Implementation Independent Architectural ComparisonReport
Users and designers of CPU architectures lack a solid means of comparing architectures, both extant and proposed. The problem is that running benchmarks, the primary method used, measures both the architecture and its implementation (CPU and enclosing system). This report describes a method we have developed to compare architectures in an implementation independent . manner. Preliminary results suggest that if the same implementation techniques are used architectures fall into performance classes with little difference between members of a class. We identify those features that distinguish the classes. The benefit of adding features to an architecture can now be analyzed: features that do not move the architecture to a better class are a waste of resources and can be avoided.
Note: Abstract extracted from PDF file via OCR
All rights reserved (no additional license for public reuse)
Grimshaw, Andrew, Dave Chisholm, Brad Segal, Ricky Benitez, Max Salinas, Peter Kester, and Stephen McCalla. "Implementation Independent Architectural Comparison." University of Virginia Dept. of Computer Science Tech Report (1990).
University of Virginia, Department of Computer Science