Internal Representation Of Numerical Types

From RAD Studio
Jump to: navigation, search

Go Up to Constants Overview Index

Integers

Integer values have the following internal representation in the architecture of the Intel Pentium family of microprocessors:

32-bit


Floating-point types

Floating-point values have the following internal representation in the architecture of the Intel Pentium family of microprocessors:

Floating point


Here we use the following terms:

s

Sign bit (0 = positive, 1 = negative)

i

Position of implicit binary point

1

Integer bit of significance. It is stored explicitly in long double and it is used implicitly in float and double

Biased exponent
(normalized values)

127 (7F) for float
1,023 (3FF) for double
16,383 (3FFF) for long double

See Also