blob: 46d9e9effd651040c509579f0bb882d459162555 [file] [log] [blame]
<?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>clGetDeviceInfo</keyword></keywordset>
</refentryinfo>
<refmeta>
<refentrytitle>clGetDeviceInfo</refentrytitle>
<refmiscinfo>
<copyright>
<year>2007-2011</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="clGetDeviceInfo">
<refname>clGetDeviceInfo</refname>
<refpurpose>Get information about an OpenCL device.</refpurpose>
</refnamediv>
<refsynopsisdiv xmlns:xlink="http://www.w3.org/1999/xlink"><title></title>
<funcsynopsis>
<funcprototype>
<funcdef>
<link xlink:href="scalarDataTypes.html">cl_int</link>
<function>clGetDeviceInfo</function>
</funcdef>
<paramdef>
<link xlink:href="abstractDataTypes.html">cl_device_id</link>
<parameter>device</parameter>
</paramdef>
<paramdef>
<link xlink:href="enums.html#cl_device_info">cl_device_info</link>
<parameter>param_name</parameter>
</paramdef>
<paramdef>
<link xlink:href="scalarDataTypes.html">size_t</link>
<parameter>param_value_size</parameter>
</paramdef>
<paramdef>
<link xlink:href="scalarDataTypes.html">void</link>
<parameter>*param_value</parameter>
</paramdef>
<paramdef>
<link xlink:href="scalarDataTypes.html">size_t</link>
<parameter>*param_value_size_ret</parameter>
</paramdef>
</funcprototype>
</funcsynopsis>
</refsynopsisdiv>
<!-- ================================ PARAMETERS -->
<refsect1 xmlns:xlink="http://www.w3.org/1999/xlink" id="parameters"><title>Parameters</title>
<variablelist>
<varlistentry>
<term><varname>device</varname></term>
<listitem>
<para>
May be a device returned by
<citerefentry><refentrytitle>clGetDeviceIDs</refentrytitle></citerefentry>
or a sub-device created by
<citerefentry><refentrytitle>clCreateSubDevices</refentrytitle></citerefentry>.
If <varname>device</varname> is a sub-device, the specific information for
the sub-device will be returned. The information that can be queried using
<function>clGetDeviceInfo</function> is specified in the table below.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><varname>param_value</varname></term>
<listitem>
<para>
A pointer to memory location where appropriate values for a given
<varname>param_name</varname> as specified in the table below will be
returned. If <varname>param_value</varname> is NULL, it is ignored.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><varname>param_value_size</varname></term> <listitem>
<para>
Specifies the size in bytes of memory pointed to by
<varname>param_value</varname>. This size in bytes must be &ge; size of
return type specified in the table below.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><varname>param_value_size_ret</varname></term> <listitem>
<para>
Returns the actual size in bytes of data being queried by
<varname>param_value</varname>. If <varname>param_value_size_ret</varname>
is NULL, it is ignored.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><varname>param_name</varname></term> <listitem>
<para>
An enumeration constant that identifies the device information being queried. It
can be one of the values as specified in the table below.
</para>
</listitem>
</varlistentry>
<!-- ================================ PARAMETER TABLE (OPTIONAL) -->
<varlistentry>
<term></term>
<listitem>
<!-- table 4.3 -->
<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>cl_device_info</entry>
<entry>Return Type</entry>
<entry>Description</entry>
</row>
</thead>
<tbody>
<row>
<entry><constant>CL_DEVICE_ADDRESS_BITS</constant></entry>
<entry>cl_uint</entry>
<entry>
The default compute device address space size specified as an unsigned
integer value in bits. Currently supported values are 32 or 64 bits.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_AVAILABLE</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if the device is available and
<constant>CL_FALSE</constant> if the device is not available.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_BUILT_IN_KERNELS</constant></entry>
<entry>char[]</entry>
<entry>
A semi-colon separated list of built-in kernels supported by the device.
An empty string is returned if no built-in kernels are supported by
the device.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_COMPILER_AVAILABLE</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_FALSE</constant> if the implementation does
not have a compiler available to compile the program source.
Is <constant>CL_TRUE</constant> if the compiler is available. This can
be <constant>CL_FALSE</constant> for the embedded platform profile only.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_DOUBLE_FP_CONFIG</constant></entry>
<entry><para>cl_device_-</para><para>fp_config</para></entry>
<entry>
Describes double precision floating-point capability of the
OpenCL device. This is a bit-field that describes one or more of the following values:
<itemizedlist mark="disc">
<listitem><constant>CL_FP_DENORM</constant> - denorms are supported.</listitem>
<listitem><constant>CL_FP_INF_NAN</constant> - INF and NaNs are supported.</listitem>
<listitem><constant>CL_FP_ROUND_TO_NEAREST</constant> - round to nearest even rounding mode supported.</listitem>
<listitem><constant>CL_FP_ROUND_TO_ZERO</constant> - round to zero rounding mode supported.</listitem>
<listitem><constant>CL_FP_ROUND_TO_INF</constant> -
round to positive and negative infinity rounding modes supported.</listitem>
<listitem><constant>CL_FP_FMA</constant> - IEEE754-2008 fused multiply-add is supported. </listitem>
<listitem><constant>CL_FP_SOFT_FLOAT</constant> - Basic floating-point operations (such as
addition, subtraction, multiplication) are implemented in software.</listitem>
</itemizedlist>
<para>
Double precision is an optional feature so the mandated minimum double
precision floating-point capability is 0.
</para>
<para>
If double precision is supported by the device, then the minimum double
precision floatingpoint capability must be: <constant>CL_FP_FMA</constant>
| <constant>CL_FP_ROUND_TO_NEAREST</constant>
| <constant>CL_FP_ROUND_TO_ZERO</constant>
| <constant>CL_FP_ROUND_TO_INF</constant> |
<constant>CL_FP_INF_NAN</constant> | <constant>CL_FP_DENORM</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_ENDIAN_LITTLE</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if the OpenCL device is a little endian
device and <constant>CL_FALSE</constant> otherwise.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_ERROR_CORRECTION_SUPPORT</constant></entry>
<entry>Return type: cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if the device implements error correction
for all accesses to compute device memory (global and constant).
Is <constant>CL_FALSE</constant> if the device does not implement such
error correction.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_EXECUTION_CAPABILITIES</constant></entry>
<entry> <para>cl_device_-</para><para>exec_capabilities</para> </entry>
<entry>
Describes the execution capabilities of the device. This is
a bit-field that describes one or more of the following values:
<para><constant>CL_EXEC_KERNEL</constant> - The OpenCL device can execute
OpenCL kernels.</para> <para><constant>CL_EXEC_NATIVE_KERNEL</constant>
- The OpenCL device can execute native kernels.</para> <para>The mandated
minimum capability is <constant>CL_EXEC_KERNEL</constant>.</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_EXTENSIONS</constant></entry>
<entry>char[]</entry>
<entry>
Returns a space separated list of extension names (the extension names themselves do
not contain any spaces) supported by the device. The list of extension names returned
can be vendor supported extension names and one or more of the following Khronos
approved extension names:
<para>
<literallayout>
<citerefentry><refentrytitle>cl_khr_int64_base_atomics</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_int64_extended_atomics</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_fp16</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_gl_sharing</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_gl_event</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_d3d10_sharing</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_dx9_media_sharing</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_d3d11_sharing</refentrytitle></citerefentry>
</literallayout>
</para>
</entry>
<entry>
The following approved Khronos extension names must be returned by all device that support OpenCL C 1.2:
<para>
<literallayout>
cl_khr_global_int32_base_atomics
cl_khr_global_int32_extended_atomics
cl_khr_local_int32_base_atomics
cl_khr_local_int32_extended_atomics
cl_khr_byte_addressable_store
cl_khr_fp64
(for backward compatibility if double precision is supported)
</literallayout>
</para>
<para>
Please refer to the OpenCL 1.2 Extension Specification for a detailed description of these extensions.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_GLOBAL_MEM_CACHE_SIZE</constant></entry>
<entry>cl_ulong</entry>
<entry>Size of global memory cache in bytes.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_GLOBAL_MEM_CACHE_TYPE</constant></entry>
<entry> <para>cl_device_mem-</para><para>_cache_type</para> </entry>
<entry>
Type of global memory cache supported. Valid values are:
<constant>CL_NONE</constant>, <constant>CL_READ_ONLY_CACHE</constant>,
and <constant>CL_READ_WRITE_CACHE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE</constant></entry>
<entry>cl_uint</entry>
<entry>Size of global memory cache line in bytes.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_GLOBAL_MEM_SIZE</constant></entry>
<entry>cl_ulong</entry>
<entry>Size of global device memory in bytes.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_HALF_FP_CONFIG</constant></entry>
<entry><para>cl_device_-</para><para>fp_config</para></entry>
<entry>
Describes the OPTIONAL half precision floating-point capability of the
OpenCL device. This is a bit-field that describes one or more of the following values:
<itemizedlist mark="disc">
<listitem><constant>CL_FP_DENORM</constant> - denorms are supported.</listitem>
<listitem><constant>CL_FP_INF_NAN</constant> - INF and NaNs are supported.</listitem>
<listitem><constant>CL_FP_ROUND_TO_NEAREST</constant> - round to nearest even rounding mode supported.</listitem>
<listitem><constant>CL_FP_ROUND_TO_ZERO</constant> - round to zero rounding mode supported.</listitem>
<listitem><constant>CL_FP_ROUND_TO_INF</constant> - round
to +ve and -ve infinity rounding modes supported.</listitem>
<listitem><constant>CL_FP_FMA</constant> - IEEE754-2008 fused multiply-add is supported. </listitem>
<listitem>
<constant>CL_FP_SOFT_FLOAT</constant> - Basic floating-point operations
(such as addition, subtraction, multiplication) are implemented in software.</listitem>
</itemizedlist>
<para>
The required minimum half precision floating-point capability as
implemented by this extension is <constant>CL_FP_ROUND_TO_ZERO</constant>
or <constant>CL_FP_ROUND_TO_INF</constant> |
<constant>CL_FP_INF_NAN</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_HOST_UNIFIED_MEMORY</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if the device and the host have a unified
memory subsystem and is <constant>CL_FALSE</constant> otherwise.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE_SUPPORT</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if images are supported by the OpenCL
device and <constant>CL_FALSE</constant> otherwise.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE2D_MAX_HEIGHT</constant></entry>
<entry>size_t</entry>
<entry>
Max height of 2D image in pixels. The minimum value is 8192 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE2D_MAX_WIDTH</constant></entry>
<entry>size_t</entry>
<entry>
Max width of 2D image or 1D image not created from a buffer object in pixels.
The minimum value is 8192 if <constant>CL_DEVICE_IMAGE_SUPPORT</constant>
is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE3D_MAX_DEPTH</constant></entry>
<entry>size_t</entry>
<entry>
Max depth of 3D image in pixels. The minimum value is 2048 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE3D_MAX_HEIGHT</constant></entry>
<entry>size_t</entry>
<entry>
Max height of 3D image in pixels. The minimum value is 2048 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE3D_MAX_WIDTH</constant></entry>
<entry>size_t</entry>
<entry>
Max width of 3D image in pixels. The minimum value is 2048 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE_MAX_BUFFER_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Max number of pixels for a 1D image created from a buffer object.
The minimum value is 65536 if <constant>CL_DEVICE_IMAGE_SUPPORT</constant>
is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IMAGE_MAX_ARRAY_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Max number of images in a 1D or 2D image array. The minimum value is 2048 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_LINKER_AVAILABLE</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_FALSE</constant> if the implementation does not
have a linker available. Is <constant>CL_TRUE</constant> if the
linker is available. This can be <constant>CL_FALSE for the embedded
platform profile only. This must be <constant>CL_TRUE</constant> if
<constant>CL_DEVICE_COMPILER_AVAILABLE</constant> is CL_TRUE</constant>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_LOCAL_MEM_SIZE</constant></entry>
<entry>cl_ulong</entry>
<entry>
Size of local memory arena in bytes. The minimum value is 32 KB for
devices that are not of type <constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_LOCAL_MEM_TYPE</constant></entry>
<entry> <para>cl_device_-</para><para>local_mem_type</para></entry>
<entry>
Type of local memory supported. This can be set to
<constant>CL_LOCAL</constant> implying dedicated local memory storage
such as SRAM, or <constant>CL_GLOBAL</constant>. For custom devices,
<constant>CL_NONE</constant> can also be returned indicating no local
memory support.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_CLOCK_FREQUENCY</constant></entry>
<entry>cl_uint</entry>
<entry>Maximum configured clock frequency of the device in MHz.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_COMPUTE_UNITS</constant></entry>
<entry>cl_uint</entry>
<entry>
The number of parallel compute units on the OpenCL device. A work-group executes
on a single compute unit. The minimum value is 1.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_CONSTANT_ARGS</constant></entry>
<entry>cl_uint</entry>
<entry>
Max number of arguments declared with the <citerefentry
href="constant"><refentrytitle>__constant</refentrytitle></citerefentry>
qualifier in a kernel. The minimum value is 8 for devices that are not
of type <constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE</constant></entry>
<entry>cl_ulong</entry>
<entry>
Max size in bytes of a constant buffer allocation. The
minimum value is 64 KB for devices that are not of type
<constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_MEM_ALLOC_SIZE</constant></entry>
<entry>cl_ulong</entry>
<entry>
Max size of memory object allocation in bytes. The minimum value is max (1/4th
of <constant>CL_DEVICE_GLOBAL_MEM_SIZE</constant>, 128*1024*1024) for devices
that are not of type <constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_PARAMETER_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Max size in bytes of the arguments that can be passed to a
kernel. The minimum value is 1024 for devices that are not of type
<constant>CL_DEVICE_TYPE_CUSTOM</constant>. For this minimum value,
only a maximum of 128 arguments can be passed to a kernel.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_READ_IMAGE_ARGS</constant></entry>
<entry>cl_uint</entry>
<entry>
Max number of simultaneous image objects that can be read by a kernel. The
minimum value is 128 if <constant>CL_DEVICE_IMAGE_SUPPORT</constant>
is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_SAMPLERS</constant></entry>
<entry>cl_uint</entry>
<entry>
Maximum number of samplers that can be used in a kernel. The minimum
value is 16 if <constant>CL_DEVICE_IMAGE_SUPPORT</constant>
is <constant>CL_TRUE</constant>. (Also see
<citerefentry><refentrytitle>sampler_t</refentrytitle></citerefentry>.)
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_WORK_GROUP_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Maximum number of work-items in a work-group executing a kernel on a
single compute unit, using the data parallel execution model. (Refer to
<citerefentry><refentrytitle>clEnqueueNDRangeKernel</refentrytitle></citerefentry>).
The minimum value is 1.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS</constant></entry>
<entry>cl_uint</entry>
<entry>
Maximum dimensions that specify the global and local
work-item IDs used by the data parallel execution model. (Refer to
<citerefentry><refentrytitle>clEnqueueNDRangeKernel</refentrytitle></citerefentry>).
The minimum value is 3 for devices that are not of type
<constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_WORK_ITEM_SIZES</constant></entry>
<entry>size_t[]</entry>
<entry>
<para>
Maximum number of work-items that can be
specified in each dimension of the work-group to
<citerefentry><refentrytitle>clEnqueueNDRangeKernel</refentrytitle></citerefentry>.
</para>
<para>
Returns <varname>n</varname> size_t entries, where
<varname>n</varname> is the value returned by the query for
<constant>CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS</constant>.
</para>
<para>
The minimum value is (1, 1, 1) for devices that are not of type
<constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_WRITE_IMAGE_ARGS</constant></entry>
<entry>cl_uint</entry>
<entry>
Max number of simultaneous image objects that can
be written to by a kernel. The minimum value is 8 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MEM_BASE_ADDR_ALIGN</constant></entry>
<entry>cl_uint</entry>
<entry>
The minimum value is the size (in bits) of the largest OpenCL built-in
data type supported by the device (<type>long16</type> in FULL profile,
<type>long16</type> or <type>int16</type> in EMBEDDED profile) for devices
that are not of type <constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MIN_DATA_TYPE_ALIGN_SIZE</constant></entry>
<entry>cl_uint</entry>
<entry>
Deprecated in OpenCL 1.2. The smallest alignment in bytes which can be
used for any data type.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_NAME</constant></entry>
<entry>char[]</entry>
<entry>Device name string.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_CHAR</constant>
<para><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_SHORT</constant></para>
<para><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_INT</constant></para>
<para><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_LONG</constant></para>
<para><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_FLOAT</constant></para>
<para><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_DOUBLE</constant></para>
<para><constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_HALF</constant></para></entry>
<entry>cl_uint</entry>
<entry>
<para>
Returns the native ISA vector width. The vector width is defined as
the number of scalar elements that can be stored in the vector.
</para>
<para>
If double precision is not supported,
<constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_DOUBLE</constant> must return 0.
</para>
<para>
If the
<citerefentry><refentrytitle>cl_khr_fp16</refentrytitle></citerefentry>
extension is not supported,
<constant>CL_DEVICE_NATIVE_VECTOR_WIDTH_HALF</constant> must return 0.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_OPENCL_C_VERSION</constant></entry>
<entry>char[]</entry>
<entry>
<para>
OpenCL C version string. Returns the highest OpenCL C version
supported by the compiler for this device that is not of type
<constant>CL_DEVICE_TYPE_CUSTOM</constant>. This version string has
the following format:
</para>
<para>
<varname>
OpenCL&lt;space&gt;C&lt;space&gt;&lt;major_version.minor_version&gt;&lt;space&gt;&lt;vendor-specific
information&gt; </varname>
</para>
<para>
The <varname>major_version.minor_version</varname> value returned must
be 1.2 if <constant>CL_DEVICE_VERSION</constant> is OpenCL 1.2.
</para>
<para>
The <varname>major_version.minor_version</varname> value returned must
be 1.1 if <constant>CL_DEVICE_VERSION</constant> is OpenCL 1.1.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PARENT_DEVICE</constant></entry>
<entry>cl_device_id</entry>
<entry>
Returns the cl_device_id of the parent device to which this sub-device
belongs. If device is a root-level device, a NULL value is returned.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PARTITION_MAX_SUB_DEVICES</constant></entry>
<entry>cl_uint</entry>
<entry>
Returns the maximum number of sub-devices that can be created
when a device is partitioned. The value returned cannot exceed
<constant>CL_DEVICE_MAX_COMPUTE_UNITS</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PARTITION_PROPERTIES</constant></entry>
<entry><para>cl_device_partit-</para><para>ion_property[]</para></entry>
<entry>
Returns the list of partition types supported by device. This is an array of
<varname>cl_device_partition_property</varname> values drawn from the following list:
<itemizedlist mark="disc">
<listitem><constant>CL_DEVICE_PARTITION_EQUALLY</constant></listitem>
<listitem><constant>CL_DEVICE_PARTITION_BY_COUNTS</constant></listitem>
<listitem><constant>CL_DEVICE_PARTITION_BY_AFFINITY_DOMAIN</constant></listitem>
</itemizedlist>
If the device does not support any partition types, a value of 0 will be returned.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PARTITION_AFFINITY_DOMAIN</constant></entry>
<entry><para>cl_device_aff-</para><para>inity_domain</para></entry>
<entry>
Returns the list of supported affinity domains for partitioning the device using
<constant>CL_DEVICE_PARTITION_BY_AFFINITY_DOMAIN</constant>.
This is a bit-field that describes one or more of the following values:
<itemizedlist mark="disc">
<listitem><constant>CL_DEVICE_AFFINITY_DOMAIN_NUMA</constant></listitem>
<listitem><constant>CL_DEVICE_AFFINITY_DOMAIN_L4_CACHE</constant></listitem>
<listitem><constant>CL_DEVICE_AFFINITY_DOMAIN_L3_CACHE</constant></listitem>
<listitem><constant>CL_DEVICE_AFFINITY_DOMAIN_L2_CACHE</constant></listitem>
<listitem><constant>CL_DEVICE_AFFINITY_DOMAIN_L1_CACHE</constant></listitem>
<listitem><constant>CL_DEVICE_AFFINITY_DOMAIN_NEXT_PARTITIONABLE</constant></listitem>
</itemizedlist>
If the device does not support any affinity domains, a value of 0 will be returned.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PARTITION_TYPE</constant></entry>
<entry><para>cl_device_part-</para><para>ition_property[]</para></entry>
<entry>
Returns the properties argument specified in
<citerefentry><refentrytitle>clCreateSubDevices</refentrytitle></citerefentry>
if <varname>device</varname> is a subdevice. Otherwise the implementation
may either return a <varname>param_value_size_ret</varname> of 0 i.e. there
is no partition type associated with <varname>device</varname> or can
return a property value of 0 (where 0 is used to terminate the partition
property list) in the memory that <varname>param_value</varname> points to.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PLATFORM</constant></entry>
<entry><link xlink:href="abstractDataTypes.html">cl_platform_id</link></entry>
<entry>The platform associated with this device.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_CHAR</constant>
<para><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_SHORT</constant></para>
<para><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_INT</constant></para>
<para><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_LONG</constant></para>
<para><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_FLOAT</constant></para>
<para><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_DOUBLE</constant></para>
<para><constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_HALF</constant></para></entry>
<entry>cl_uint</entry>
<entry>
<para>
Preferred native vector width size for built-in scalar types that can
be put into vectors. The vector width is defined as the number of scalar
elements that can be stored in the vector.
</para>
<para>
If double precision is not supported,
<constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_DOUBLE</constant> must return 0.
</para>
<para>
If the
<citerefentry><refentrytitle>cl_khr_fp16</refentrytitle></citerefentry>
extension is not supported,
<constant>CL_DEVICE_PREFERRED_VECTOR_WIDTH_HALF</constant> must return 0.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PRINTF_BUFFER_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Maximum size of the internal buffer that holds the output of <citerefentry
href="printfFunction"><refentrytitle>printf</refentrytitle></citerefentry>
calls from a kernel. The minimum value for the FULL profile is 1 MB.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PREFERRED_INTEROP_USER_SYNC</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if the device's preference is for the user
to be responsible for synchronization, when sharing memory objects between
OpenCL and other APIs such as DirectX, <constant>CL_FALSE</constant> if the
device / implementation has a performant path for performing synchronization
of memory object shared between OpenCL and other APIs such as DirectX
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PROFILE</constant></entry>
<entry>char[]</entry>
<entry>
<para>
OpenCL profile string. Returns the profile name supported by the device
(see note). The profile name returned can be one of the following strings:
</para>
<para>
FULL_PROFILE - if the device supports the OpenCL specification
(functionality defined as part of the core specification and does not
require any extensions to be supported).
</para>
<para>
EMBEDDED_PROFILE - if the device supports the OpenCL embedded profile.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PROFILING_TIMER_RESOLUTION</constant></entry>
<entry>size_t</entry>
<entry>Describes the resolution of device timer. This is measured in nanoseconds.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_QUEUE_PROPERTIES</constant></entry>
<entry>
<para>cl_command_-</para>
<para>queue_properties</para>
</entry>
<entry>
<para>
Describes the command-queue properties supported by the device. This is
a bit-field that describes one or more of the following values:
</para>
<para><constant>CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE</constant></para>
<para><constant>CL_QUEUE_PROFILING_ENABLE</constant></para>
<para>
These properties are described in the table for
<citerefentry><refentrytitle>clCreateCommandQueue</refentrytitle></citerefentry>.
The mandated minimum capability is
<constant>CL_QUEUE_PROFILING_ENABLE</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_REFERENCE_COUNT</constant></entry>
<entry>cl_uint</entry>
<entry>
Returns the device reference count. If the device is a root-level device,
a reference count of one is returned.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_SINGLE_FP_CONFIG</constant></entry>
<entry><para>cl_device_-</para><para>fp_config</para></entry>
<entry>
<para>
Describes single precision floating-point capability of the device. This
is a bit-field that describes one or more of the following values:
</para>
<para><constant>CL_FP_DENORM</constant> - denorms are supported</para>
<para><constant>CL_FP_INF_NAN</constant> - INF and quiet NaNs are supported</para>
<para><constant>CL_FP_ROUND_TO_NEAREST</constant> - round to nearest even rounding mode supported</para>
<para><constant>CL_FP_ROUND_TO_ZERO</constant> - round to zero rounding mode supported</para>
<para><constant>CL_FP_ROUND_TO_INF</constant> - round to +ve and -ve infinity rounding modes supported</para>
<para><constant>CL_FP_FMA</constant> - IEEE754-2008 fused multiply-add is supported</para>
<para>
<constant>CL_FP_CORRECTLY_ROUNDED_DIVIDE_SQRT</constant> - divide and
sqrt are correctly rounded as defined by the IEEE754 specification.
</para>
<para>
<constant>CL_FP_SOFT_FLOAT</constant> - Basic floating-point operations
(such as addition, subtraction, multiplication) are implemented in
software.
</para>
<para>
The mandated minimum floating-point capability for devices that
are not of type <constant>CL_DEVICE_TYPE_CUSTOM</constant>
is <constant>CL_FP_ROUND_TO_NEAREST</constant> |
<constant>CL_FP_INF_NAN</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_TYPE</constant></entry>
<entry>cl_device_type</entry>
<entry>
The OpenCL device type. Currently supported values are one of or
a combination of: <constant>CL_DEVICE_TYPE_CPU</constant>,
<constant>CL_DEVICE_TYPE_GPU</constant>,
<constant>CL_DEVICE_TYPE_ACCELERATOR</constant>,
<constant>CL_DEVICE_TYPE_DEFAULT</constant>, a combination of the above
types, or <constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_VENDOR</constant></entry>
<entry>char[]</entry>
<entry>Vendor name string.</entry>
</row>
<row>
<entry><constant>CL_DEVICE_VENDOR_ID</constant></entry>
<entry>cl_uint</entry>
<entry>
A unique device vendor identifier. An example of a unique device identifier
could be the PCIe ID.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_VERSION</constant></entry>
<entry>char[]</entry>
<entry>
<para>
OpenCL version string. Returns the OpenCL version supported by the
device. This version string has the following format:
</para>
<para>
<varname>
OpenCL&lt;space&gt;&lt;major_version.minor_version&gt;&lt;space&gt;&lt;vendor-specific
information&gt; </varname>
</para>
<para>
The <varname>major_version.minor_version</varname> value returned will
be 1.2.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DRIVER_VERSION</constant></entry>
<entry>char[]</entry>
<entry>
OpenCL software driver version string in the form
<varname>major_number.minor_number</varname>.
</entry>
</row>
</tbody>
</tgroup>
</informaltable>
</listitem>
</varlistentry>
<!-- ================================ END PARAMETER TABLE -->
</variablelist>
</refsect1>
<!-- ================================ NOTES -->
<refsect1 id="notes"><title>Notes</title>
<para>
<constant>CL_DEVICE_PROFILE</constant>: The platform profile returns the profile
that is implemented by the OpenCL framework. If the platform profile returned is
FULL_PROFILE, the OpenCL framework will support devices that are FULL_PROFILE and may
also support devices that are <constant>EMBEDDED_PROFILE</constant>. The compiler must
be available for all devices i.e. <constant>CL_DEVICE_COMPILER_AVAILABLE</constant>
is <constant>CL_TRUE</constant>. If the platform profile returned
is <constant>EMBEDDED_PROFILE</constant>, then devices that are only
<constant>EMBEDDED_PROFILE</constant> are supported.
</para>
<para>
The device queries described the table above should return the
same information for a root-level device i.e. a device returned by
<citerefentry><refentrytitle>clGetDeviceIDs</refentrytitle></citerefentry> and any
sub-devices created from this device except for the following queries:
<itemizedlist mark="disc">
<listitem><constant>CL_DEVICE_GLOBAL_MEM_CACHE_SIZE</constant></listitem>
<listitem><constant>CL_DEVICE_BUILT_IN_KERNELS</constant></listitem>
<listitem><constant>CL_DEVICE_PARENT_DEVICE</constant></listitem>
<listitem><constant>CL_DEVICE_PARTITION_TYPE</constant></listitem>
<listitem><constant>CL_DEVICE_REFERENCE_COUNT</constant></listitem>
</itemizedlist>
</para>
</refsect1>
<!-- ================================ ERRORS -->
<refsect1 id="errors"><title>Errors</title>
<para>
<function>clGetDeviceInfo</function> returns <errorname>CL_SUCCESS</errorname> if the
function is executed successfully. Otherwise, it returns the following:
</para>
<itemizedlist mark="disc">
<listitem>
<errorname>CL_INVALID_DEVICE</errorname> if <varname>device</varname> is not valid.
</listitem>
<listitem>
<errorname>CL_INVALID_VALUE</errorname> if <varname>param_name</varname>
is not one of the supported values or if size in bytes specified by
<varname>param_value_size</varname> is less than size of return type as shown in the
table above and <varname>param_value</varname> is not a <errorname>NULL</errorname>
value or if param_name is a value that is available as an extension and the corresponding
extension is not supported by the device.
</listitem>
<listitem>
<errorname>CL_OUT_OF_RESOURCES</errorname> if there is a failure to allocate resources
required by the OpenCL implementation on the device.
</listitem>
<listitem>
<errorname>CL_OUT_OF_HOST_MEMORY</errorname> if there is a failure to allocate
resources required by the OpenCL implementation on the host.
</listitem>
</itemizedlist>
</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="clGetDeviceInfo">OpenCL Specification</olink>
</para>
</refsect1>
<!-- ================================ ALSO SEE -->
<refsect1 id="seealso"><title>Also see</title>
<para>
<citerefentry><refentrytitle>clGetDeviceIDs</refentrytitle></citerefentry>,
<citerefentry><refentrytitle>cl_khr_fp64</refentrytitle></citerefentry>,
<citerefentry href="constant"><refentrytitle>__constant</refentrytitle></citerefentry>,
<citerefentry><refentrytitle>clCreateCommandQueue</refentrytitle></citerefentry>,
<citerefentry><refentrytitle>clRetainCommandQueue</refentrytitle></citerefentry>,
<citerefentry><refentrytitle>clEnqueueNDRangeKernel</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>
<!-- 28-Oct-2011 -->
</refentry>