I am not getting the documented pipelined performance for atan2 (or other functions like sqrt). (SPRUGG9)
This is what I tried --
for (i=0; col <SizeX; i++)
{
phiRadQ15[i] = _IQ15atan2(xRFxd[i],yRFxd[i]);
}
I get 118 cycles per point which is the non pipelined performace. (processor used: c6455)
I tried #pragma MUST_ITERATE. No difference in cycle count.
Can someone please tell me what I need to do to get the pipelined performace?
Thanks,
SBanik