Image reconstruction is observed to be one of the most common problem because of it's large data movement and non-trivial data dependencies. In the past, these problems were tackled by many high performance hardware such as FPGA's and GPGPU's. This also reflects the investemts to be made in these supercomputers for real time reconstruction of clinical computed tomography (CT) applications. Medical imaging systems are employing high performance computing (HPC) technology to meet their time constraints. This paper presents different optimizations to the volume reconstruction and implement it on a commodity hardware such as x86 based multicore system. This paper chooses to perform its implementaion on Intel Xeon X5365 multicore processor. We perform different levels of parallelization and analyse each of them and report their results with respect to serial implementation. The objective of this paper is to understand the constraints of volume reconstruction in multicore architecture and optimize them while preserving the quality of the reconstructed image.
The Computer Journal, published by Oxford University Press, is one of the longest-established journals serving all branches of the academic computer science community.
For more information about the journal see http://comjnl.oxfordjournals.org/.
This paper will be looking at the development of random Fibonacci sequences throughout history and investigating the various mathematical methods used by many mathematicians to determine important qualities about the sequence, which all lead to the growth rate.