GMS30C2116 Datasheet, PDF (112/322 Pages) Hynix Semiconductor – USERS MANUAL

3-48
CHAPTER 3
3.33.2 Floating-Point Instructions (continued)
Invalid Operation Exception (V)
The Invalid Operation exception is signaled if one or both of the operand are invalid for an
implemented operations. The MIPS ISA defines the result, when the exception occurs
without a trap, as a quiet Not a Number (NaN). The invalid operations are:
l Addition or subtraction: magnitude subtraction of infinities, such as: ( +inf ) + ( -
inf ) or ( - inf ) - ( -inf )
l Mutiplication: 0 times +inf, with any signs.
l Division: 0/0, or inf/inf, with any signs.
l Conversion of a floating-point number to a fixed-point format when an overflow, or
operand value of infinity or NaN, precludes a faithful representation in that format.
l Comparison of predicates involving < or > without ?, when the operands are
unordered.
l Any arithmetic operation on a signaling NaN. A move (MOV) operation is not
considered to be an arithmetic operation, but absolute value (ABS) and negate
(NEG) are considered to be arithmetic operations and cause this exception if one or
both operands is a signaling NaN.
l Square root: sqrt(x), where x is less than zero.
Floating-point Trap Enabled Results: The original operand values are undistrurbed.
Floating-point Trap Disabled Results: The FPU always signals an Unimplemented
exception because it does not create the NaN that the IEEE standard specifies should be
returned these circumstances.
Underflow Exception (U)
Two related events contribute to the Underflow exception:
l The creation of a tiny non-zero result between +2^Emin and -2^Emin that can cause some
later exception because it is so tiny.
l The extraordinary loss of accuracy during the approximation of such tiny numbers
by demoralized numbers.
Floating-point Trap Enabled Results: When an underflow trap is enabled, underflow is
signaled when tininess is detected regardless of loss of accuracy. If underflow traps are
enabled, the result register is not modified, and the source registers are preserved.
Floating-point Trap Disabled Results: When an underflow trap is not enabled, underflow is
signaled (using the underflow flag) only when both tininess and loss of accuracy have been
detected. The delivered result might be zero, demoralized, or +2^Emin and -2^Emin .