Modeling Shared Cache Performance of OpenMP Programs using Reuse Distance

TL;DR

This paper introduces a scalable probabilistic model to predict shared cache performance of OpenMP programs by analyzing reuse distance profiles from sequential memory traces, enabling accurate performance estimation on multicore systems.

Contribution

It presents a novel static, probabilistic approach to model shared cache reuse distances for parallel applications, improving prediction efficiency and accuracy.

Findings

Accurate shared cache hit-rate predictions

Effective static analysis from sequential traces

Scalable performance modeling approach

Abstract

Performance modeling of parallel applications on multicore computers remains a challenge in computational co-design due to the complex design of multicore processors including private and shared memory hierarchies. We present a Scalable Analytical Shared Memory Model to predict the performance of parallel applications that runs on a multicore computer and shares the same level of cache in the hierarchy. This model uses a computationally efficient, probabilistic method to predict the reuse distance profiles, where reuse distance is a hardware architecture-independent measure of the patterns of virtual memory accesses. It relies on a stochastic, static basic block-level analysis of reuse profiles measured from the memory traces of applications ran sequentially on small instances rather than using a multi-threaded trace. The results indicate that the hit-rate predictions on the shared cache are accurate.