ERROR: floating constant is out of range

I have created a algorithm in Simulink. I tried to implement this model to TI TMS320F28379D LAUNCHXL board embedded target. When I try to build the model, I receive the following error messages.

This is one of the code segments that caused error

[Walter Roberson]

I wonder where that constant comes from; too bad it isn't documented. It happens to be 0x1cc0000000000000 but that does not tell me anything about what it represents.

Ah... with further checking I see that it is

format long g
eps(sqrt(realmin))

ans = 
     3.31216864211124e-170

and that suggests its replacement value should be

eps(sqrt(realmin('single')))

ans = single
     1.29247e-26

Answer 1

[Mike Hosea]

That number is well within the range of IEEE double precision floating point numbers, but the TI TMS320F28379D has native 32bit IEEE single precision floating point. I'm not sure what the C compiler for it does with double precision data, but based on what you have shown, it doesn't seem to support IEEE double precision floating point, even in emulation. Try using single precision.

[Sahil Talathi]

How can we be able to change double precision to single precision ?

[Mike Hosea]

You could start here.