Other Parts Discussed in Thread: AWR2243, MMWCAS-RF-EVM, MMWCAS-DSP-EVM
Hi Experts,
I'm working with the cascade_radar_object_detect example from PROCESSOR_SDK_RADAR_03_08_00_00 and am using the AWR2243 MMWCAS-RF-EVM along with the companion MMWCAS-DSP-EVM(TDA2x). I have a few questions regarding FFT implementations in this use case:
1. In this use case, how is FFT carried out? Is it through creating an FFT object in radarfft2.c and then submitting the instance for execution to the EVE kernel?
(PROCESSOR_SDK_RADAR_03_08_00_00\vision_sdk\apps\src\rtos\radar\src\alg_plugins\alg_fxns\radarfft2)
Is vcop_fft_256_16ix16o_kernel.k the actual execution function?
(PROCESSOR_SDK_RADAR_03_08_00_00\ti_components\algorithms\eve_sw_01_20_01_00\kernels\radarlib\vcop_fft_npt_16ix16o\src_kernelC\vcop_fft_256_16ix16o_kernel.k)
2. What is the function of the fft_alg.c code found in this path?
(PROCESSOR_SDK_RADAR_03_08_00_00\ti_components\algorithms\eve_sw_01_20_01_00\apps\fft\algo\src\fft_alg.c)
Is it called during FFT execution in this use case? I've invoked Vps_printf function in it but nothing is displayed.
3. I noticed that numLines is scaled up in fft_alg.c. Is this in line with what have been mentioned in the following thread that says, "For example we have 16kb of IBHF memory so we will try to get as many lines as we can to exhaust complete 16 kb"? Is it possible to just process one line instead?
PROCESSOR-SDK-TDAX: Different output by same input values at TDAx2P EVE core and FFT example of TI
Thank you for your time and expertise.
Liam
