On Thu, 18 Apr 2002 at 11:48 +0400, Dmitry <[email protected]> wrote:

> NO... TI fp implementation is not IEEE 754 complaint... TI uses 40 bits 
> floats while gcc uses 32 bits IEEE 754 fp concepts.

I just am looking at the manual for the free demo (IAR) compiler
for the msp430, and it is documenting their floating point format
as 4 bytes

31 - sign
30 - 23 exponent
22 - 0 Mantissa

(-1)**s * s**(exponent-127) * 1.Mantissa

It states that the values are stored in IEEE format




Reply via email to