Designing high-performance microprocessors in 3-dimensional integration technology.

摘要

In this dissertation, we demonstrate the impact of a new emerging technology called 3D-integration technology on microprocessors. 3D-integration technology stacks transistors in the vertical dimension in addition to the conventional horizontal plane. The additional degree of connectivity in the vertical dimension enables circuit designers to replace global wires with short vertical interconnects, thus reducing delay, power consumption, and area. To adapt planar microarchitectures to 3D-integrated designs, we study several building blocks that together comprise a substantial portion of a processor's total transistor count. In particular, we focus our attention on three basic circuit classes: static random access memory circuits, associative logic circuits, and data processing circuits. We propose different 3D-integrated circuit designs to deal with the constraints of the conventional planar technology. Based on the 3D-integrated circuits, we propose high-performance 3D-integrated microprocessors and evaluate the impact on performance, power, and temperature. We demonstrate two different approaches to improve performance: clock speed improvement and IPC improvement. We demonstrate the simultaneous benefits of the 3D-integration and highlight the power density issue related to the 3D-integration technology. Next, we propose novel microarchitectural techniques to address the challenge of power density. We demonstrate that microarchitecture-level techniques can effectively control the power density in the 3D-integrated processors. The 3D-integrated processors provide a significant performance benefit over the planar processors while simultaneously reducing the total power. One of the key contributions of this dissertation is the temperature analysis that shows that the worst-case temperatures on the 3D-integrated processors can be effectively controlled using microarchitecture level techniques. The 3D-integration technology may extend the applicability of Moore's law for a few more technology generations.