Compiler/OMAPL138B-EP: Why DSP(TMS320C6748) of OMAPL138 is taking more clocks than TI simulator TMS320C6748 for same intrinsic functions?

Part Number:OMAPL138B-EP

Tool/software: TI C/C++ Compiler

Hi

We had developed Melpe Vocoder and simulated with TI simulator for TMS320C6748, When the same Melpe Vocoder code is ported on DSP(TMS320C6748) portion of OMAPL138, it is taking 6 times more clocks than TI simulator for TMS320C6748 . All the intrinsic functions taking too much clocks in OMAPL138 compared to TMS320C6748.

For example:

TI simulator  TMS320C6748:

c = _sadd2( a, b); // It is taking 7 clocks

LCDKOMAPL138 Kit  on TMS320C6748:

c = _sadd2( a, b); // It is taking 18 clocks

Pls help in this regards