Advanced alterable pipeline timer (adapt): a tool to design a high performance PowerPC/sup TM/ microprocessor

Abstract

The microprocessor discussed in this paper is a member of the G3 family of PowerPC processors, the third generation of PowerPC microprocessor products. It provides the performance levels required for high end desktop systems while offering the low typical power dissipation and small die size that make it very attractive for portable systems. It is an advanced superscalar design with six execution units, aggressive upstream branch processing, out-oforder instruction execution, and a tightly integrated "backside" L2 cache. Most notably it achieves workstation/ server class performance while only dissipating 5 watts. A major portion of the design effort involved architectural performance modeling, making cost/power/performance trade-offs, and verifying performance of the implementation.

Building a design support tool for superscalar processors and its case studies

“…According to Ref. 19, errors compared with the real machine are within 10%, and simulation speed is 25,000 cycles/s, which is rather fast. Though the instruction set is fixed by Paratool, configuration of components (such as operation units and cache) and latency are variable.…”

“…But their details are rarely made public. Here, we will explain VMW [16] as a system that seeks generality, ADAPT [19] as a system dedicated to a specific processor architecture, and Mikage [20] as a system targeted mainly for educational use.…”

A G3 PowerPC/sup TM/ superscalar low-power microprocessor

“…The simulator is referred to as an "ADvanced Alterable Pipeline Timer" (or "ADAPT") [4] and is what guided many of the design decisions and trade-offs. ADAPT is designed in such a way that various parameters (such as the number of instructions fetched per cycle, cache sizes, number of renames, etc.)…”

Techniques for implementing fast processor simulators

“…For instance, the next fastest commercial processor simulator for which simulator performance numbers have been reported that we are aware of is ADAPT [2]. ADAPT has less than half the performance of Turandot, though it is modeling a simpler processor.…”