why fft core's output is conjugate

[Altera_Forum]

Dear all, 

When I using quartu12.0's fft megacore IP, I found such problems:  chose variable streaming and natural order for both input and output order, set ifft, make sink real = impulse, sink_imag =0, 1024 point, the ifft output from the simulation shows that the (source_real + 1j*soure_imag) is the conjugate of the ideal result from matlab. Please help me. Thank you very much.  And the result is caputured as follows:  https://www.alteraforum.com/forum/attachment.php?attachmentid=6874

[Altera_Forum]

It could be your understanding of ofrequency sense is wrong between matlab model and ifft core.  

You got cos waveform on real ifft output and -sin waveform on imag ifft output. This means your so called impulse is one nonzero sample at a negative frequency location i.e. after midpoint of 1024 (well further from sample 512) .  So in which sample did you put that impulse?