sdk/2.0/docs/man/vectorDataTypes.xml

<?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>Vector Data Types</keyword>
        </keywordset>
    </refentryinfo>

    <refmeta>
        <refentrytitle>Vector Data Types</refentrytitle>

        <refmiscinfo>
            <copyright>
                <year>2007-2013</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>3</manvolnum>
    </refmeta>

<!-- ================================ SYNOPSIS -->

    <refnamediv id="VectorDataTypes">
        <refname>Vector Data Types</refname>

        <refpurpose>
            Built-in vector data types.
        </refpurpose>
    </refnamediv>

<!-- ================================ DESCRIPTION  -->

    <refsect1 id="description"><title>Description</title>
      <para>
        The <type>char</type>, <type>unsigned char</type>, <type>short</type>, <type>unsigned
        short</type>, <type>integer</type>, <type>unsigned integer</type>, <type>long</type>,
        <type>unsigned long</type>, and <type>float</type> vector data types are supported. The
        vector data type is defined with the type name i.e. <type>char</type>, <type>uchar</type>,
        <type>short</type>, <type>ushort</type>, <type>int</type>, <type>uint</type>,
        <type>float</type>, <type>long</type>, or <type>ulong</type> followed by a literal value
        <emphasis>n</emphasis> that defines the number of elements in the vector. Supported
        values of <emphasis>n</emphasis> are 2, 3, 4, 8, and 16 for all vector data types.
      </para>

      <para>
        The built-in vector data types are also declared as appropriate types in the OpenCL API
        (and header files) that can be used by an application. The following table describes
        the built-in vector data type in the OpenCL C programming language and the corresponding
        data type available to the application:
      </para>

        <informaltable frame="all">
            <tgroup cols="3" 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><type>char<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 8-bit signed two's complement integer values.</entry>
                        <entry><type>cl_char<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>uchar<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 8-bit unsigned integer values.</entry>
                        <entry><type>cl_uchar<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>short<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 16-bit signed two's complement integer values.</entry>
                        <entry><type>cl_short<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>ushort<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 16-bit unsigned integer values.</entry>
                        <entry><type>cl_ushort<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>int<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 32-bit signed two's complement integer values.</entry>
                        <entry><type>cl_int<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>uint<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 32-bit unsigned integer values.</entry>
                        <entry><type>cl_uint<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>long<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 64-bit signed two's complement integer values.</entry>
                        <entry><type>cl_long<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>ulong<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 64-bit unsigned integer values.</entry>
                        <entry><type>cl_ulong<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>float<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 32-bit floating-point values.</entry>
                        <entry><type>cl_float<replaceable>n</replaceable></type></entry>
                    </row>

                    <row>
                        <entry><type>double<replaceable>n</replaceable></type></entry>
                        <entry>A vector of <replaceable>n</replaceable> 64-bit floating-point values.</entry>
                        <entry><type>cl_double<replaceable>n</replaceable></type></entry>
                    </row>
                </tbody>
            </tgroup>
        </informaltable>
    </refsect1>

<!-- ================================ NOTES  -->

    <refsect1 id="notes"><title>Notes</title>
        <para>
          Built-in vector data types are supported by the OpenCL implementation even if the
          underlying compute device does not support any or all of the vector data types. These
          are to be converted by the device compiler to appropriate instructions that use
          underlying built-in types supported natively by the compute device.
        </para>

        <para>
          The <type>double</type> vector type is an optional type that is supported if
          <constant>CL_DEVICE_DOUBLE_FP_CONFIG</constant> for a device is not zero. See
          <citerefentry><refentrytitle>clGetDeviceInfo</refentrytitle></citerefentry>.
        </para>
    </refsect1>

    <refsect1 id="optional"><title>Optional Half Floating Point</title>
        <para>
          The <type>half</type> floating-point is supported as an optional
          extension.  An application that wants to use <type>half</type>
          and <type>half<replaceable>n</replaceable></type> will need to include the
          <citerefentry href="cl_khr_fp16"><refentrytitle>#pragma OPENCL EXTENSION cl_khr_fp16
          : enable</refentrytitle></citerefentry> directive. This will extended the list of
          built-in vector and scalar data types to include the following:
        </para>

        <para>
          <informaltable frame="all">
            <tgroup cols="3" 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><type>half2</type></entry>
                        <entry>A 2-component half-precision floating-point vector.</entry>
                        <entry><type>cl_half2</type></entry>
                    </row>

                    <row>
                        <entry><type>half3</type></entry>
                        <entry>A 3-component half-precision floating-point vector.</entry>
                        <entry><type>cl_half3</type></entry>
                    </row>

                    <row>
                        <entry><type>half4</type></entry>
                        <entry>A 4-component half-precision floating-point vector.</entry>
                        <entry><type>cl_half4</type></entry>
                    </row>

                    <row>
                        <entry><type>half8</type></entry>
                        <entry>An 8-component half-precision floating-point vector.</entry>
                        <entry><type>cl_half8</type></entry>
                    </row>

                    <row>
                        <entry><type>half16</type></entry>
                        <entry>A 16-component half-precision floating-point vector.</entry>
                        <entry><type>cl_half16</type></entry>
                    </row>
                </tbody>
            </tgroup>
          </informaltable>
        </para>

        <para>
          The relational, equality, logical and logical unary operators can be used with
          <type>half</type> scalar and <type>half<replaceable>n</replaceable></type>
          vector types and shall produce a scalar <type>int</type> and vector
          <type>short<replaceable>n</replaceable></type> result respectively.
        </para>

        <para>
            The OpenCL compiler accepts an h and H suffix on floating point literals,
            indicating the literal is typed as a <type>half</type>.
        </para>

    </refsect1>

<!-- ================================ 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="vectorDataTypes">OpenCL Specification</olink>
        </para>
    </refsect1>

<!-- ================================ ALSO SEE  -->

    <refsect1 id="seealso"><title>Also see</title>
        <para>
            <citerefentry href="scalarDataTypes"><refentrytitle>Scalar Data Types</refentrytitle></citerefentry>,
            <citerefentry href="reservedDataTypes"><refentrytitle>Reserved Data Types</refentrytitle></citerefentry>,
            <citerefentry href="otherDataTypes"><refentrytitle>Other Data Types</refentrytitle></citerefentry>,
            <citerefentry href="abstractDataTypes"><refentrytitle>Abstract Data Types</refentrytitle></citerefentry>,
            <citerefentry href="enums"><refentrytitle>Enumerated Data Types</refentrytitle></citerefentry>,
            <citerefentry><refentrytitle>cl_khr_fp16</refentrytitle></citerefentry>,
            <citerefentry><refentrytitle>cl_khr_fp64</refentrytitle></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>

<!-- 4-Jun-2013 -->
</refentry>