d1_struct, d2_struct, d3_struct, d4_struct, d1_t, d2_t, d3_t, d4_t, d1_ptr, d2_ptr, d3_ptr, d4_ptr, d1_srcptr, d2_srcptr, d3_srcptr, d4_srcptr
#include <flint/dfloat.h>d1_structd2_structd3_structd4_structd1_td2_td3_td4_td1_ptrd2_ptrd3_ptrd4_ptrd1_srcptrd2_srcptrd3_srcptrd4_srcptrA dN_struct is a struct with a single member double d[N], the components of the expansion x = d_0 + \cdots + d_{N-1}, and a dN_t is a length-one array of it. The components are roughly nonoverlapping (|d_{i+1}| is at most about \operatorname{ulp}(d_i)), the head d_0 is the double nearest to x (up to ties), and d_0 = 0 implies x = 0. Results of the arithmetic are weakly normalised; see dN_canonicalise for the canonical form. d1 is a plain double, provided for uniformity.
Guides#
The FLINT team, “d1_struct, d2_struct, d3_struct, d4_struct, d1_t, d2_t, d3_t, d4_t, d1_ptr, d2_ptr, d3_ptr, d4_ptr, d1_srcptr, d2_srcptr, d3_srcptr, d4_srcptr (dfloat.h),” FLINT: Fast Library for Number Theory, documentation, version 3.7.0-dev, 2026. https://flintlib.org/doc/dfloat/d1_struct.html (generated 2026-10-20).
@misc{flint-doc-dfloat-d1-struct,
author = {{The FLINT team}},
title = {{d1\_struct, d2\_struct, d3\_struct, d4\_struct, d1\_t, d2\_t, d3\_t, d4\_t, d1\_ptr, d2\_ptr, d3\_ptr, d4\_ptr, d1\_srcptr, d2\_srcptr, d3\_srcptr, d4\_srcptr (dfloat.h)}},
howpublished = {FLINT: Fast Library for Number Theory, documentation, version 3.7.0-dev},
year = {2026},
url = {https://flintlib.org/doc/dfloat/d1_struct.html},
note = {Generated 2026-10-20}
}