
    <para><!-- core spec. p. 259, section 6.12.5 -->
      Built-in geometric functions operate component-wise.  The description
      is per-component.  <type>float<replaceable>n</replaceable></type> is
      <type>float</type>, <type>float2</type>, <type>float3</type>, or <type>float4</type> and
      <type>double<replaceable>n</replaceable></type> is <type>double</type>, <type>double2</type>,
      <type>double3</type>, or <type>double4</type>. The built-in geometric functions are
      implemented using the round to nearest even rounding mode.
    </para>

    <para><!-- core spec. p. 259, section 6.12.5, footnote 53 associated with the section header -->
        The geometric functions can be implemented using contractions such
        as <citerefentry><refentrytitle>mad</refentrytitle></citerefentry> or
        <citerefentry><refentrytitle>fma</refentrytitle></citerefentry>.
    </para>

<!-- 17-Oct-2011 -->