TDA2PXEVM: TDA2PXEVM

Thanks Dave.

I have one doubt regarding #pragma MUST_ITERATE.

My code contains 4 number of "nested for" loop running sequentially.

for (row...)
{
for (col)
{

//some operations

}

}

The loop iterates 450X1800 times.

The loop takes 114ms. If i use this pragma will there be any improvement in execution time?

Soumya,

The MUST_ITERATE pragma give the minimum number of iterations through the loop. So the number of nested loops doesn't necessarily apply there but if the outer loop is iterating 450x1800 times you should be able to apply it.

Note that you can look to additionally use _nassert() on the loop counters for the nested and outer loops, and also on your address pointers to give hints to the compiler for ways to unroll and parallelize.

Another optimization technique is to merge two (nested) loops into one.

Best regards,

Dave