Resources

Potential Scenario

2015-Beckmann-Sanchez-Cache Calculus-Modeling Caches through Differential Equations

Author(s): Nathan Beckmann

NA

Keywords: discrete cache cache calculus

110 total view(s), 70 download(s)

Abstract

Resource Image We present cache calculus, a technique that models cache behavior as a system of ordinary differential equations, letting standard calculus techniques find simple and accurate solutions of cache performance for common access patterns.

Citation

Researchers should cite this work as follows:

Article Context

Resource Type
Differential Equation Type
Technique
Qualitative Analysis
Application Area
Course
Course Level
Lesson Length
Technology
Approach
Skills

Description

Beckmann, Nathan and  Daniel Sanchez.  2015. Cache Calculus: Modeling Caches through Differential Equations. Computer Science and Artifical Intelligence Laboratory, MIT.

See https://people.csail.mit.edu/sanchez/papers/2016.cachecalc.cal.pdf.  Access 28 March 2023.

Abstract: —Caches are critical to performance, yet their behavior is hard to understand and model. In particular, prior work does not provide closed-form solutions of cache performance, i.e. simple expressions for the miss rate of a specific access pattern. Existing cache models instead use numerical methods that, unlike closed-form solutions, are computationally expensive and yield limited insight. We present cache calculus, a technique that models cache behavior as a system of ordinary differential equations, letting standard calculus techniques find simple and accurate solutions of cache performance for common access patterns.

Keywords: cache, differential equation, model, discrete, difference equation

 

Article Files

Authors

Author(s): Nathan Beckmann

NA

Comments

Comments

There are no comments on this resource.