Intel® C++ Compiler
Support and discussions for creating C++ code that runs on platforms based on Intel® processors.
7650 Discussions

Exponent range in the Intel Decimal Floating-Point Math Library

cppcoder
Beginner
165 Views
I'm sorry if this is not the right forum to ask, but couldn't find a more suitable one. I was wondering what are the exponent ranges (maximum and minimum exponents) supported by theIntel Decimal Floating-Point Math Library

Thanks in advance
0 Kudos
3 Replies
Om_S_Intel
Employee
165 Views
The floating point exponent is 8 bit long for ia-32. You may learn more about this by going throgh-fp-model compiler option.

timintel
Beginner
165 Views
Om, how would fp-model be relevant to the decimal library?
Are you saying that the decimal library uses an 8-bit binary exponent for all 3 decimal floating point types?
tydeman
Beginner
165 Views
If you go to the C standard web site:
http://www.open-std.org/jtc1/sc22/wg14/www/documents
and look at the paper:
N1107 07-Mar-2005 Kwok, Decimal-floating point TR document rev 5
you will find a table of the minimum and maximum exponents for
each of the three decimal floating-point types. There also are macros
for the minimum and maximum exponent:

  • minimum exponent
  • DEC32_MIN_EXP -95
    DEC64_MIN_EXP -383
    DEC128_MIN_EXP -6143

  • maximum exponent
  • DEC32_MAX_EXP 96
    DEC64_MAX_EXP 384
    DEC128_MAX_EXP 6144

Reply