Management Of Microprocessors Research Articles

Development of Analytical Model to a Temperature Distribution of a First Level Package With a Nonuniformly Powered Die

Microprocessors continue to grow in capabilities, complexity, and performance. Microprocessors typically integrate functional components such as logic and level two cache memory in their architecture. This functional integration of logic and memory results in improved performance of the microprocessor. However, the integration also introduces a layer of complexity in the thermal design and management of microprocessors. As a direct result of functional integration, the power map on a microprocessor is typically highly nonuniform, and the assumption of a uniform heat flux across the die surface has been shown to be invalid post Pentium II architecture. The active side of the die is divided into several functional blocks with distinct power assigned to each functional block. Previous work (Kaisare et al., 2005, “Thermal Based Optimization of Functional Block Distributions in a Non-Uniformly Powered Die,” InterPACK 2005, San Francisco, CA, Jul. 17–22) has been done, which includes numerical analysis and thermal based optimization of a typical package consisting of a nonuniformly powered die, heat spreader, thermal interface materials I and II, and the base of the heat sink. In this paper, an analytical approach to temperature distribution of a first level package with a nonuniformly powered die is carried out for the first time. The analytical model for two-layer bodies developed by Haji-Sheikh et al. (2003, “Steady-State Heat Conduction in Multi-Layer Bodies,” Int. J. Heat Mass Transfer, 46(13), pp. 2363–2379) is extended to this typical package, which is a multilayer body. The solution is to begin by designating each surface heat flux as a volumetric heat source. An inverse methodology is applied to solve the equations for various surfaces to calculate the maximum junction temperature for a given multilayer body. Finally validation of the analytical solution is carried out using previously developed numerical model.

The Management of Microprocessors

1. Preface There are by now no sentences left with which to make a dramatic introduction to a paper on microprocessors. The very best sentences have long since become part of common parlance, and most of those were never true anyway. The real facts about microprocessors are increasingly mundane and certainly socially depressing: as things stand right now it is quite uncertain that Britain will be a richer, more contented land by the year 2,000. There is no point in pretending that everything is fine. Most things are fairly bad: the rate at which British industry is making any attempt to come to terms with the microprocessor; the rate at which microprocessor engineers are being produced; the relatively low status afforded to anything practical; the evidence to date of union attitudes towards technology‐originated redundancies; all of these are cause for concern. We are in the middle of something qualitatively different from the last 50 years of industrial change, and the social implications are more than most will admit to. Employment will drop. To believe that workers replaced by microprocessor systems will inevitably find jobs in the construction of those same microprocessor systems is as foolish as thinking that horses replaced by motor cars could find jobs in the various branches of the auto industry.