Multi-GPU Kinetic Solvers using MPI and CUDA

被引：2

作者：

Zabelok, Sergey ^{[1
]}

Arslanbekov, Robert ^{[1
]}

Kolobov, Vladimir ^{[1
]}

机构：

[1] Russian Acad Sci, Ctr Comp, Moscow 119333, Russia

来源：

PROCEEDINGS OF THE 29TH INTERNATIONAL SYMPOSIUM ON RAREFIED GAS DYNAMICS | 2014年 / 1628卷

关键词：

unified flow solver; adaptive mesh refinement; discrete velocity method; boltzmann kinetic equation; direct simulation monte carlo; lattice boltzmann method; graphics processing units; CUDA; MPI; ALGORITHM;

D O I：

10.1063/1.4902640

中图分类号：

O59 [应用物理学];

学科分类号：

摘要：

This paper describes recent progress towards porting a Unified Flow Solver OHS to heterogeneous parallel computing. The main challenge of porting CIF'S to graphics processing units (GPUs) comes from the dynamically adapted mesh, which causes irregular data access. We describe the implementation of CUDA kernels for three modules in UFS: the direct Boltzmann solver using discrete velocity method (DVM), the DSMC module, and the Lattice Boltzmann Method (LBM) solver, all using octree Cartesian mesh with adaptive Mesh Refinement (AMR). Double digit speedup on single GPU and good scaling for multi-GPU has been demonstrated.

引用

页码：539 / 546

页数：8