pub fn scale_expansion_zeroelim(e: &[f64], b: f64, h: &mut [f64]) -> usize
Expand description
Multiply an expansion by a scalar, eliminating zero components from the output expansion.
Sets h = be
. See either [1] or [2] for details.
Maintains the nonoverlapping property. If round-to-even is used (as
with IEEE 754), maintains the strongly nonoverlapping and nonadjacent
properties as well. (That is, if e
has one of these properties, so
will h
.)