their utility to the thermal designer of air-cooled PCBs.