Large-Scale direct numerical simulations of turbulence using GPUs and modern Fortran

被引：3

作者：

Karp, Martin ^{[1
,5
]}

Massaro, Daniele ^{[2
]}

Jansson, Niclas ^{[3
]}

Hart, Alistair ^{[4
]}

Wahlgren, Jacob ^{[1
]}

Schlatter, Philipp ^{[2
]}

Markidis, Stefano ^{[1
]}

机构：

[1] KTH Royal Inst Technol, Div Computat Sci & Technol, EECS, Stockholm, Sweden

[2] KTH Royal Inst Technol, SimEx, FLOW, Engn Mech, Stockholm, Sweden

[3] KTH Royal Inst Technol, PDC Ctr High Performance Comp, EECS, Stockholm, Sweden

[4] Hewlett Packard Enterpise HPE, London, England

[5] KTH Royal Inst Technol, Div Computat Sci & Technol, EECS, Lindstedtsvagen 5, S-10044 Stockholm, Sweden

来源：

INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS | 2023年 / 37卷 / 05期

关键词：

GPU; Fortran; spectral element method; computational fluid dynamics; direct numerical simulation; high performance computing; ACCELERATION; ALGORITHM;

D O I：

10.1177/10943420231158616

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

We present our approach to making direct numerical simulations of turbulence with applications in sustainable shipping. We use modern Fortran and the spectral element method to leverage and scale on supercomputers powered by the Nvidia A100 and the recent AMD Instinct MI250X GPUs, while still providing support for user software developed in Fortran. We demonstrate the efficiency of our approach by performing the world's first direct numerical simulation of the flow around a Flettner rotor at Re = 30,000 and its interaction with a turbulent boundary layer. We present a performance comparison between the AMD Instinct MI250X and Nvidia A100 GPUs for scalable computational fluid dynamics. Our results show that one MI250X offers performance on par with two A100 GPUs and has a similar power efficiency based on readings from on-chip energy sensors.

引用

页码：487 / 502

页数：16

共 50 条

[1] large-scale direct numerical simulations of turbulence using GPUs and modern Fortran (feb, 10.1177/10943420231158616, 2023)
不详
[J]. INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS, 2023,
[2] A study of turbulence structure and large-scale motion in the Ekman layer through direct numerical simulations
Shingai, K
Kawamura, H
[J]. JOURNAL OF TURBULENCE, 2004, 5
[3] Accurate Parallel Algorithm for Tracking Inertial Particles in Large-Scale Direct Numerical Simulations of Turbulence
Ishihara, Takashi
Enohata, Kei
Morishita, Koji
Yokokawa, Mitsuo
Ishii, Katsuya
[J]. PARALLEL COMPUTING TECHNOLOGIES (PACT 2015), 2015, 9251 : 522 - 527
[4] NUMERICAL SIMULATIONS OF LARGE-SCALE PLASMA TURBULENCE IN THE DAYTIME EQUATORIAL ELECTROJET
RONCHI, C
SUDAN, RN
FARLEY, DT
[J]. JOURNAL OF GEOPHYSICAL RESEARCH-SPACE PHYSICS, 1991, 96 (A12) : 21263 - 21279
[5] Large-Scale Compressible-Flow Direct Numerical Simulations
Sandberg, R. D.
[J]. DIRECT AND LARGE-EDDY SIMULATION X, 2018, 24 : 25 - 33
[6] Semi-automatic porting of a large-scale Fortran CFD code to GPUs
Corrigan, Andrew
Camelli, Fernando
Loehner, Rainald
Mut, Fernando
[J]. INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN FLUIDS, 2012, 69 (02) : 314 - 331
[7] Large-Scale Stochastic Learning using GPUs
Parnell, Thomas
Dunner, Celestine
Atasu, Kubilay
Sifalakis, Manolis
Pozidis, Haris
[J]. 2017 IEEE INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM WORKSHOPS (IPDPSW), 2017, : 419 - 428
[8] Large-scale simulations on multiple Graphics Processing Units (GPUs) for the direct simulation Monte Carlo method
Su, C. -C.
Smith, M. R.
Kuo, F. -A.
Wu, J. -S.
Hsieh, C. -W.
Tseng, K. -C.
[J]. JOURNAL OF COMPUTATIONAL PHYSICS, 2012, 231 (23) : 7932 - 7958
[9] Preserving large-scale features in simulations of elastic turbulence
Yerasi, Sumithra R.
Picardo, Jason R.
Gupta, Anupam
Vincenzi, Dario
[J]. arXiv, 2023,
[10] Optimizing Pairwise Box Intersection Checking on GPUs for Large-Scale Simulations
Lo, Shih-Hsiang
Lee, Che-Rung
Chung, I-Hsin
Chung, Yeh-Ching
[J]. ACM TRANSACTIONS ON MODELING AND COMPUTER SIMULATION, 2013, 23 (03):

← 1 2 3 4 5 →