Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | Bildverarbeitung fur die Medizin 2016 |
| Untertitel | Algorithmen – Systeme – Anwendungen - Proceedings des Workshops |
| Herausgeber/-innen | Thomas M. Deserno, Heinz Handels, Thomas Tolxdorff, Hans-Peter Meinzer |
| Herausgeber (Verlag) | Kluwer Academic Publishers |
| Seiten | 206-211 |
| Seitenumfang | 6 |
| ISBN (Print) | 9783662494646 |
| Publikationsstatus | Veröffentlicht - 1 Jan. 2017 |
| Extern publiziert | Ja |
| Veranstaltung | Workshops on Image processing for the medicine, 2016 - Berlin, Deutschland Dauer: 13 März 2016 → 15 März 2016 |
Publikationsreihe
| Name | Informatik aktuell |
|---|---|
| ISSN (Print) | 1431-472X |
Abstract
Driven by improved computational throughput, multi- and many-core processors have been increasingly used in medical image processing. As these systems contain a discrete memory node, programmers have to manually manage the data transfer. To improve throughput by overlapping data transfers and task execution, special hardware details have to be known and should be considered with care. Data management could be even more tedious when the data size exceeds the GPU memory. In this work, we present a library that provides a convenient interface for CT reconstruction. Further, it contains a transparent data management, automatic data partitioning in case the GPU memory is insufficient, and overlapping techniques for improved performance. Our evaluations reveal that the library is able to reduce the amount of necessary code lines by ≈ 63% with respect to a comparable manual implementation. Additionally, a speedup of 38.1% for a volume size of 256 (10.7% for a volume size of 512) could be achieved by the library’s overlapping technique.
ASJC Scopus Sachgebiete
- Mathematik (insg.)
- Modellierung und Simulation
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Bildverarbeitung fur die Medizin 2016: Algorithmen – Systeme – Anwendungen - Proceedings des Workshops. Hrsg. / Thomas M. Deserno; Heinz Handels; Thomas Tolxdorff; Hans-Peter Meinzer. Kluwer Academic Publishers, 2017. S. 206-211 (Informatik aktuell).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - A Memory Management Library for CT-Reconstruction on GPUs
AU - Wu, Hao
AU - Berger, Martin
AU - Maier, Andreas
AU - Lohmann, Daniel
PY - 2017/1/1
Y1 - 2017/1/1
N2 - Driven by improved computational throughput, multi- and many-core processors have been increasingly used in medical image processing. As these systems contain a discrete memory node, programmers have to manually manage the data transfer. To improve throughput by overlapping data transfers and task execution, special hardware details have to be known and should be considered with care. Data management could be even more tedious when the data size exceeds the GPU memory. In this work, we present a library that provides a convenient interface for CT reconstruction. Further, it contains a transparent data management, automatic data partitioning in case the GPU memory is insufficient, and overlapping techniques for improved performance. Our evaluations reveal that the library is able to reduce the amount of necessary code lines by ≈ 63% with respect to a comparable manual implementation. Additionally, a speedup of 38.1% for a volume size of 256 (10.7% for a volume size of 512) could be achieved by the library’s overlapping technique.
AB - Driven by improved computational throughput, multi- and many-core processors have been increasingly used in medical image processing. As these systems contain a discrete memory node, programmers have to manually manage the data transfer. To improve throughput by overlapping data transfers and task execution, special hardware details have to be known and should be considered with care. Data management could be even more tedious when the data size exceeds the GPU memory. In this work, we present a library that provides a convenient interface for CT reconstruction. Further, it contains a transparent data management, automatic data partitioning in case the GPU memory is insufficient, and overlapping techniques for improved performance. Our evaluations reveal that the library is able to reduce the amount of necessary code lines by ≈ 63% with respect to a comparable manual implementation. Additionally, a speedup of 38.1% for a volume size of 256 (10.7% for a volume size of 512) could be achieved by the library’s overlapping technique.
UR - http://www.scopus.com/inward/record.url?scp=85019848963&partnerID=8YFLogxK
U2 - 10.1007/978-3-662-49465-3_37
DO - 10.1007/978-3-662-49465-3_37
M3 - Conference contribution
AN - SCOPUS:85019848963
SN - 9783662494646
T3 - Informatik aktuell
SP - 206
EP - 211
BT - Bildverarbeitung fur die Medizin 2016
A2 - Deserno, Thomas M.
A2 - Handels, Heinz
A2 - Tolxdorff, Thomas
A2 - Meinzer, Hans-Peter
PB - Kluwer Academic Publishers
T2 - Workshops on Image processing for the medicine, 2016
Y2 - 13 March 2016 through 15 March 2016
ER -