Abstract
Back-Projection is the major algorithm in Computed Tomography to reconstruct images from a set of recorded projections. It is used for both fast analytical methods and high-quality iterative techniques. X-ray imaging facilities rely on Back-Projection to reconstruct internal structures in material samples and living organisms with high spatial and temporal resolution. Fast image reconstruction is also essential to track and control processes under study in real-time. In this article, we present efficient implementations of the Back-Projection algorithm for parallel hardware. We survey a range of parallel architectures presented by the major hardware vendors during the last 10 years. Similarities and differences between these architectures are analyzed and we highlight how specific features can be used to enhance the reconstruction performance. In particular, we build a performance model to find hardware hotspots and propose several optimizations to balance the load between texture engine, computational and special function units, as well as different types of memory maximizing the utilization of all GPU subsystems in parallel. We further show that targeting architecture-specific features allows one to boost the performance 2–7 times compared to the current state-of-the-art algorithms used in standard reconstructions codes. The suggested load-balancing approach is not limited to the back-projection but can be used as a general optimization strategy for implementing parallel algorithms.
Highlights
X-ray tomography is a powerful tool to investigate materials and small animals at the micro- and nano-scale [1]
Our results show that all NVIDIA GPUs starting with Fermi benefit from the 64-bit texture fetches if requests are properly localized
The type-conversions are executed at a half rate of the peak floating-point performance on AMD GCN GPUs, but only a single type-conversion instruction can be executed per 12 floating-point operations on NVIDIA Kepler GPUs
Summary
X-ray tomography is a powerful tool to investigate materials and small animals at the micro- and nano-scale [1]. A recent study suggests to implement back projection as convolution in log-polar coordinates in order to gain high reconstruction speed with interpolation in the image domain [23] This new method has not yet been adopted in production environments. Multiple papers perform a general analysis of a range of GPU architectures, reveal undisclosed details trough micro-benchmarking, and propose guidelines for performance optimization [27,28,29] This information is invaluable to understand factors limiting performance on a specific architecture and to find an alternative approach to achieve a better performance. In [31], we presented two highly-optimized back-projection algorithms for NVIDIA Pascal GPUs and a hybrid approach to balance the load between different GPU subsystems using both in parallel.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
