Floating-Point Subexpression Elimination; Check Argument Range Of Trigonometric Instructions; Efficiently - AMD Athlon Processor x86 Optimization Manual

22007E/0—November 1999

Floating-Point Subexpression Elimination

Check Argument Range of Trigonometric Instructions

Efficiently

Floating-Point Subexpression Elimination
There are cases which do not require an FXCH instruction after
every instruction to allow access to two new stack entries. In the
cases where two instructions share a source operand, an FXCH
is not required between the two instructions. When there is an
opportunity for subexpression elimination, reduce the number
of superfluous FXCH instructions by putting the shared source
operand at the top of the stack. For example, using the function:
func( (x*y), (x+z) )
Example 1 (Avoid):
FLD Z
FLD Y
FLD X
FADD ST, ST(2)
FXCH ST(1)
FMUL ST, ST(2)
CALL FUNC
FSTP ST(0)
Example 2 (Preferred):
FLD Z
FLD Y
FLD X
FMUL ST(1), ST
FADDP ST(2), ST
CALL FUNC
The transcendental instructions FSIN, FCOS, FPTAN, and
FSINCOS are architecturally restricted in their argument
range. Only arguments with a magnitude of <= 2^63 can be
evaluated. If the argument is out of range, the C2 bit in the FPU
status word is set, and the argument is returned as the result.
Software needs to guard against such (extremely infrequent)
cases.
AMD Athlon™ Processor x86 Code Optimization
103