.. index:: pair: namespace; rgpot::simd
.. _doxid-namespacergpot_1_1simd:

namespace rgpot::simd
=====================

.. toctree::
	:hidden:

	namespace_rgpot_simd_detail.rst

Overview
~~~~~~~~




.. ref-code-block:: cpp
	:class: doxyrest-overview-code-block

	
	namespace simd {

	// namespaces

	namespace :ref:`rgpot::simd::detail<doxid-namespacergpot_1_1simd_1_1detail>`;

	// global functions

	double :ref:`exact_fma<doxid-exact__fma_8hpp_1a0ec343a5f4a0560c589ebc170d4eb1ca>`(double a, double b, double c);
	double :target:`fma_trunc<doxid-exact__fma_8hpp_1a189fbb77dbdcc92c88ae4331c966616c>`(double a, double b, double c);
	double :target:`fma_quick<doxid-exact__fma_8hpp_1ae9371c811aa36d48483ed35b0efd12d4>`(double a, double b, double c);
	double :ref:`mul_add<doxid-exact__fma_8hpp_1a6400dc19dfc6964f5daa7da968fc9e12>`(double a, double b, double c);
	double :ref:`r2_of<doxid-exact__fma_8hpp_1a894a4d861296108dba7ed726b526790c>`(double dx, double dy, double dz);
	double :ref:`exp_fixed<doxid-exp_8cc_1a3b98ce20e45a2530b231bb9d1911f83b>`(double x);
	RGPOT_SIMD_CLONES void :ref:`exp_fixed_n<doxid-exp_8cc_1a350d02044999bf40739e4c634df3484b>`(const double* x, double* out, std::size_t n);

	} // namespace simd
.. _details-namespacergpot_1_1simd:

Detailed Documentation
~~~~~~~~~~~~~~~~~~~~~~



Global Functions
----------------

.. index:: pair: function; exact_fma
.. _doxid-exact__fma_8hpp_1a0ec343a5f4a0560c589ebc170d4eb1ca:

.. ref-code-block:: cpp
	:class: doxyrest-title-code-block

	double exact_fma(double a, double b, double c)

a\*b+c with one rounding, the same bits as a hardware fused multiply-add. Normals in range use a Dekker product and a round-to-odd sum (Boldo and Melquiond). Inputs outside that range and subnormal results use std::fma.

.. index:: pair: function; mul_add
.. _doxid-exact__fma_8hpp_1a6400dc19dfc6964f5daa7da968fc9e12:

.. ref-code-block:: cpp
	:class: doxyrest-title-code-block

	double mul_add(double a, double b, double c)

a\*b+c, one rounding. A hardware instruction when this translation unit has one, or when fma_hw_active(). Otherwise fma_quick.

.. index:: pair: function; r2_of
.. _doxid-exact__fma_8hpp_1a894a4d861296108dba7ed726b526790c:

.. ref-code-block:: cpp
	:class: doxyrest-title-code-block

	double r2_of(double dx, double dy, double dz)

dx^2 + dy^2 + dz^2. The first square is a multiply. The two sums fuse.

.. index:: pair: function; exp_fixed
.. _doxid-exp_8cc_1a3b98ce20e45a2530b231bb9d1911f83b:

.. ref-code-block:: cpp
	:class: doxyrest-title-code-block

	double exp_fixed(double x)

e^x with the same bits on every instruction set and C library. Within 1.24 ulp for normal results (see docs/orgmode/reference/simd_exp.org).

.. index:: pair: function; exp_fixed_n
.. _doxid-exp_8cc_1a350d02044999bf40739e4c634df3484b:

.. ref-code-block:: cpp
	:class: doxyrest-title-code-block

	RGPOT_SIMD_CLONES void exp_fixed_n(const double* x, double* out, std::size_t n)

out[i] = exp_fixed(x[i]) for i < n. Vectorised and dispatched on the CPU at run time; every variant returns the bits of exp_fixed.