The rounding mode for floating-point arithmetic is characterized by the
values:
- round_indeterminate
if the rounding style is indeterminable
- round_toward_zero
if the rounding style is toward zero
- round_to_nearest
if the rounding style is to the nearest representable value
- round_toward_infinity
if the rounding style is toward infinity
- round_toward_neg_infinity
if the rounding style is toward negative infinity