GPU computing for patient-specific model of pulmonary airflow

Takami Yamaguchi; Y. Imai; T. Miki; Takuji Ishikawa

GPU computing for patient-specific model of pulmonary airflow

Takami Yamaguchi, Y. Imai, T. Miki, Takuji Ishikawa

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We propose an implementation of Lattice Boltzmann (LB) method on GPUs for simulating airflow in pulmonary airways with complex branches. An adaptive meshing method is developed for optimizing memory accessing, where the global domain comprises unstructured subdomains, while the local subdomain consists of a structured grids.We also develop a multi-GPU computing method based on a domain decomposition. For strong scaling tests with a subject-specific geometry (12 million LB nodes), the performance on 8 GPUs is approximately 200 GFLOPS, which is 100 times faster computation than 8 CPU cores.

Original language	English
Title of host publication	Computational Modelling of Objects Represented in Images
Subtitle of host publication	Fundamentals, Methods and Applications III - Proceedings of the International Symposium, CompIMAGE 2012
Pages	239-242
Number of pages	4
Publication status	Published - 2012 Oct 26
Event	3rd International Symposium on Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications, CompIMAGE 2012 - Rome, Italy Duration: 2012 Sept 5 → 2012 Sept 7

Other

Other	3rd International Symposium on Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications, CompIMAGE 2012
Country/Territory	Italy
City	Rome
Period	12/9/5 → 12/9/7

ASJC Scopus subject areas

Computational Theory and Mathematics
Modelling and Simulation

Cite this

GPU computing for patient-specific model of pulmonary airflow. / Yamaguchi, Takami; Imai, Y.; Miki, T. et al.

Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications III - Proceedings of the International Symposium, CompIMAGE 2012. 2012. p. 239-242.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Yamaguchi, T, Imai, Y, Miki, T & Ishikawa, T 2012, GPU computing for patient-specific model of pulmonary airflow. in Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications III - Proceedings of the International Symposium, CompIMAGE 2012. pp. 239-242, 3rd International Symposium on Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications, CompIMAGE 2012, Rome, Italy, 12/9/5.

@inproceedings{e0b5070a2b7e4a78a99357c03f64a412,

title = "GPU computing for patient-specific model of pulmonary airflow",

abstract = "We propose an implementation of Lattice Boltzmann (LB) method on GPUs for simulating airflow in pulmonary airways with complex branches. An adaptive meshing method is developed for optimizing memory accessing, where the global domain comprises unstructured subdomains, while the local subdomain consists of a structured grids.We also develop a multi-GPU computing method based on a domain decomposition. For strong scaling tests with a subject-specific geometry (12 million LB nodes), the performance on 8 GPUs is approximately 200 GFLOPS, which is 100 times faster computation than 8 CPU cores.",

author = "Takami Yamaguchi and Y. Imai and T. Miki and Takuji Ishikawa",

year = "2012",

month = oct,

day = "26",

language = "English",

isbn = "9780415621342",

pages = "239--242",

booktitle = "Computational Modelling of Objects Represented in Images",

note = "3rd International Symposium on Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications, CompIMAGE 2012 ; Conference date: 05-09-2012 Through 07-09-2012",

}

TY - GEN

T1 - GPU computing for patient-specific model of pulmonary airflow

AU - Yamaguchi, Takami

AU - Imai, Y.

AU - Miki, T.

AU - Ishikawa, Takuji

PY - 2012/10/26

Y1 - 2012/10/26

N2 - We propose an implementation of Lattice Boltzmann (LB) method on GPUs for simulating airflow in pulmonary airways with complex branches. An adaptive meshing method is developed for optimizing memory accessing, where the global domain comprises unstructured subdomains, while the local subdomain consists of a structured grids.We also develop a multi-GPU computing method based on a domain decomposition. For strong scaling tests with a subject-specific geometry (12 million LB nodes), the performance on 8 GPUs is approximately 200 GFLOPS, which is 100 times faster computation than 8 CPU cores.

AB - We propose an implementation of Lattice Boltzmann (LB) method on GPUs for simulating airflow in pulmonary airways with complex branches. An adaptive meshing method is developed for optimizing memory accessing, where the global domain comprises unstructured subdomains, while the local subdomain consists of a structured grids.We also develop a multi-GPU computing method based on a domain decomposition. For strong scaling tests with a subject-specific geometry (12 million LB nodes), the performance on 8 GPUs is approximately 200 GFLOPS, which is 100 times faster computation than 8 CPU cores.

UR - http://www.scopus.com/inward/record.url?scp=84867738203&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867738203&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84867738203

SN - 9780415621342

SP - 239

EP - 242

BT - Computational Modelling of Objects Represented in Images

T2 - 3rd International Symposium on Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications, CompIMAGE 2012

Y2 - 5 September 2012 through 7 September 2012

ER -

GPU computing for patient-specific model of pulmonary airflow

Abstract

Other

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this