Acceleration of Heat Diffusion Simulation on Single-Node Systems

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F23%3APU151516" target="_blank" >RIV/00216305:26230/23:PU151516 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.fit.vut.cz/research/publication/13099/" target="_blank" >https://www.fit.vut.cz/research/publication/13099/</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Acceleration of Heat Diffusion Simulation on Single-Node Systems

Popis výsledku v původním jazyce

k-Wave is a MATLAB toolbox for simulating the propagation of acoustic waves in the time domain field. It consists of several computationally intensive simulations, one of which is a heat diffusion simulation. To reduce the overall computation time and cost, an accelerated implementation targeting single-node systems was created. It allows to exploit the full potential of the CPUs by using OpenMP utilizing SIMD units and all cores though thread-level parallelism, and FFTW-like libraries for FFT computation. Alternatively, performing the computation on single and multiple GPUs is also supported on NVIDIA platforms such as DGX-2 or HGX A100 by using CUDA and the cuFFT library. As a result, the simulation output can be obtained faster at a lower price.

Název v anglickém jazyce

Acceleration of Heat Diffusion Simulation on Single-Node Systems

Popis výsledku anglicky

k-Wave is a MATLAB toolbox for simulating the propagation of acoustic waves in the time domain field. It consists of several computationally intensive simulations, one of which is a heat diffusion simulation. To reduce the overall computation time and cost, an accelerated implementation targeting single-node systems was created. It allows to exploit the full potential of the CPUs by using OpenMP utilizing SIMD units and all cores though thread-level parallelism, and FFTW-like libraries for FFT computation. Alternatively, performing the computation on single and multiple GPUs is also supported on NVIDIA platforms such as DGX-2 or HGX A100 by using CUDA and the cuFFT library. As a result, the simulation output can be obtained faster at a lower price.

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

R - Projekt Ramcoveho programu EK

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů