About the Author

Next-generation FFT Algorithms in Theory and Practice at CSE23: Parallel Implementations and Applications

By Samar Aseeri

During the 2023 SIAM Conference on Computational Science and Engineering, which recently took place in Amsterdam, the Netherlands, an international minisymposium session explored the latest advancements in fast Fourier transform (FFT) algorithms and their implementations. The event was an excellent opportunity for professionals to exchange ideas and discuss new possibilities for ongoing research, practical applications, and future directions.

Daisuke Takahashi (University of Tsukuba in Japan), Franz Franchetti (Carnegie Mellon University), and Samar A. Aseeri (King Abdullah University of Science and Technology (KAUST) in Saudi Arabia)—all of whom are experts in the area of FFT—organized the engaging session. The minisymposium included the following four presentations:

In the first presentation, Takahashi described a six-step FFT algorithm to parallelize a parallel number theoretic transform on manycore clusters and commented on the performance results. Het Mankad (Carnegie Mellon University)’s talk then overviewed the goal of extending and modernizing FFTW for the exascale era—in the context of FFTX—while maintaining backwards compatibility. She also spoke about NNTX, a variant of FFTX that focuses on the number theoretic transform. Aseeri delivered the third presentation and addressed the way in which profiles of a FFT-based solver for the Klein-Gordon equation can help explain the better scaling of FFTE when compared to 2DECOMP&FFT. Finally, Pierre Balty’s (Université catholique de Louvain in Belgium) concluding talk introduced a Fourier-based Library of Unbounded Poisson Solvers (FLUPS) for two- and three-dimensional distributed uniform grids that is tailored for massively parallel architectures.

Overall, this minisymposium was a great success and provided attendees with a comprehensive understanding of FFT algorithms and implementations. The session opened up new possibilities for both research and practical applications and created a platform for applied mathematicians and computational scientists to share their experiences and knowledge.

Samar Aseeri is a computational scientist at King Abdullah University of Science and Technology (KAUST) in Saudi Arabia. She earned her undergraduate, graduate, and doctoral degrees in applied mathematics from Umm Al-Qura University in Saudi Arabia, and completed supercomputing training at IBM in New York. Aseeri is currently leading two initiatives to establish high-performance computing communities: Benchmarking in the Data Center and FFT in the Exascale Era.