float>half-bits ( x -- bits )

Vocabulary

math.floats.small

Inputs

x

a real

Outputs

bits	an integer

Word description

Encodes IEEE binary16 with round-to-nearest, ties-to-even. Conversion is direct from binary64, independent of the host rounding mode. Signed zeros, infinities and subnormals are preserved. NaNs become 0x7e00.

Definition

USING: math.floats.small.private ;

IN: math.floats.small

: float>half-bits ( x -- bits ) 10 5 float>small-bits ;