Abstract #5005

Order of Magnitude Speedup for Iterative Algorithms in Quantitative Susceptibility Mapping Using Multi-Core Parallel Programming

Deqi Cui^1,2, Tian Liu¹, Pascal Spincemaille³, Yi Wang

¹Biomedical Engineering, Cornell University, New York, NY, United States; ²Optic electronics, Beijing Institute of Technology, Beijing, China; ³Radiology, Weill Cornell Medical College, New York, NY, United States

Iterative susceptibility reconstructions requires four Fast Fourier Transforms in every iteration and hundreds of iterations to converge. Although the iteration is hard to disperse, FFT can be computed in parallel. In this study, a parallelized FFT based on OpenMP was implemented to achieve quasi real-time susceptibility map reconstructions.