Texas Instruments TMS320C64X Programmer's Reference Manual page 89

Implementation Notes
Benchmarks
-
Bank Conflicts: No bank conflicts occur if the input arrays x[ ] and y[ ] are
offset by 4 half-words (8 bytes).
-
Interruptibility: The code is fully interruptible.
-
The code is unrolled 4 times to enable full memory and multiplier
bandwidth to be utilized.
-
Interrupts are masked by branch delay slots only.
-
Prolog collapsing has been performed to reduce codesize.
Cycles nx / 4 + 14
Codesize 64 bytes
DSP_dotprod
C64x+ DSPLIB Reference 4-61