IEEE 754 or 854 FP representation standard 
Author Message
 IEEE 754 or 854 FP representation standard

Hello netters,
I'm looking for information on IEEE 754 or IEEE 854 binary floating point
representation. I won't need the whole standard (that's why I don't want
to buy one); I just need to know how bits are allocated for sign, mantissa,
and exponent.
Any hints or pointers are appreciated.

Sergei



Mon, 01 Dec 1997 03:00:00 GMT  
 IEEE 754 or 854 FP representation standard

Quote:

>I'm looking for information on IEEE 754 or IEEE 854 binary floating point
>representation. I won't need the whole standard (that's why I don't want
>to buy one); I just need to know how bits are allocated for sign, mantissa,
>and exponent.

Well, since Intel (I think) claims that the FPU of the 80486 is 754
compliant, why not just look at one of their data sheets?  I suspect
that there are other FPU chips made by others that might have the info
you need in their data sheets.

 ...

--
Fritz Whittington Texas Instruments, P.O. Box 655474, MS 446 Dallas, TX 75265
    Shipping address: 13510 North Central Expressway, MS 446 Dallas, TX 75243  

Since I am not an official TI spokesperson, these opinions contain no spokes.



Mon, 01 Dec 1997 03:00:00 GMT  
 IEEE 754 or 854 FP representation standard

    Sergei> Hello netters, I'm looking for information on IEEE 754 or IEEE 854
    Sergei> binary floating point representation. I won't need the whole
    Sergei> standard (that's why I don't want to buy one); I just need to know
    Sergei> how bits are allocated for sign, mantissa, and exponent.  Any
    Sergei> hints or pointers are appreciated.

Well, since 854 is the radix-independent standard I assume that you mean the
754 standard.  My copy of 754 (-1985) shows:

                           Exponent        
Format           Exponent    Bias       Mantissa  Sign
------           --------  --------     --------  ----
Single               8       +127          23       1
Extended-single    >10     unspecified    >31       1
Double              11      +1023          52       1
Extended-double    >14     unspecified    >63       1

Note that the mantissas are normalized and that the leading 1 is suppressed
giving an effective mantissa size one larger than the actual size.  (This need
not be true in the case of extended formats).

An implementation must support Single and must support the Extended format
corresponding to the largest base format supported.  Bit order is not defined.

John



Tue, 02 Dec 1997 03:00:00 GMT  
 
 [ 3 post ] 

 Relevant Pages 

1. IEEE 754 or 854 FP representation standard

2. IEEE 754 float representation of a number

3. PEP 754: IEEE 754 Floating Point Special Values

4. Ada/IEEE binding (IEEE 754)

5. need IEEE 754 <-> F/PC

6. Some Q&A with Kahan on IEEE 754 (long)

7. FPU IEEE-754 calculation

8. IEEE 754 Floating point VHDL functions / MATH package

9. IEEE 754

10. IEEE 754 format

11. IEEE Std. 754-1985

12. IEEE Std.754-1985

 

 
Powered by phpBB® Forum Software