3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

Reconstructing 3D parameter maps of huge volumes entirely on the GPU is highly desirable due to the offeredcomputation speed-up. However, GPU memory restrictions limit the coverable volume. To overcome thislimitation, a double-buffering strategy in combination with model-based parameter quantification and 3D-TGVregularization is proposed. This combination warrants whole volume reconstruction while maintaining thespeed advantages of GPU-based computation. In contrast to sequential transfers, double-buffering splits thevolume into blocks and overlaps memory transfer and kernel execution concurrently, hiding memory latency.The proposed method is able to reconstruct arbitrary large volumes within 5.3 min/slice, even on a single GPU.
Original languageEnglish
Title of host publicationProceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019
Publication statusPublished - 2019
EventISMRM 27th Annual Meeting & Exhibition - Palais des congrès de Montréal, Montréal, Canada
Duration: 11 May 201916 May 2019
https://www.ismrm.org/19m/

Conference

ConferenceISMRM 27th Annual Meeting & Exhibition
Abbreviated titleISMRM 2019
CountryCanada
CityMontréal
Period11/05/1916/05/19
Internet address

Fingerprint

Hardware
Data storage equipment
Graphics processing unit

Cite this

Maier, O., & Stollberger, R. (2019). 3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering. In Proceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019 [4839]

3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering. / Maier, Oliver; Stollberger, Rudolf.

Proceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019. 2019. 4839.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Maier, O & Stollberger, R 2019, 3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering. in Proceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019., 4839, ISMRM 27th Annual Meeting & Exhibition, Montréal, Canada, 11/05/19.
Maier O, Stollberger R. 3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering. In Proceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019. 2019. 4839
@inproceedings{d694f9e24c4f4385abc100941a152932,
title = "3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering",
abstract = "Reconstructing 3D parameter maps of huge volumes entirely on the GPU is highly desirable due to the offeredcomputation speed-up. However, GPU memory restrictions limit the coverable volume. To overcome thislimitation, a double-buffering strategy in combination with model-based parameter quantification and 3D-TGVregularization is proposed. This combination warrants whole volume reconstruction while maintaining thespeed advantages of GPU-based computation. In contrast to sequential transfers, double-buffering splits thevolume into blocks and overlaps memory transfer and kernel execution concurrently, hiding memory latency.The proposed method is able to reconstruct arbitrary large volumes within 5.3 min/slice, even on a single GPU.",
author = "Oliver Maier and Rudolf Stollberger",
year = "2019",
language = "English",
booktitle = "Proceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019",

}

TY - GEN

T1 - 3D Model-Based Parameter Quantification on Resource Constrained Hardware using Double-Buffering

AU - Maier, Oliver

AU - Stollberger, Rudolf

PY - 2019

Y1 - 2019

N2 - Reconstructing 3D parameter maps of huge volumes entirely on the GPU is highly desirable due to the offeredcomputation speed-up. However, GPU memory restrictions limit the coverable volume. To overcome thislimitation, a double-buffering strategy in combination with model-based parameter quantification and 3D-TGVregularization is proposed. This combination warrants whole volume reconstruction while maintaining thespeed advantages of GPU-based computation. In contrast to sequential transfers, double-buffering splits thevolume into blocks and overlaps memory transfer and kernel execution concurrently, hiding memory latency.The proposed method is able to reconstruct arbitrary large volumes within 5.3 min/slice, even on a single GPU.

AB - Reconstructing 3D parameter maps of huge volumes entirely on the GPU is highly desirable due to the offeredcomputation speed-up. However, GPU memory restrictions limit the coverable volume. To overcome thislimitation, a double-buffering strategy in combination with model-based parameter quantification and 3D-TGVregularization is proposed. This combination warrants whole volume reconstruction while maintaining thespeed advantages of GPU-based computation. In contrast to sequential transfers, double-buffering splits thevolume into blocks and overlaps memory transfer and kernel execution concurrently, hiding memory latency.The proposed method is able to reconstruct arbitrary large volumes within 5.3 min/slice, even on a single GPU.

M3 - Conference contribution

BT - Proceedings of the ISMRM 27TH Annual Meeting & Exhibition, 2019

ER -