Inline Caching Meets Quickening

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Inline caches effectively eliminate the overhead implied by dynamic typing. Yet, inline caching is mostly used in code generated by just-in-time compilers. We present efficient implementation techniques for using inline caches without dynamic translation, thus enabling future interpreter implementers to use this important optimization technique—we report speedups of up to a factor of 1.71—without the additional implementation and maintenance costs incurred by using a just-in-time compiler.