Parallel butterfly structure for multiple point (rank-at-a-time) processing performance.
The Parallel Butterfly FFT IP Core provides efficient continuous data FFT calculations at the rate of multiple points per clock cycle. A bank of parallel butterflies executes the FFT one rank at a time on multiple I/O streams.
Parallel Butterfly FFT
Overview
Key Features
- ParaCore Architect parametric-based core provides maximum adaptability and flexibility (see details on the FFT Parameters)
- Completely proven in many real-world applications
- Supports any radix-2 length FFT and IFFT transformations
- Variable length option for runtime per-transform length select
- Fixed- or floating-point complex I/O of any width
- Integrated Hanning Window (user defined)
- Optional magnitude output
- Designed for continuous transformations with extremely fast conversion times
- Naturally ordered I/O streams
- Parallel I/O paths for maximum throughput
- Massively parallel butterfly architecture, from 1 butterfly up to the full logic capacity of the target device.
- Available in generic HDL or targeted EDIF formats
- Delivered as a completely self contained module with a full testbench