MILC staggered conjugate gradient performance on Intel KNL
Department of Physics, Indiana University, Bloomington IN 47405, USA
arXiv:1611.00728 [hep-lat], (2 Nov 2016)
@article{detar2016milc,
title={MILC staggered conjugate gradient performance on Intel KNL},
author={DeTar, Carleton and Doerfler, Douglas and Gottlieb, Steven and Jha, Ashish and Kalamkar, Dhiraj and Li, Ruizi and Toussaint, Doug},
year={2016},
month={nov},
archivePrefix={"arXiv"},
primaryClass={hep-lat}
}
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 generation 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.
November 3, 2016 by hgpu
