DSP_ifft32x32

-The FFT coefficients (twiddle factors) are generated using the program tw_fft32x32 provided in the directory ‘support\fft’. The scale factor must be 2147483647.5. No scaling is done with the function; thus the input data must be scaled by 2log2(nx) to completely prevent overflow.

Implementation Notes

 

- Bank Conflicts: No bank conflicts occur.

 

- Interruptibility: The code is interruptible.

 

- The routine uses log4(nx) − 1 stages of radix-4 transform and performs

 

either a radix-2 or radix-4 transform on the last stage depending on nx. If

 

nx is a power of 4, then this last stage is also a radix-4 transform, otherwise

 

it is a radix-2 transform.

 

- See the fft16x16t implementation notes, as similar ideas are used.

Benchmarks

Cycles

(13*nx/8 + 28) * ceil(log4(nx) − 1) + 6 * nx/4 + 39

 

Codesize

960 bytes

C64x+ DSPLIB Reference

4-37

Page 65
Image 65
Texas Instruments TMS320C64X manual DSPifft32x32