Part Number: TMS320F28379D
Hello,
I have a question regarding constants in the CLA programming.
I often use operations like this:
arg = __mmaxf32(arg, 0.1e-3);
To load the constant in one of the MRx registers two operations (0.1e-3 --> 0x38b1b717) are needed. Very often a lower precise would be enough. In the current case 0x38b10000 --> 9.9659e-05 would also be OK and the constant can be load in one step.
Is it possible to truncate the float constants to 16Bit values?
Thank you very much
Ralf
