CLA

The MMACF32 instruction must perform a data move in parallel with the multiply-accumulate.  You're missing an instruction after the MMACF32 which does this.  Take a look at the MMACF32 instruction description in the CLA User's Guide (spruge6a, p.87).  There is always a "||MMOV32 in the next line.  

I think you can do this in the same inline assembly, something like this:

__asm(" MMACF32 MR3, MR2, MR2, MR0, MR1 || MMOV32 MRx *MARy");

If not, try separating the data move in another inline assembly, but without the leading whitespace:

__asm("|| MMOV32 MRx var1");

I hope this helps.

Regards,

Richard

Hi Richard ,

Thanks for the response. Im still having problems, I had orignally copied that piece of code from the user guide:

__asm(" MMACF32 MR3, MR2, MR2, MR0, MR1");
__asm("|| MMOV32 MR0, *MAR0");
__asm(" MMOV32 MR1, *MAR1");

Thanks