Fast GPU-beamforming of\\Row-Column Addressed Probe Data

Matthias Bo Stuart; Patrick Møller Jensen; Julian Thomas Reckeweg  Olsen; Alexander Borch Kristensen; Mikkel Schou; Bernd Dammann; Hans Henrik Brandenborg Sørensen; Jørgen Arendt Jensen

doi:10.1109/ultsym.2019.8925802

Fast GPU-beamforming of\\Row-Column Addressed Probe Data

Matthias Bo Stuart, Patrick Møller Jensen, Julian Thomas Reckeweg Olsen, Alexander Borch Kristensen, Mikkel Schou, Bernd Dammann, Hans Henrik Brandenborg Sørensen, Jørgen Arendt Jensen

Technical University of Denmark

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

288 Downloads (Pure)

Abstract

A delay-and-sum beamformer for 3D imaging using row-column arrays and written in CUDA is presented and compared to an existing similar GPU-based beamformer written in the MATLAB programming language. Data from a 192+192 rowcolumn array single element emission sequence is simulated and beamformed. The two beamformers’ performance is evaluated in two synthetic aperture setups comprised of 1) two orthogonal planes and 2) a full volume on three different NVIDIA GPUs: a 1050 Ti, a 1080 Ti, and a TITAN V. The execution time and the sample throughput (samples beamformed per second) are reported. The CUDA beamformer performs consistently better than the MATLAB beamformer with speed-ups ranging from 1.9 to 64.6 times, and the worst-case throughput of the CUDA beamformer exceeds the best-case of the MATLAB beamformer. High-resolution images of crossing planes can be beamformed at up to 13 Hz, while a 50-by-50-by-20 cubic-millimeter highresolution volume sampled at one quarter of a millimeter is beamformed in 3 seconds.

Original language	English
Title of host publication	Proceedings of 2019 IEEE International Ultrasonics Symposium
Number of pages	4
Publisher	IEEE
Publication date	2019
Pages	1497-1500
ISBN (Print)	9781728145952
DOIs	https://doi.org/10.1109/ultsym.2019.8925802
Publication status	Published - 2019
Event	2019 IEEE International Ultrasonics Symposium - SEC Glasgow, Glasgow, United Kingdom Duration: 6 Oct 2019 → 9 Oct 2019 http://attend.ieee.org/ius-2019/

Conference

Conference	2019 IEEE International Ultrasonics Symposium
Location	SEC Glasgow
Country/Territory	United Kingdom
City	Glasgow
Period	06/10/2019 → 09/10/2019
Internet address	http://attend.ieee.org/ius-2019/

Access to Document

10.1109/ultsym.2019.8925802

FulltextAccepted author manuscript, 218 KB

OpenUrl availability

Full text

Cite this

@inproceedings{48f640801c094cd7876ba272d0f1bd04,

title = "Fast GPU-beamforming of\\Row-Column Addressed Probe Data",

abstract = "A delay-and-sum beamformer for 3D imaging using row-column arrays and written in CUDA is presented and compared to an existing similar GPU-based beamformer written in the MATLAB programming language. Data from a 192+192 rowcolumn array single element emission sequence is simulated and beamformed. The two beamformers{\textquoteright} performance is evaluated in two synthetic aperture setups comprised of 1) two orthogonal planes and 2) a full volume on three different NVIDIA GPUs: a 1050 Ti, a 1080 Ti, and a TITAN V. The execution time and the sample throughput (samples beamformed per second) are reported. The CUDA beamformer performs consistently better than the MATLAB beamformer with speed-ups ranging from 1.9 to 64.6 times, and the worst-case throughput of the CUDA beamformer exceeds the best-case of the MATLAB beamformer. High-resolution images of crossing planes can be beamformed at up to 13 Hz, while a 50-by-50-by-20 cubic-millimeter highresolution volume sampled at one quarter of a millimeter is beamformed in 3 seconds.",

author = "Stuart, {Matthias Bo} and Jensen, {Patrick M{\o}ller} and Olsen, {Julian Thomas Reckeweg} and Kristensen, {Alexander Borch} and Mikkel Schou and Bernd Dammann and S{\o}rensen, {Hans Henrik Brandenborg} and Jensen, {J{\o}rgen Arendt}",

year = "2019",

doi = "10.1109/ultsym.2019.8925802",

language = "English",

isbn = "9781728145952 ",

pages = "1497--1500",

booktitle = "Proceedings of 2019 IEEE International Ultrasonics Symposium",

publisher = "IEEE",

address = "United States",

note = "2019 IEEE International Ultrasonics Symposium, IUS 2019 ; Conference date: 06-10-2019 Through 09-10-2019",

url = "http://attend.ieee.org/ius-2019/",

}

Stuart, MB, Jensen, PM, Olsen, JTR, Kristensen, AB, Schou, M, Dammann, B , Sørensen, HHB & Jensen, JA 2019, Fast GPU-beamforming of\\Row-Column Addressed Probe Data. in Proceedings of 2019 IEEE International Ultrasonics Symposium. IEEE, pp. 1497-1500, 2019 IEEE International Ultrasonics Symposium, Glasgow, United Kingdom, 06/10/2019. https://doi.org/10.1109/ultsym.2019.8925802

TY - GEN

T1 - Fast GPU-beamforming of\\Row-Column Addressed Probe Data

AU - Stuart, Matthias Bo

AU - Jensen, Patrick Møller

AU - Olsen, Julian Thomas Reckeweg

AU - Kristensen, Alexander Borch

AU - Schou, Mikkel

AU - Dammann, Bernd

AU - Sørensen, Hans Henrik Brandenborg

AU - Jensen, Jørgen Arendt

PY - 2019

Y1 - 2019

N2 - A delay-and-sum beamformer for 3D imaging using row-column arrays and written in CUDA is presented and compared to an existing similar GPU-based beamformer written in the MATLAB programming language. Data from a 192+192 rowcolumn array single element emission sequence is simulated and beamformed. The two beamformers’ performance is evaluated in two synthetic aperture setups comprised of 1) two orthogonal planes and 2) a full volume on three different NVIDIA GPUs: a 1050 Ti, a 1080 Ti, and a TITAN V. The execution time and the sample throughput (samples beamformed per second) are reported. The CUDA beamformer performs consistently better than the MATLAB beamformer with speed-ups ranging from 1.9 to 64.6 times, and the worst-case throughput of the CUDA beamformer exceeds the best-case of the MATLAB beamformer. High-resolution images of crossing planes can be beamformed at up to 13 Hz, while a 50-by-50-by-20 cubic-millimeter highresolution volume sampled at one quarter of a millimeter is beamformed in 3 seconds.

AB - A delay-and-sum beamformer for 3D imaging using row-column arrays and written in CUDA is presented and compared to an existing similar GPU-based beamformer written in the MATLAB programming language. Data from a 192+192 rowcolumn array single element emission sequence is simulated and beamformed. The two beamformers’ performance is evaluated in two synthetic aperture setups comprised of 1) two orthogonal planes and 2) a full volume on three different NVIDIA GPUs: a 1050 Ti, a 1080 Ti, and a TITAN V. The execution time and the sample throughput (samples beamformed per second) are reported. The CUDA beamformer performs consistently better than the MATLAB beamformer with speed-ups ranging from 1.9 to 64.6 times, and the worst-case throughput of the CUDA beamformer exceeds the best-case of the MATLAB beamformer. High-resolution images of crossing planes can be beamformed at up to 13 Hz, while a 50-by-50-by-20 cubic-millimeter highresolution volume sampled at one quarter of a millimeter is beamformed in 3 seconds.

U2 - 10.1109/ultsym.2019.8925802

DO - 10.1109/ultsym.2019.8925802

M3 - Article in proceedings

SN - 9781728145952

SP - 1497

EP - 1500

BT - Proceedings of 2019 IEEE International Ultrasonics Symposium

PB - IEEE

T2 - 2019 IEEE International Ultrasonics Symposium

Y2 - 6 October 2019 through 9 October 2019

ER -

Fast GPU-beamforming of\\Row-Column Addressed Probe Data

Abstract

Conference

Access to Document

OpenUrl availability

Fingerprint

Cite this