I thought I would share a documentation mistake in the v9.0 FFT IP core. This relates in particular to the bit-accurate C Model for the FFT core.
Xilinx provides this command for compiling the example under GCC:
gcc -x c++ -I. -L. -lIp_xfft_v9_0_bitacc_cmodel -Wl,-rpath,. -o run_bitacc_cmodel run_bitacc_cmodel.c
However, this will fail as the linked libraries should be placed at the end of the gcc command as shown below:
g++ -I. -L. -Wl,-rpath,. -o run_bitacc_cmodel run_bitacc_cmodel.c -lIp_xfft_v9_0_bitacc_cmodel
Additionally it is better to compile under g++ rather than gcc however the previous will still work.
Thanks for sharing this, we can compile the C-model using the command in the user guide, do you set the LD_LIBRARY_PATH environment variable?