Motorola chips and floating point registers

[Martin_BF]

Hi,

I am just getting into the freescale world and am enjoying learning about the HC908GP32 chip. The CWX IDE is great and I am using the full chip simulation currently.

Are there any of the 8bit micro family that would encorporate a floating point register(s) ?

I ask because when I try and get some math based processing, it would seem I will blow away all code size limits really quickly using the 908GP32. This is only a subset of the programming I was hoping to do.

If the 8bit micro world is not compatible, which micro family should I be looking at ?

Cheers

Martin

[mke_et]

I haven't tried doing any floating point on an 08 part, but I use it on the 12 all the time. Motorola used to have a floating point math package on their web site for the 6811 in assembly. It wasn't hard to convert that into the assembler in CodeWarrior to run on the 12.

Mike

[Martin_BF]

Hi Rocco,

Thanks for the response. I had forgotten that option.

I will give it a go and so you may hear from me later.

Martin

[rocco]

Hi, Martin:

In my humble (but often overbearing) opinion, The 8 and 16 bit micros are not suited for floating point. But I feel floating point is not suited for almost all embedded applications. You rarely need the dynamic range of floating-point.

I use the GP32 a lot with 16, 24 and 32 bit fixed-point math. My experience is that there are few embedded applications where the decimal point really has to float. I have to do some awfully precise FFTs in a DSP56303, and it's all fixed-point, down to eight decimal places. Even the trig.

I find that the integer math capabilities of the HC08s are pretty darn efficient for doing fixed-point math.