Quantifying Load Stream Behavior

The increasing performance gap between processors and memory will force future architectures to devote significant resources towards removing and hiding memory latency. The two major architectural features used to address this grow-ing gap are caches and prefetching. In this paper we perform a detailed quantification of the cache miss patterns for the Olden benchmarks, SPEC 2000 benchmarks, and a collection of pointer based applications. We classify misses into one of four categories correspond-ing to the type of access pattern. These are next-line, stride, same-object (additional misses that occur to a recently ac-cessed object), or pointer-based transitions. We then pro-pose and evaluate a hardware profiling architecture to cor-rectly identify which type of access pattern is being seen. This access pattern identification could be used to help guide and allocate prefetching resources, and provide information to feedback-directed optimizations. A second goal of this paper is to identify a suite of chal-lenging pointer-based benchmarks that can be used to fo-cus the development of new software and hardware prefetch-ing algorithms, and identify the challenges in performing prefetching for these applications using new metrics.