FCompare
Summary:
This function compares two floating point inputs, A and B, to form a set
of comparision output bits. IEEE standard floating point is implemented. However,
you have the option to create new floating point formats and to select rounding
and status options that go beyond the standard. Certain IEEE specified exceptions
are not handled in hardware and are left to software exception handlers.
The particular floating point format of the input signals is
programmable via MantWidth and ExpWidth. MantWidth sets the width of the
mantissa or fractional part of the floating point value. ExpWidth sets the
width of the exponent part of the value. The floating point value has a total width
of ExpWidth+MantWidth+1 bits because there is also a sign bit.
Some implementation architecture variations can be controlled via the
Arch parameter.
See the Floating Point Library User Guide for more details on floating point
representation, rounding options, status bits and architecture options.
Function:
AGTB = A>B
AEQB = A==B
ALTB = A**B.
AEQB: 1-bit output indicates A==B.
ALTB: 1-bit output indicates A**