arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.30128cond-mat.mtrl-sci
收藏

迈向基于混合精度计算和残差子空间迭代压缩通信的百亿亿次全相对论赝势密度泛函理论计算

Towards exascale fully relativistic pseudopotential density functional theory calculations enabled by mixed-precision computation and compressed-communication using residual based subspace iteration

Department of Computational and Data Sciences, Indian Institute of Science(印度科学学院计算与数据科学系);Department of Mechanical Engineering, University of Michigan(密歇根大学机械工程系);Argonne National Laboratory(阿贡国家实验室)

Nikhil Kodali, Gourab Panigrahi, Nishant Gupta, Kartick Ramakrishnan, Sundaresan G, Rudra Panch, Sambit Das, Vishwas Rao, Phani Motamarri

AI总结:

针对非共线磁性和自旋轨道耦合带来的计算挑战,提出结合残差Chebyshev滤波子空间迭代、混合精度计算和压缩通信的GPU高性能框架,实现百亿亿次规模的全相对论赝势DFT模拟。

摘要:

非共线(NC)磁性和自旋轨道耦合(SOC)对于具有显著相对论效应和磁挫折的预测性从头算材料模拟不可或缺,然而它们通过引入每个电子的复杂二分量波函数以及随之而来的更大特征问题,显著增加了立方标度密度泛函理论(DFT)的成本。我们提出了一个以GPU为中心的高性能NC-SOC DFT框架,该框架结合了:(i)求解有限元(FE)离散化DFT方程的算法进展;(ii)基于残差的Chebyshev滤波子空间迭代(R-ChFSI),对不精确的矩阵-向量积具有容忍性,用于求解得到的稀疏广义特征问题;(iii)加速FE泊松求解器的无矩阵策略;(iv)R-ChFSI支持的混合精度计算,配合压缩比超过4倍的块浮点压缩MPI通信,在降低计算和数据移动成本的同时保持双精度鲁棒性;以及(v)一种通信高效的能带划分算法以提高可扩展性。数值结果表明,在百亿亿次架构上实现了改进的求解时间和出色的可扩展性,使得能够对多达100,000个电子进行全相对论赝势DFT模拟。

英文摘要:

Materials exhibiting noncollinear magnetism or strong spin-orbit-coupling underpin many spintronic and topological applications, but their simulations require complex two-component spinors and costs substantially more than scalar density functional theory (DFT). We present a GPU-centric exascale finite-element DFT framework with noncollinear magnetism and spin-orbit-coupling, combining (i) adaptive higher-order finite-element discretization, (ii) a matrix-free Poisson solver, (iii) residual-based Chebyshev filtered subspace iteration (R-ChFSI) for sparse generalized eigenproblems, (iv) R-ChFSI-enabled mixed precision computation with block floating-point compressed MPI communication, and (v) communication-efficient band partitioning. R-ChFSI permits inexact matrix-multivector products and $4\times$--$6.4\times$ compression of communicated data relative to FP64 while preserving double-precision robustness and reducing computation and data movement. Results demonstrate strong scaling on Aurora and Frontier, with up to $3.4\times$ faster Chebyshev filtering. On Aurora, total self-consistent-field (SCF) solve wall time reduces by $2\times$. An 800-node (9,600-GPU) calculation demonstrates fully relativistic pseudopotential DFT for $\sim$80,000 electrons in under 7 minutes per SCF iteration.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑