GPU-optimized Direct Fourier Method for On-line Tomography; Fundamenta Informaticae; Vol. 141, iss. 2-3

GPU-optimized Direct Fourier Method for On-line Tomography; Fundamenta Informaticae; Vol. 141, iss. 2-3

Manylion Llyfryddiaeth
Parent link:Fundamenta Informaticae
Vol. 141, iss. 2-3.— 2015.— [P. 245-258]
Awdur Corfforaethol: Национальный исследовательский Томский политехнический университет (ТПУ) Институт кибернетики (ИК) Кафедра автоматики и компьютерных систем (АИКС)
Awduron Eraill: Шкарин А. В. Андрей Владимирович, Ametova E. S. Evelina Serverovna, Chilingaryan S. Suren, Dritschler T. Timo, Kopmann A. Andreas, Vogelgesang M. Matthias, Shkarin R. V. Roman Vladimirovich, Tsapko S. G. Sergey Gennadyevich, Mirone A. Alessandro
Crynodeb:Title screen
On-line monitoring of synchrotron 3D-imaging experiments requires very fast tomographic reconstruction. Direct Fourier methods (DFM) have the potential to be faster than standard Filtered Backprojection. We have evaluated multiple DFMs using various interpolation techniques. We compared reconstruction quality and studied the parallelization potential. A method using Direct Fourier Inversion (DFI) and a sinc-based interpolation was selected and parallelized for execution on GPUs. Several optimization steps were considered to boost the performance. Finally we evaluated the achieved performance for the latest generation of GPUs from NVIDIA and AMD. The results show that tomographic reconstruction with a throughput of more than 1.5 GB/sec on a single GPU is possible.
Режим доступа: по договору с организацией-держателем ресурса
Iaith:Saesneg
Cyhoeddwyd: 2015
Pynciau:
электронный ресурс
труды учёных ТПУ
компьютерная томография
томографическая реконструкция
Mynediad Ar-lein:http://dx.doi.org/10.3233/FI-2015-1274
Fformat: Electronig Pennod Llyfr
KOHA link:https://koha.lib.tpu.ru/cgi-bin/koha/opac-detail.pl?biblionumber=646120
Disgrifiad
Crynodeb:Title screen
On-line monitoring of synchrotron 3D-imaging experiments requires very fast tomographic reconstruction. Direct Fourier methods (DFM) have the potential to be faster than standard Filtered Backprojection. We have evaluated multiple DFMs using various interpolation techniques. We compared reconstruction quality and studied the parallelization potential. A method using Direct Fourier Inversion (DFI) and a sinc-based interpolation was selected and parallelized for execution on GPUs. Several optimization steps were considered to boost the performance. Finally we evaluated the achieved performance for the latest generation of GPUs from NVIDIA and AMD. The results show that tomographic reconstruction with a throughput of more than 1.5 GB/sec on a single GPU is possible.
Режим доступа: по договору с организацией-держателем ресурса
DOI:10.3233/FI-2015-1274

Eitemau Tebyg