Tropic Networks Inc.

Leonard Marziliano

Ping Wai Wan

Dongxing Jin

Derrick Remedios

A method and apparatus for performing a fast Fourier transform (FFT) computation on large data sizes in real time is provided. The speed at which a FFT is performed is increased by reducing the number of times a Direct Memory Access (DMA) unit must transfer data between an internal memory and an external memory. This is achieved through an algorithm in which data is imported from the external memory into the internal memory and having the CPU perform several calculations on the imported data before it is exported back to the external memory. The imported data in the internal memory has a structure that results in a reduction of the number of times the imported data is swapped between different layers of the internal memory during the FFT computation. Furthermore, the DMA unit import/exports data between the internal memory and the external memory while at the same time having the CPU perform calculations on other data in the internal memory.

Real-time method and apparatus for performing a large size fast fourier transform

Various methods and apparatuses are provided for performing a radix-M FFT (Fast Fourier Transform) upon N time domain samples to produce N/S frequency domain samples for detecting tones of dithers impressed on channels of a WDM (wavelength Division Multiplexed) optical signal. Successive tones have a tone frequency spacing, &Dgr;f , and a sampling frequency, f , is chosen so that f &equals;N&Dgr;f /S. S is a spacing given by S&equals;M with w being an integer. The radix-M FFT is performed in k&equals;log (N) stages and within the stages a reduced number of radix-M computations, when compared to the number of radix-M computations of a conventional radix-M FFT, are performed on data points associated with the N time domain samples. This is possible because successive frequency domain samples of the N/S frequency domain samples differ by &Dgr;f &equals;S&Dgr;f where &Dgr;f is a frequency bandwidth.

Growing community of inventors

Hull, Canada

Leonard Marziliano