Lines Matching full:rounding
124 * at 1.0 and 16.0 is due to rounding on conversion to float. See float_from_q4_27() for details.
138 * No rounding is needed as the representation is exact.
151 * No rounding is needed as the representation is exact.
165 * No rounding is needed as the representation is exact.
178 * The data is truncated without rounding.
315 * The data is clamped, and truncated without rounding.
330 * No rounding is needed as the representation is exact for nominal values.
331 * Rounding for overflow values is to nearest, ties to even.
356 * If rounding is needed on truncation, the fractional lsb is rounded to nearest,
371 * Rounding is done according to float_from_i32().
600 * Rounding of 0.5 lsb is to even (default for IEEE 754).
639 * Rounding of 0.5 lsb is to even (default for IEEE 754).
772 * Note the closed range at 1.0 and 16.0 is due to rounding on conversion to float.
782 * is no rounding due to the multiplication - the bit pattern is preserved.
783 * However, there may be rounding due to the fixed-point to float conversion,
796 * Note the closed range at 1.0 and 16.0 is due to rounding on conversion to float.
873 * There is no rounding, the conversion and representation is exact.
881 * is no rounding due to the multiplication - the bit pattern is preserved.
921 * Rounding may occur in the least significant 8 bits for large fixed point
923 * Rounding will be to nearest, ties to even.
937 * There is no rounding, the conversion and representation is exact.
949 * There is no rounding in the nominal range, the conversion and representation
950 * is exact. For values outside the nominal range, rounding is to nearest, ties to even.
