Lines Matching full:rounding
756 -mfp-trap-mode=MODE -mfp-rounding-mode=MODE
6494 `-ffinite-math-only', `-fno-rounding-math', `-fno-signaling-nans'
6547 which relies on rounding behavior like `(x + 2**52) - 2**52)'.
6604 floating point rounding behavior. This is round-to-zero for all
6607 for programs that change the FP rounding mode dynamically, or that
6608 may be executed with a non-default rounding mode. This option
6610 compile-time (which may be affected by rounding mode) and
6612 sign-dependent rounding modes.
6614 The default is `-fno-rounding-math'.
6617 disable all GCC optimizations that are affected by rounding mode.
10175 `-mfp-rounding-mode=ROUNDING-MODE'
10176 Selects the IEEE rounding mode. Other Alpha compilers call this
10177 option `-fprm ROUNDING-MODE'. The ROUNDING-MODE can be one of:
10180 Normal IEEE rounding mode. Floating point numbers are
10188 Chopped rounding mode. Floating point numbers are rounded
10192 Dynamic rounding mode. A field in the floating point control
10194 controls the rounding mode in effect. The C library
10195 initializes this register for rounding towards plus infinity.
11252 Setting the rounding of floating-point operations to less than the
11496 instructions have a different rounding behavior compared to
16248 * `The rounding behaviors characterized by non-standard values of
16258 * `The direction of rounding when an integer is converted to a
16264 * `The direction of rounding when a floating-point number is
16289 * `Additional floating-point exceptions, rounding modes,
17541 * Pragmas to control overflow and rounding behaviors are not
21027 there is a system register which can be set to control the rounding
24162 rounding are made.
41417 * mfp-rounding-mode: DEC Alpha Options. (line 85)
