Standard

Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors. / Akimova, Elena N.; Misilov, Vladimir E.
AIP Conference Proceedings: book. Том 3094 1. ред. American Institute of Physics Inc., 2024. 190001 (AIP Conference Proceedings; Том 3094, № 1).

Результаты исследований: Глава в книге, отчете, сборнике статей › Материалы конференции › Рецензирование

Harvard

Akimova, EN & Misilov, VE 2024, Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors. в AIP Conference Proceedings: book. 1 изд., Том. 3094, 190001, AIP Conference Proceedings, № 1, Том. 3094, American Institute of Physics Inc., International Conference of Numerical Analysis and Applied Mathematics 2022, ICNAAM 2022, Heraklion, Греция, 19/09/2022. https://doi.org/10.1063/5.0210370

APA

Akimova, E. N., & Misilov, V. E. (2024). Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors. в AIP Conference Proceedings: book (1 ред., Том 3094). [190001] (AIP Conference Proceedings; Том 3094, № 1). American Institute of Physics Inc.. https://doi.org/10.1063/5.0210370

Vancouver

Akimova EN, Misilov VE. Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors. в AIP Conference Proceedings: book. 1 ред. Том 3094. American Institute of Physics Inc. 2024. 190001. (AIP Conference Proceedings; 1). doi: 10.1063/5.0210370

Author

Akimova, Elena N. ; Misilov, Vladimir E. / Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors. AIP Conference Proceedings: book. Том 3094 1. ред. American Institute of Physics Inc., 2024. (AIP Conference Proceedings; 1).

BibTeX

@inproceedings{bd75f963746a41cdae05e747e404a9a3,
title = "Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors",
abstract = "The paper presents the efficient parallel algorithms for solving the heat equations for the permafrost models. By using the finite difference scheme, we transform the Stefan problem into the problem of solving multiple systems of linear algebraic equations. These systems have tridiagonal coefficient matrices and are solved by the sweep method. On the base of the parallel algorithms, we developed the codes for multicore CPU with OpenMP technology and for GPU with CUDA technology. The novel approach was proposed to achieve the efficient memory access patterns. Numerical experiments show that our new code is up to 2x faster than the previous one, and the GPU implementation achieves a tenfold speedup compared to the serial CPU code.",
author = "Akimova, {Elena N.} and Misilov, {Vladimir E.}",
year = "2024",
doi = "10.1063/5.0210370",
language = "English",
isbn = "978-073544954-1",
volume = "3094",
series = "AIP Conference Proceedings",
publisher = "American Institute of Physics Inc.",
number = "1",
booktitle = "AIP Conference Proceedings",
address = "United States",
edition = "1",
note = "International Conference of Numerical Analysis and Applied Mathematics 2022, ICNAAM 2022 ; Conference date: 19-09-2022 Through 25-09-2022",

}

RIS

TY - GEN

T1 - Solving the Heat Transfer Equations for Permafrost Models Using Multicore and Graphics Processors

AU - Akimova, Elena N.

AU - Misilov, Vladimir E.

PY - 2024

Y1 - 2024

N2 - The paper presents the efficient parallel algorithms for solving the heat equations for the permafrost models. By using the finite difference scheme, we transform the Stefan problem into the problem of solving multiple systems of linear algebraic equations. These systems have tridiagonal coefficient matrices and are solved by the sweep method. On the base of the parallel algorithms, we developed the codes for multicore CPU with OpenMP technology and for GPU with CUDA technology. The novel approach was proposed to achieve the efficient memory access patterns. Numerical experiments show that our new code is up to 2x faster than the previous one, and the GPU implementation achieves a tenfold speedup compared to the serial CPU code.

AB - The paper presents the efficient parallel algorithms for solving the heat equations for the permafrost models. By using the finite difference scheme, we transform the Stefan problem into the problem of solving multiple systems of linear algebraic equations. These systems have tridiagonal coefficient matrices and are solved by the sweep method. On the base of the parallel algorithms, we developed the codes for multicore CPU with OpenMP technology and for GPU with CUDA technology. The novel approach was proposed to achieve the efficient memory access patterns. Numerical experiments show that our new code is up to 2x faster than the previous one, and the GPU implementation achieves a tenfold speedup compared to the serial CPU code.

UR - http://www.scopus.com/inward/record.url?partnerID=8YFLogxK&scp=85196480495

U2 - 10.1063/5.0210370

DO - 10.1063/5.0210370

M3 - Conference contribution

SN - 978-073544954-1

VL - 3094

T3 - AIP Conference Proceedings

BT - AIP Conference Proceedings

PB - American Institute of Physics Inc.

T2 - International Conference of Numerical Analysis and Applied Mathematics 2022, ICNAAM 2022

Y2 - 19 September 2022 through 25 September 2022

ER -

ID: 58895809