Asynchronous distributed-memory task-parallel algorithm for compressible flows on unstructured 3D Eulerian grids

被引：8

作者：

Bakosi, J. ^{[1
]}

Bird, R. ^{[1
]}

Gonzalez, F. ^{[3
]}

Junghans, C. ^{[1
]}

Li, W. ^{[2
]}

Luo, H. ^{[2
]}

Pandare, A. ^{[1
]}

Waltz, J. ^{[1
]}

机构：

[1] Los Alamos Natl Lab, Los Alamos, NM 87544 USA

[2] North Carolina State Univ, Raleigh, NC USA

[3] Strong Analyt, Chicago, IL USA

来源：

ADVANCES IN ENGINEERING SOFTWARE | 2021年 / 160卷

关键词：

Shock hydrodynamics; Finite element method; Flux-corrected transport; Charm plus; Automatic load balancing; ELEMENT; SYSTEMS;

D O I：

10.1016/j.advengsoft.2020.102962

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

We discuss the implementation of a finite element method, used to numerically solve the Euler equations of compressible flows, using an asynchronous runtime system (RTS). The algorithm is implemented for distributed-memory machines, using stationary unstructured 3D meshes, combining data-, and task-parallelism on top of the Charm++ RTS. Charm++'s execution model is asynchronous by default, allowing arbitrary overlap of computation and communication. Task-parallelism allows scheduling parts of an algorithm independently of, or dependent on, each other. Built-in automatic load balancing enables continuous redistribution of computational load by migration of work units based on real-time CPU load measurement. The RTS also features automatic checkpointing, fault tolerance, resilience against hardware failure, and supports power-, and energy-aware computation. We demonstrate scalability up to 25 x 10(9) cells at O(10(4)) compute cores and the benefits of automatic load balancing for irregular workloads. The full source code with documentation is available at https://quinoacomputing.org.

引用

页数：21

共 50 条

[41] Cell-vertex based parallel and adaptive explicit 3D flow solution on unstructured grids
Yilmaz, E
Kavsaoglu, MS
Akay, HU
Akmandor, IS
INTERNATIONAL JOURNAL OF COMPUTATIONAL FLUID DYNAMICS, 2001, 14 (04) : 271 - 286
[42] Large scale parallel distributed solution of 3D microgravity flows.
Carey, GF
McLay, R
Bicken, G
Harlé, C
Swift, S
PROCEEDINGS OF THE HIGH-PERFORMANCE COMPUTING (HPC'98), 1998, : 39 - 44
[43] parMATT: parallel multiple alignment of protein 3D-structures with translations and twists for distributed-memory systems
Shegay, Maksim V.
Suplatov, Dmitry A.
Popova, Nina N.
Svedas, Vytas K.
Voevodin, Vladimir V.
BIOINFORMATICS, 2019, 35 (21) : 4456 - 4458
[44] Underwater 3D imaging by distributed and parallel subarray beamforming algorithm
Tian, X. (tianx@mail.bme.zju.edu.cn), 1600, Zhejiang University (48):
[45] Adaptive deformation of 3D unstructured meshes with curved body fitted boundaries with application to unsteady compressible flows
Cirrottola, Luca
Ricchiuto, Mario
Froehly, Algiane
Re, Barbara
Guardone, Alberto
Quaranta, Giuseppe
JOURNAL OF COMPUTATIONAL PHYSICS, 2021, 433
[46] A parallel implementation of a 3D-BEM-algorithm using distributed memory algorithms
Mader, G
Uhlmann, FH
IEEE TRANSACTIONS ON MAGNETICS, 1997, 33 (02) : 1796 - 1799
[47] A general reconstruction algorithm for simulating flows with complex 3D immersed boundaries on Cartesian grids
Gilmanov, A
Sotiropoulos, F
Balaras, E
JOURNAL OF COMPUTATIONAL PHYSICS, 2003, 191 (02) : 660 - 669
[48] Scalable 3D Hybrid Parallel Delaunay Image-to-Mesh Conversion Algorithm for Distributed Shared Memory Architectures
Feng, Daming
Tsolakis, Christos
Chernikov, Andrey N.
Chrisochoides, Nikos P.
24TH INTERNATIONAL MESHING ROUNDTABLE, 2015, 124 : 18 - 30
[49] Scalable 3D hybrid parallel Delaunay image-to-mesh conversion algorithm for distributed shared memory architectures
Feng, Daming
Tsolakis, Christos
Chernikov, Andrey N.
Chrisochoides, Nikos P.
COMPUTER-AIDED DESIGN, 2017, 85 : 10 - 19
[50] A 3D parallel boundary element method on unstructured triangular grids for fully nonlinear wave-body interactions
Dombre, E.
Harris, J. C.
Benoit, M.
Violeau, D.
Peyrard, C.
OCEAN ENGINEERING, 2019, 171 : 505 - 518

← 1 2 3 4 5 →