I am fairly new to VisualDSP++ and Blackfin processors.
VisualDSP++ provides many optimisation #pragmas for 'for' loops, such as 'vector_for'.
A lot of the code I have iterates over a range between two pointers using a while loop.
Would it be advantageous to convert these loops to indexed 'for' loops in order to use these #pragmas?
Or is VisualDSP++ likely to be smart enough to deduce the best optimisation?
while (pBegin != pEnd)
*pDestination++ = *pBegin++;
size_t count = pEnd - pBegin;;
for (size_t i = 0; i < count; ++i)
pDestination[i] = pBegin[i];