The
FLOATINMATH option specifies that the precision that the compiler
should use when invoking the mathematical built-in functions.

.-ASIS-----.
>>-FLOATINMATH--(--+-LONG-----+--)-----------------------------><
'-EXTENDED-'
- ASIS
- Arguments to the mathematical built-in functions will not
be forced to have long or extended floating-point precision.
- LONG
- Any argument to a mathematical built-in function with short
floating-point precision will be converted to the maximum long floating-point
precision to yield a result with the same maximum long floating-point
precision.
- EXTENDED
- Any argument to a mathematical built-in function with short
or long floating-point precision will be converted to the maximum
extended floating-point precision to yield a result with the same
maximum extended floating-point precision.
A FLOAT DEC expression with precision p has short floating-point
precision if p <= 6, long floating-point precision
if 6 < p <= 16 and extended floating-point
precision if p > 16.
A FLOAT BIN expression with precision p has short floating-point
precision if p <= 21, long floating-point precision
if 21 <p <= 53 and extended floating-point
precision if p > 53.
The maximum extended floating-point precision depends on the
platform.
|
This information center is powered by Eclipse technology. (http://www.eclipse.org)