Tool/software: TI C/C++ Compiler
I have this loop that I am trying to optimize:
for (j=0; j<L_SUBFR; j++){ pG729ADecoder->Postfilt.scal_res2[j] = pG729ADecoder->Postfilt.res2[j]>>2; }
where L_SUBFR is 40. Here is the pipeline information:
If I try the restrict keyword on scal_res2 and res2, this is what I get with no decrease in cycle count:
I have also tried the MUST_ITERATE and UNROLL pragmas with different unroll factors, but the cycle count only increases.
Is there a way to optimize this loop? I would appreciate any suggestion!