TMS320F280049C: TMU introduces many NOP instruction, it causes code size become larger.

Part Number: TMS320F280049C

Hi Experts,

I am asking for my customer here.

They use TMU Optimize code in their control code, they find when the TMU instruction is called continuously, the compiler inserts a NOP instruction as below:

#define     TMUDIV(NUM, DENUM) (__divf32((FLOAT32)(NUM),(FLOAT32)(DENUM)))

#define     TMUINV(SRC) (TMUDIV(1.0,(SRC)))

#define     TMUSQRT(SRC) (__sqrt((FLOAT32)(SRC)))

#define     TMUINVSQRTTABLE(IND) ((INT16)(TMUINV(TMUSQRT(TMUDIV((IND),256.0))*2048.0)))

i16tmpAx_0 = TMUINVSQRTTABLE(objInverter.m_i16VinMFromCLA_0);

Complier Version：TI v21.12.0.STS

--tmu_support=tmu0

--float_support=fpu32

--opt_level=4

--opt_for_speed=2

--fp_mode=relaxed

You can see the above NOP  instruction, it causes code size become larger. So do you have any suggestion or Guide? Thanks.