Re: compilation problem with module function interface definition

James Giles wrote:
glen herrmannsfeldt wrote: 

(snip)

Maybe in a few years decimal floating point will get popular, and
this problem will go away.  That is, 1.23 will be exactly 1.23,
single or double precision.  There are people working on an
efficient float decimal system.


But then what will be the rule for mixed mode between binary
floating point and decimal floating point?  Not that it's a bad idea,
but it's something that will have to be addressed.


fixed/float conversion can be pretty slow on many systems now.

The suggested format is fairly easy to convert, but most likely
there will be hardware assist for the conversion.  Note that
x87 includes conversions between fixed decimal and its internal
floating point format, even from the 8087 days.  It can't be
so hard to do.

Storing three digits in 10 bits, with additional BCD digits when
the number of digits is not a multiple of three is only slightly
less efficient than pure binary.  For a given range, fewer exponent
bits are needed, which pretty much makes up for the lost bits due
to decimal normalization.

-- glen


.

Relevant Pages

  • Re: printf "%.100f" the value 2 ^ (-127)
    ... >conversion function that's doing the rounding. ... Floating point numbers are ... number actually contains hundreds of decimal digits. ...
    (microsoft.public.vc.language)
  • Re: Interesting math
    ... Floating point number represents a real number with 6 digits precision. ... Floating point numbers are denoted by the keyword float. ...
    (alt.usage.english)
  • Re: Gentler Decimal Floating-Point
    ... effectively change the radix of a decimal floating point ... under the heading "Logarithms all the time", ... logarithmic representation of numbers to be made workable for all the ... The middle digits are shifted only when the exponent, ...
    (comp.arch.arithmetic)
  • Re: Linear Algebra Challenge
    ... Since I'm using floating point, so I'll never be able to calculate one ... \<< aeq vals SUBST EVAL vals AXL + AXL 'vals' STO pert vals SUBST ... I switched over to Maxima to avoid the memory problem, ... floating point math set to 99 digits. ...
    (comp.sys.hp48)
  • Re: Floating point subtraction rounding error (NOT display error)
    ... and apparently it's because floating point subtraction is having ... Notice that, to align both numbers at their decimal point, we had to discard the two less significant binary digits of 0.6. ... Note that this is the best result you can get with a limited precision of 53 bits, and it's not even Python who computes the value, but the underlying C math library. ...
    (comp.lang.python)