blob: 96ad6c4da923418b8519bacf1cb1a73e171453b4 [file] [log] [blame]
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "xhtml1-transitional.dtd">
<!-- saved from url=(0013)about:internet -->
<?xml-stylesheet type="text/xsl" href="mathml.xsl"?><html xmlns="http://www.w3.org/1999/xhtml" xmlns:pref="http://www.w3.org/2002/Math/preference" xmlns:xlink="http://www.w3.org/1999/xlink" pref:renderer="mathplayer-dl">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
<style xmlns="" type="text/css">
/* This style sets a margin around the entire page */
html, body {
margin: 10px;
}
p {
font: normal 16px verdana, sans-serif;
margin: 0;
padding-bottom:12px;
}
h1 {
font: bold 25px verdana, sans-serif;
margin-top: 0;
margin-bottom: 3px;
padding-top: 0;
padding-bottom: 0;
}
h2 {
font: bold 19px verdana, sans-serif;
margin-top: 28px;
margin-bottom: 3px;
padding-top: 0;
padding-bottom: 0;
}
h3 {
font: bold 19px verdana, sans-serif !important;
margin-top: 28px;
margin-bottom: 3px;
padding-top: 0;
padding-bottom: 0;
}
li {
font: normal 16px verdana, sans-serif;
margin-top: 0;
margin-bottom: 18px;
padding-top: 0;
padding-bottom: 0;
}
.pdparam {
font: italic 16px verdana, sans-serif;
}
.term {
font: italic 16px verdana, sans-serif;
font-weight: normal;
}
.type {
font: normal 16px verdana, sans-serif !important;
}
.parameter {
font-style: italic;
}
a:link, a:visited {
color: blue;
text-decoration: none;
font: normal 16px;
}
a:hover {
background-color: #FFFF99;
font: normal 16px;
}
div.funcsynopsis {
text-align: left;
background-color: #e6e6e6;
font: normal 16px verdana, sans-serif;
padding-top: 10px;
padding-bottom: 10px;
}
div.funcsynopsis table {
border-collapse: separate;
font: normal 16px verdana, sans-serif;
}
div.funcsynopsis td {
background-color: #e6e6e6;
border: 0 solid #000;
padding: 1px;
font: normal 16px verdana, sans-serif;
}
div.refsect1 {
font-family: verdana, sans-serif;
font-size: 16px;
}
code.constant {
font: normal 16px courier new, monospace !important;
}
span.errorname {
font: normal 16px verdana, sans-serif !important;
}
code.function {
font: bold 16px verdana, sans-serif !important;
}
b.fsfunc {
font: bold 16px verdana, sans-serif !important;
}
code.varname {
font: italic 16px verdana, sans-serif;
}
code.replaceable {
font: italic 16px courier new, monospace;
}
code.funcdef {
font: normal 16px verdana, sans-serif !important;
}
.citerefentry {
font: normal 16px verdana, sans-serif !important;
}
.parameter {
font-style: italic;
}
code.fsfunc {
font: normal 16px verdana, sans-serif !important;
}
/* PARAMETER: This style controls spacing between the terms in Parameter section */
dt {
margin-top: 15px;
}
/* TABLES: These styles apply to all tables OTHER than the Synopsis and Example tables */
div.refsect1 table {
width: 100%;
margin-top: 10px;
background-color: #FFF;
border-collapse: collapse;
border-color: #000;
border-width: 1px;
font: normal 16px verdana, sans-serif;
}
div.refsect1 th {
border-collapse: collapse;
border-color: #000;
border-width: 1px;
font: bold 16px verdana, sans-serif;
}
div.refsect1 td {
background-color: #FFF;
padding: 5px;
vertical-align: text-top;
border-collapse: collapse;
border-color: #000;
border-width: 1px;
font: normal 16px verdana, sans-serif;
}
div.refsect1 p{
font: normal 16px verdana, sans-serif;
margin-top: 8px;
margin-bottom: 8px;
padding-top: 0;
padding-bottom: 0;
}
/* EXAMPLE: These styles apply only to the Example section */
div.refsect2 {
font: normal 16px courier new, monospace !important;
}
div.refsect2 table {
margin-top: 0;
background-color: #e6e6e6;
width: 100%;
border: 0 solid #000;
padding: 2px;
font: normal 16px courier new, monospace !important;
}
div.refsect2 td {
background-color: #e6e6e6;
font: normal 16px courier new, monospace !important;
white-space:pre;
}
/* COPYRIGHT: This style formats the text of the copyright statement at the bottom of the page */
div.refsect3 {
font: normal 11px verdana, sans-serif;
margin-top: 50px;
margin-bottom: 20px;
padding-top: 0;
padding-bottom: 0;
}
</style>
<title>Scalar Data Types</title>
<meta name="generator" content="DocBook XSL Stylesheets V1.79.1" />
<meta name="keywords" content="Scalar Data Types" />
</head>
<body>
<div class="refentry">
<a id="id-1"></a>
<div class="titlepage"></div>
<div xmlns="" class="refnamediv">
<a xmlns="http://www.w3.org/1999/xhtml" id="Built-inScalarDataTypes"></a>
<h1>Scalar Data Types</h1>
<p>
Built-in scalar data types.
</p>
</div>
<div class="refsect1">
<a id="description"></a>
<h2>Description</h2>
<p>
These are the data types available in the OpenCL C programming language used to create
kernels that are executed on OpenCL device(s). The OpenCL C programming language (also
referred to as OpenCL C) is based on the ISO/IEC 9899:1999 C language specification
(a.k.a. C99 specification) with specific extensions and restrictions. Please refer
to the ISO/IEC 9899:1999 specification for a detailed description of the language
grammar. This section describes modifications and restrictions to ISO/IEC 9899:1999
supported in OpenCL C.
</p>
<p>
Most built-in scalar 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 scalar data type in the OpenCL C programming language and the corresponding
data type available to the application:
</p>
<div class="informaltable">
<table class="informaltable" border="1">
<colgroup>
<col align="left" class="col1" />
<col align="left" class="col2" />
<col align="left" class="col3" />
</colgroup>
<thead>
<tr>
<th align="left">Type in OpenCL C Language</th>
<th align="left">Description</th>
<th align="left">API type for application</th>
</tr>
</thead>
<tbody>
<tr>
<td align="left"> <span class="type">bool</span> </td>
<td align="left">
A conditional data type which is either <code class="varname">true</code> or
<code class="varname">false</code>. The value <code class="varname">true</code> expands
to the integer constant 1 and the value <code class="varname">false</code>
expands to the integer constant 0. When any scalar value is converted
to bool, the result is 0 if the value compares equal to 0; otherwise,
the result is 1.
</td>
<td align="left"> n/a </td>
</tr>
<tr>
<td align="left"> <span class="type">char</span> </td>
<td align="left"> A signed two's complement 8-bit integer. </td>
<td align="left"> <span class="type">cl_char</span> </td>
</tr>
<tr>
<td align="left">
<div class="literallayout"><p><span class="type">unsigned char</span>,<br />
<span class="type">uchar</span></p></div>
</td>
<td align="left"> An unsigned 8-bit integer. </td>
<td align="left"> <span class="type">cl_uchar</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">short</span> </td>
<td align="left"> A signed two's complement 16-bit integer. </td>
<td align="left"> <span class="type">cl_short</span> </td>
</tr>
<tr>
<td align="left">
<div class="literallayout"><p><span class="type">unsigned short</span>,<br />
<span class="type">ushort</span></p></div>
</td>
<td align="left"> An unsigned 16-bit integer. </td>
<td align="left"> <span class="type">cl_ushort</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">int</span> </td>
<td align="left"> A signed two's complement 32-bit integer. </td>
<td align="left"> <span class="type">cl_int</span> </td>
</tr>
<tr>
<td align="left">
<div class="literallayout"><p><span class="type">unsigned int</span>,<br />
<span class="type">uint</span></p></div>
</td>
<td align="left"> An unsigned 32-bit integer. </td>
<td align="left"> <span class="type">cl_uint</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">long</span> </td>
<td align="left"> A signed two's complement 64-bit integer. </td>
<td align="left"> <span class="type">cl_long</span> </td>
</tr>
<tr>
<td align="left">
<div class="literallayout"><p><span class="type">unsigned long</span>,<br />
<span class="type">ulong</span></p></div>
</td>
<td align="left"> An unsigned 64-bit integer. </td>
<td align="left"> <span class="type">cl_ulong</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">float</span> </td>
<td align="left">
A 32-bit floating point. The float data type must conform to the
IEEE 754 single precision storage format.
</td>
<td align="left"> <span class="type">cl_float</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">double</span> </td>
<td align="left">
A 64-bit floating point. The <span class="type">double</span> data type must conform to the
IEEE 754 double precision storage format. The <span class="type">double</span> scalar
type is an optional type that is supported if <code class="constant">CL_DEVICE_DOUBLE_FP_CONFIG</code>
for a device is not zero. See the table for <code class="varname">param_name</code> for
<a class="citerefentry" href="clGetDeviceInfo.html"><span class="citerefentry"><span class="refentrytitle">clGetDeviceInfo</span></span></a>
for details.
</td>
<td align="left"> <span class="type">cl_double</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">half</span> </td>
<td align="left">
A 16-bit float. The <span class="type">half</span> data type must conform to the IEEE 754-2008
half precision storage format.
</td>
<td align="left"> <span class="type">cl_half</span> </td>
</tr>
<tr>
<td align="left"> <span class="type">size_t</span> </td>
<td align="left">
The unsigned integer type of the result of the sizeof
operator. This is a 32-bit unsigned integer if <code class="constant">CL_DEVICE_ADDRESS_BITS</code>
defined in <a class="citerefentry" href="clGetDeviceInfo.html"><span class="citerefentry"><span class="refentrytitle">clGetDeviceInfo</span></span></a>
is 32-bits and is a 64-bit unsigned integer if <code class="constant">CL_DEVICE_ADDRESS_BITS</code>
is 64-bits.
</td>
<td align="left"> n/a </td>
</tr>
<tr>
<td align="left"> <span class="type">ptrdiff_t</span> </td>
<td align="left">
A signed integer type that is the result of subtracting two pointers.
This is a 32-bit signed integer if <code class="constant">CL_DEVICE_ADDRESS_BITS</code> defined in
<a class="citerefentry" href="clGetDeviceInfo.html"><span class="citerefentry"><span class="refentrytitle">clGetDeviceInfo</span></span></a>
is 32-bits and is a 64-bit signed integer if <code class="constant">CL_DEVICE_ADDRESS_BITS</code>
is 64-bits.
</td>
<td align="left"> n/a </td>
</tr>
<tr>
<td align="left"> <span class="type">intptr_t</span> </td>
<td align="left">
A signed integer type with the property that any valid pointer to
<span class="type">void</span> can be converted to this type, then converted back
to pointer to <span class="type">void</span>, and the result will compare equal
to the original pointer. This is a 32-bit signed integer if
<code class="constant">CL_DEVICE_ADDRESS_BITS</code> defined in table 4.3 is
32-bits and is a 64-bit signed integer if <code class="constant">CL_DEVICE_ADDRESS_BITS</code>
is 64-bits.
</td>
<td align="left"> n/a </td>
</tr>
<tr>
<td align="left"> <span class="type">uintptr_t</span> </td>
<td align="left">
An unsigned integer type with the property that any valid pointer to
<span class="type">void</span> can be converted to this type, then converted back
to pointer to <span class="type">void</span>, and the result will compare equal to
the original pointer. This is a 32-bit signed integer if
<code class="constant">CL_DEVICE_ADDRESS_BITS</code> defined in table 4.3 is
32-bits and is a 64-bit signed integer if <code class="constant">CL_DEVICE_ADDRESS_BITS</code>
is 64-bits.
</td>
<td align="left"> n/a </td>
</tr>
<tr>
<td align="left"> <span class="type">void</span> </td>
<td align="left">
The <span class="type">void</span> type comprises an empty set of values; it is an
incomplete type that cannot be completed.
</td>
<td align="left"> <span class="type">void</span> </td>
</tr>
</tbody>
</table>
</div>
</div>
<div class="refsect1">
<a id="optional"></a>
<h2>Optional Half Floating Point</h2>
<p>
The <span class="type">half</span> floating-point is supported as an optional
extension. An application that wants to use <span class="type">half</span>
and <span class="type">half<em class="replaceable"><code>n</code></em></span> will need to include the
<a class="citerefentry" href="cl_khr_fp16.html"><span class="citerefentry"><span class="refentrytitle">#pragma OPENCL EXTENSION cl_khr_fp16
: enable</span></span></a> directive. This will extended the list of
built-in vector and scalar data types to include the following:
</p>
<p>
</p>
<div class="informaltable">
<table class="informaltable" border="1">
<colgroup>
<col align="left" class="col1" />
<col align="left" class="col2" />
<col align="left" class="col3" />
</colgroup>
<thead>
<tr>
<th align="left">Type in OpenCL Language</th>
<th align="left">Description</th>
<th align="left">API type for application</th>
</tr>
</thead>
<tbody>
<tr>
<td align="left">
<span class="type">half2</span>
</td>
<td align="left">A 2-component half-precision floating-point vector.</td>
<td align="left">
<span class="type">cl_half2</span>
</td>
</tr>
<tr>
<td align="left">
<span class="type">half3</span>
</td>
<td align="left">A 3-component half-precision floating-point vector.</td>
<td align="left">
<span class="type">cl_half3</span>
</td>
</tr>
<tr>
<td align="left">
<span class="type">half4</span>
</td>
<td align="left">A 4-component half-precision floating-point vector.</td>
<td align="left">
<span class="type">cl_half4</span>
</td>
</tr>
<tr>
<td align="left">
<span class="type">half8</span>
</td>
<td align="left">An 8-component half-precision floating-point vector.</td>
<td align="left">
<span class="type">cl_half8</span>
</td>
</tr>
<tr>
<td align="left">
<span class="type">half16</span>
</td>
<td align="left">A 16-component half-precision floating-point vector.</td>
<td align="left">
<span class="type">cl_half16</span>
</td>
</tr>
</tbody>
</table>
</div>
<p>
</p>
<p>
The relational, equality, logical and logical unary operators can be used with
<span class="type">half</span> scalar and <span class="type">half<em class="replaceable"><code>n</code></em></span>
vector types and shall produce a scalar <span class="type">int</span> and vector
<span class="type">short<em class="replaceable"><code>n</code></em></span> result respectively.
</p>
<p>
The OpenCL compiler accepts an h and H suffix on floating point literals, indicating
the literal is typed as a <span class="type">half</span>.
</p>
</div>
<div class="refsect1">
<a id="specification"></a>
<h2>Specification</h2>
<p>
<img src="pdficon_small1.gif" />
<a href="https://www.khronos.org/registry/cl/specs/opencl-2.0-openclc.pdf#namedest=scalarDataTypes" target="OpenCL Spec">OpenCL Specification</a>
</p>
</div>
<div class="refsect1">
<a id="seealso"></a>
<h2>Also see</h2>
<p>
<a class="citerefentry" href="macroLimits.html"><span class="citerefentry"><span class="refentrytitle">Macros and Limits</span></span></a>,
<a class="citerefentry" href="vectorDataTypes.html"><span class="citerefentry"><span class="refentrytitle">Vector Data Types</span></span></a>,
<a class="citerefentry" href="reservedDataTypes.html"><span class="citerefentry"><span class="refentrytitle">Reserved Data Types</span></span></a>,
<a class="citerefentry" href="otherDataTypes.html"><span class="citerefentry"><span class="refentrytitle">Other Data Types</span></span></a>,
<a class="citerefentry" href="abstractDataTypes.html"><span class="citerefentry"><span class="refentrytitle">Abstract Data Types</span></span></a>,
<a class="citerefentry" href="enums.html"><span class="citerefentry"><span class="refentrytitle">Enumerated Data Types</span></span></a>,
<a class="citerefentry" href="cl_khr_fp16.html"><span class="citerefentry"><span class="refentrytitle">cl_khr_fp16</span></span></a>,
<a class="citerefentry" href="cl_khr_fp64.html"><span class="citerefentry"><span class="refentrytitle">cl_khr_fp64</span></span></a>
</p>
</div>
<div xmlns="" class="refsect3" lang="en" xml:lang="en"><a xmlns="http://www.w3.org/1999/xhtml" id="Copyright"></a><h4 xmlns="http://www.w3.org/1999/xhtml"></h4><img xmlns="http://www.w3.org/1999/xhtml" src="KhronosLogo.jpg" /><p xmlns="http://www.w3.org/1999/xhtml"></p>Copyright © 2007-2015 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.
</div>
</div>
</body>
</html>