 |
Processor—One
option for
each
supported
x86/Pentium
model. This
setting
determines
the
instructions
permitted,
as well as
the pipeline
optimization
strategy
used. |
 |
fprecise—Stores
all floating
point
calculations
in
memory to
ensure
precise
truncation.
Fully
compatible
with the
GNU
convention
for handling
floating
point
comparisons,
including
comparisons
that involve
NANs. |
 |
ffunctions—Enables
the compiler
to use the
387 hardware
instructions
for certain
floating
point
functions
instead
of calling
them in the
library. |
 |
Long
Long Support—Fully
supports
64-bit data
types,
constants,
and
expressions
through the
long long
data
type. |
 |
Start-up
Code—Start-up
code
supports the
romcopy
layout. |
 |
Hardware
Coprocessor—In
hardware
floating-point
(HFP)
mode, the
compiler
uses
floating-point
registers
to hold
floating-point
data and to
perform
floatingpoint
operations. |
 |
Floating-Point
Built-ins—Controls
the use of
floatingpoint
intrinsics,
such as FABS,
FSQRT, and
FSIN. |
 |
Placement of
Zero-Initialized
Data—Controls
the
allocation
of variables
and arrays
explicitly
initialized
to
zero.
Allocating
such objects
to an
uninitialized
section
will
generally
reduce the
size of the
executable
ROM
image. |