Block-by-Block Configurable Fast Fourier Transform Implementation on AI Engine (XAPP1356)

Document ID
Release Date
1.0 English

The Fast Fourier Transform (FFT) has wide applications in signal processing algorithms that often require high throughput with configurable FFT sizes on a block-by-block basis. It is shown that ten AI Engines packed in a 5x2 array deliver 3.7 GSPS throughput. The multi-AI Engine FFT architecture is highly scalable and can be extended to even larger FFT sizes for many applications including, but not limited to, 5G wireless.