MILC staggered conjugate gradient performance on Intel KNL

TL;DR

This paper discusses optimizing the staggered conjugate gradient algorithm in the MILC code for Intel KNL architecture, focusing on performance improvements for scientific computing applications.

Contribution

It presents an optimized implementation of the staggered CG algorithm tailored for Intel KNL, comparing MPI+OpenMP and QPhiX-based versions for enhanced performance.

Findings

QPhiX solver improves performance over baseline

Multi-node runs show scalability benefits

Optimization reduces computation time significantly

Abstract

We review our work done to optimize the staggered conjugate gradient (CG) algorithm in the MILC code for use with the Intel Knights Landing (KNL) architecture. KNL is the second gener- ation Intel Xeon Phi processor. It is capable of massive thread parallelism, data parallelism, and high on-board memory bandwidth and is being adopted in supercomputing centers for scientific research. The CG solver consumes the majority of time in production running, so we have spent most of our effort on it. We compare performance of an MPI+OpenMP baseline version of the MILC code with a version incorporating the QPhiX staggered CG solver, for both one-node and multi-node runs.