Analysis of Intel's Haswell Microarchitecture Using The ECM Model and Microbenchmarks

TL;DR

This paper provides a detailed analysis of Intel's Haswell microarchitecture, using the ECM model and microbenchmarks to quantify its efficiency and improvements in features like the Uncore design, memory hierarchy, and execution units.

Contribution

It introduces a comprehensive evaluation of Haswell's microarchitecture features using the ECM model and a standardized set of microbenchmarks for streaming loop kernels.

Findings

Quantified the impact of dual-ring Uncore design

Assessed the effects of Cluster-on-Die mode and Uncore Frequency Scaling

Provided a blueprint for analyzing streaming loop kernels

Abstract

This paper presents an in-depth analysis of Intel's Haswell microarchitecture for streaming loop kernels. Among the new features examined is the dual-ring Uncore design, Cluster-on-Die mode, Uncore Frequency Scaling, core improvements as new and improved execution units, as well as improvements throughout the memory hierarchy. The Execution-Cache-Memory diagnostic performance model is used together with a generic set of microbenchmarks to quantify the efficiency of the microarchitecture. The set of microbenchmarks is chosen such that it can serve as a blueprint for other streaming loop kernels.