Research on FFT Algorithms for Peta-Scale Computing Environment

在这项研究中，我们研究了一种并行快速傅立叶变换（FFT）算法的千万亿次计算环境的实现。由于最佳的缓存块大小可能取决于问题的大小，我们提出了一种方法来确定最佳的缓存块大小，最大限度地减少缓存未命中的数量。此外，并行FFT需要密集的全对全通信，这影响了FFT的性能。还实现了全对全通信的自动调谐。性能测试结果表明，所提出的并行FFT的自动性能调整的实现是有效的，以提高性能。

In this research, we investigated an implementation of parallel fast Fourier transform (FFT) algorithm for peta-scale computing environment. Since the optimal cache block size may depend on the problem size, we proposed a method to determine the optimal cache block size that minimizes the number of cache misses. In addition, parallel FFTs require intensive all-to-all communication, which affects the performance of FFTs. An automatic tuning of all-to-all communication was also implemented. The performance results demonstrate that the proposed implementation of parallel FFTs with automatic performance tuning is efficient for improving the performance.