A comparative study of an X-ray tomography reconstruction algorithm in accelerated and cloud computing systems

摘要

With the increase of resolution in medical image scanners and the need of faster reconstruction methods,rnnew ways of exploiting the inherent parallelism of reconstruction algorithms have arisen. In this paper, wernpresent Mangoose++, an application to perform X-ray computed tomography that supports multiple gradesrnof parallelism. This parallelism is tackled with two different approaches: the usage of parallel nodes withrnmulticore CPUs in a cloud environment and the usage of high-performance computing (HPC)-based parallelrnarchitectures such as general-purpose computing on graphics processing unit (GPGPU) or Intel Xeon Phi.rnIn this paper, we show the design and implementation of the application in three types of platforms relatedrnto the previous mentioned approaches, comparing and analyzing the performance, resource utilization, andrnscalability of each platform. Accelerators offer high performance for data sizes that fit inside the acceleratorrnmemory. This is the main advantage of Intel Xeon Phi that, in this work, obtains similar performance resultsrnthan compute unified device architecture (CUDA)-based GPGPU versions, comparing with cards with lessrnmemory capacity. In our evaluation experiments, we additionally analyze and discuss the costs and efficiencyrnof Mangoose++ over Amazon Compute Cloud platform, demonstrating that lower times can be achieved inrna reasonable price compared with owned HPC-based hardware. A comparison between distinct hardwarernconfigurations is provided for emphasizing on the advantages and disadvantages of each one.