How can I use floating point math on a floating point F28335 target?

Altair Forum User

Submitted by Joe on Thu, 06/13/2013 - 07:36

 

 

I understand that the Delfino F28335 has a floating point unit inside. In the VisSim menu, under 'blocks', I see fixed point math blocks. Do I have to use those fixed-point blocks for an embedded target? If not, how can I use floating point math with the 28335 as the target?

Altair Forum User

Submitted by Anders89 on Thu, 06/13/2013 - 07:59.

 

Aside from the fixed-point blocks, VisSim blocks can be floating-point, matrix, string or integer. You can use all blocks on either a fixed point MCU or a floating point MCU. It's just that floating point blocks will run pretty slowly on a fixed-point MCU. If you turn on menu item View > Data Types, the block connectors will be colored by data type:

Type	Color
floating-point	red
matrix	purple
integer	green
string	blue
fixed-point/scaled-integer	yellow

 

So if you see any red or purple in your diagram, you can rest easy knowing you are using the FPU on your MCU.

Altair Forum User

Submitted by rt54321 on Thu, 06/13/2013 - 08:40.

Thank you for your fast response!

--> Does this mean that, for example, if I use the multiplier block under 'Blocks -> Arithmetic -> *', that this will automatically be a floating point multiplier, if I have floating point inputs to it?

Altair Forum User

Submitted by Anders89 on Thu, 06/13/2013 - 09:03.

If you use the Blocks -> Arithmetic -> multiplier block, it will generate a floating point multiply regardless of the input type. Even if you use fixed-point inputs, they will be converted to floating point, then multiplied in floating point.