Hi Bruno, Thank you very much!!!

I am using Raisonance Kit 6.1, Version 06.10.15, Build number : BN732-XE8000.

Thanks again!!!

Hi

In this case the compiler you use is the GNU GCC for Coolrisc.
I have checked in the actual version of our software and we do not force
any special option for floating handling.
We just pass to the compiler the define of teh target and so the default value
for this target will be apply.

To have more information about flaot representation you will need to refer to the GCC documentation.
The fastest way to find the representation used is to check in \GNU\c816-gcc\c816\include\float.h
You should find there a description of the different paramater.

Regards,
Matloub

Hi Matloub,

Thanks a lot!

I will check these files as you mentioned above, pls. tell if you have any other suggestion or example!

Best Regards!

Dears,

I found an example in other place, I tried but it does not work good.

Could you give me some suggestions?

Thanks!!!

//convert IEEE-Xemics to IEEE_PC
float convertToIEEEPC(unsigned char *data) {
float res;
unsigned char *pK;

pK=(unsigned char*)&res;

*pK=*(data+3);
*(pK+1)=*(data+2);
*(pK+2)=*(data+1) & 0x7f;
*(pK+3)=((*data) >> 1 )- 1;
if(*data & 0x01) *(pK+2) |=0x80;
if(*(data+1) & 0x80) *(pK+3) |=0x80;
return res;
}

//convert IEEE-PC to IEEE_Xemics

convertToIEEEDev(unsigned char *dest,float ff) {
unsigned char *f;
f=(unsigned char*)&ff;

*dest=((*(f+3))<<1)+2;
*(dest+1)=(*(f+2)) & 0x7f;
*(dest+2)=*(f+1);
*(dest+3)=*f;

if(*(f+3) & 0x80)
*(dest+1) |= 0x80;

if(*(f+2) & 0x80)
*dest |=1;
}

// :-))))

I am so sorry!

The abouve example is right, I made some mistakes.

Thanks for the author of example!

Thanks for others!!!

Hi Kevin,

Alternatively you could use the ldexp() and frexp() functions from math.h, which are standard C. They are intended for float conversion (separating the mantissa from the exponent).

Regards,
Bruno