| <?xml version="1.0" encoding="UTF-8"?> |
| <!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook MathML Module V1.1b1//EN" |
| "http://www.oasis-open.org/docbook/xml/mathml/1.1CR1/dbmathml.dtd"> |
| <refentry> |
| <refentryinfo> |
| <keywordset> |
| <keyword> |
| cl_khr_fp64 |
| </keyword> |
| </keywordset> |
| </refentryinfo> |
| |
| <refmeta> |
| <refentrytitle>cl_khr_fp64</refentrytitle> |
| |
| <refmiscinfo> |
| <copyright> |
| <year>2007-2009</year> |
| <holder>The Khronos Group Inc. |
| Permission is hereby granted, free of charge, to any person obtaining a |
| copy of this software and/or associated documentation files (the |
| "Materials"), to deal in the Materials without restriction, including |
| without limitation the rights to use, copy, modify, merge, publish, |
| distribute, sublicense, and/or sell copies of the Materials, and to |
| permit persons to whom the Materials are furnished to do so, subject to |
| the condition that this copyright notice and permission notice shall be included |
| in all copies or substantial portions of the Materials.</holder> |
| </copyright> |
| </refmiscinfo> |
| <manvolnum>2</manvolnum> |
| </refmeta> |
| |
| <!-- ================================ SYNOPSIS --> |
| |
| <refnamediv id="cl_khr_fp64"> |
| <refname>cl_khr_fp64</refname> |
| |
| <refpurpose> |
| Support for <type>double</type> floating-point precision. |
| </refpurpose> |
| </refnamediv> |
| |
| <!-- ALTERNATIVE SYNTAX SYNOPSIS (NON-FUNCTION) --> |
| <refsect2 id="synopsis"> |
| <title> |
| </title> |
| |
| <informaltable frame="none"> |
| <tgroup cols="1" align="left" colsep="0" rowsep="0"> |
| <colspec colname="col1" colnum="1" /> |
| <tbody> |
| <row> |
| <entry> |
| #pragma OPENCL EXTENSION cl_khr_fp64 : enable |
| </entry> |
| </row> |
| </tbody> |
| </tgroup> |
| </informaltable> |
| </refsect2> |
| |
| |
| <!-- ================================ DESCRIPTION --> |
| |
| <refsect1 id="description"><title>Description</title> |
| <para> |
| Support for double floating-point precision is a requirement for a class of scientific computing algorithms/applications. This class of applications can be enabled by adding support for double precision floating-point as an optional extension. |
| </para> |
| <para> |
| OpenCL 1.0 adds support for double precision floating-point as an optional extension. An application that wants to use <type>double</type> will need to include the <function>#pragma OPENCL EXTENSION cl_khr_fp64 : enable</function> directive before any double precision data type is declared in the kernel code. |
| </para> |
| <para> |
| The lists of built-in <citerefentry href="scalarDataTypes"><refentrytitle>scalar data types</refentrytitle></citerefentry> and <citerefentry href="vectorDataTypes"><refentrytitle>vector data types</refentrytitle></citerefentry> are extended to include the following: |
| <informaltable frame="all"> |
| <tgroup cols="2" align="left" colsep="1" rowsep="1"> |
| <colspec colname="col1" colnum="1" /> |
| <colspec colname="col2" colnum="2" /> |
| <colspec colname="col3" colnum="3" /> |
| |
| <thead> |
| <row> |
| <entry>Type in OpenCL Language</entry> |
| <entry>Description</entry> |
| <entry>API type for application</entry> |
| </row> |
| </thead> |
| |
| <tbody> |
| <row> |
| <entry>double</entry> |
| <entry>A double precision vector</entry> |
| <entry>cl_double</entry> |
| </row> |
| <row> |
| <entry>double2</entry> |
| <entry>2-component double vector</entry> |
| <entry>cl_double2</entry> |
| </row> |
| |
| <row> |
| <entry>double4</entry> |
| <entry>4-component double vector</entry> |
| <entry>cl_double4</entry> |
| </row> |
| |
| <row> |
| <entry>double8</entry> |
| <entry>8-component double vector</entry> |
| <entry>cl_double8</entry> |
| </row> |
| |
| <row> |
| <entry>double16</entry> |
| <entry>16-component double vector</entry> |
| <entry>cl_double16</entry> |
| </row> |
| |
| </tbody> |
| </tgroup> |
| </informaltable> |
| </para> |
| |
| <para> |
| The double data type must confirm to the IEEE-754 double precision storage format. |
| </para> |
| |
| <para> |
| Conversions from <type>double</type> to <type>half</type> are correctly rounded. Conversions from <type>half</type> to <type>double</type> are lossless. |
| </para> |
| </refsect1> |
| |
| |
| <!-- ================================ EXAMPLE --> |
| <!-- DO NOT DELETE IN CASE AN EXAMPLE IS ADDED IN THE FUTURE --> |
| <!-- |
| <refsect2 id="example1"> |
| <title> |
| Example |
| </title> |
| |
| <informaltable frame="none"> |
| <tgroup cols="1" align="left" colsep="0" rowsep="0"> |
| <colspec colname="col1" colnum="1" /> |
| <tbody> |
| <row> |
| <entry> |
| Example goes here - it will be set in "code" type with white space preserved. |
| </entry> |
| </row> |
| </tbody> |
| </tgroup> |
| </informaltable> |
| </refsect2> |
| --> |
| |
| |
| <!-- ================================ SPECIFICATION --> |
| <!-- Set the "uri" attribute in the <olink /> element to the "named destination" for the PDF page |
| --> |
| <refsect1 id="specification"><title>Specification</title> |
| <para> |
| <imageobject> |
| <imagedata fileref="pdficon_small1.gif" format="gif" /> |
| </imageobject> |
| |
| <olink uri="cl_khr_fp64">OpenCL Specification</olink> |
| </para> |
| </refsect1> |
| |
| |
| <!-- ================================ ALSO SEE --> |
| |
| <refsect1 id="seealso"><title>Also see</title> |
| <para> |
| <citerefentry> |
| <refentrytitle>clGetDeviceInfo</refentrytitle>, <citerefentry><refentrytitle>EXTENSION</refentrytitle></citerefentry> |
| </citerefentry> |
| </para> |
| </refsect1> |
| |
| |
| |
| <!-- ============================== COPYRIGHT --> |
| <!-- Content included from copyright.inc.xsl --> |
| |
| <refsect3 id="Copyright"><title></title> |
| <imageobject> |
| <imagedata fileref="KhronosLogo.jpg" format="jpg" /> |
| </imageobject> |
| <para /> |
| </refsect3> |
| |
| </refentry> |
| |