Architectural characterization of XQuery workloads on modern processors

Abstract

As XQuery rapidly emerges as the standard for querying XML documents, it is very important to understand the architectural characteristics and behaviors of such workloads. A lot of efforts are focused on the implementation, optimization, and evaluation of XQuery tools. However, few or no prior work studies the architectural and memory system behaviors of XQuery workloads on modern hardware platforms. This makes it unclear whether modern CPU techniques, such as the multi-level caches and hardware branch predictors, can support such workloads well enough. This paper presents a detailed characterization of the architectural behavior of XQuery workloads. We examine four XQuery tools on three hardware platforms (AMD, Intel, and Sun) using well-designed XQuery queries. We report measured architectural data, including the L1/L2 cache misses, TLB misses, and branch mispredictions. We believe that the information will be useful in understanding XQuery workloads and analyzing the potential architectural optimization opportunities of improving XQuery performance.