Informing memory operations: providing memory performance feedback in modern processors

A new class of memory operations called informing memory operations is proposed, which essentially consist of a memory operation combined either implicitly or explicitly with a conditional branch-and-link operation that is taken only if the reference suffers a cache miss. Two different implementations of informing memory operations are described, one based on cache-outcome condition code and another based on low-overhead traps; it is found that modern in-order-issue and out-of-order issue superscalar processors already contain the bulk of the necessary hardware support. This paper describes how a number of software-based memory optimizations can exploit informing memory operations to enhance performance, and takes a look at cache coherence with fine-grained access control as a case study.