blob: 02816ec9eae3898da930fd4793fd177635748f8c [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>2007-2013</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="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> <!-- The first sentence is not compatible with the second, but that's
how the API spec rev 19 reads, p. 66 -->
A device returned by
<citerefentry><refentrytitle>clGetDeviceIDs</refentrytitle></citerefentry>.
</para>
<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 (Table 4.3).
</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 -->
<para>
Table 4.3:
</para>
<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 of the global address space 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> otherwise. A device is
considered to be available if the device can
be expected to successfully execute
commands enqueued to the device
</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 floating-point capability must be: <constant>CL_FP_FMA</constant>
| <constant>CL_FP_ROUND_TO_NEAREST</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>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>
<citerefentry><refentrytitle>cl_khr_gl_depth_images</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_gl_msaa_sharing</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_initialize_memory</refentrytitle></citerefentry>
cl_khr_context_abort
<citerefentry><refentrytitle>cl_khr_spir</refentrytitle></citerefentry>
<citerefentry><refentrytitle>cl_khr_srgb_image_writes</refentrytitle></citerefentry>
</literallayout>
</para>
<para>
The following approved Khronos extension names
must be returned by all device that support OpenCL C 2.0:</para>
<para>
<literallayout>
cl_khr_byte_addressable_store
<citerefentry><refentrytitle>cl_khr_fp64</refentrytitle></citerefentry> (for backward compatibility if double precision is supported)
<citerefentry><refentrytitle>cl_khr_3d_image_writes</refentrytitle></citerefentry>
cl_khr_image2d_from_buffer
cl_khr_depth_images
</literallayout>
</para>
<para>
Please refer to the OpenCL 2.0 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_GLOBAL_VARIABLE_
PREFERRED_TOTAL_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Maximum preferred total size, in bytes, of
all program variables in the global address
space. This is a performance hint. An
implementation may place such variables
in storage with optimized device access.
This query returns the capacity of such
storage. The minimum value is 0.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_IL_VERSION</constant></entry>
<entry>char []</entry>
<entry>
The intermediate languages that can be
supported by
<citerefentry><refentrytitle>clCreateProgramWithIL</refentrytitle></citerefentry>
for this device. Returns a space-separated
list of IL version strings of the form
&lt;IL_Prefix>_&lt;Major_Version>.&lt;Minor_Version>.
For OpenCL 2.1, "SPIR-V" is a
required IL prefix.
</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 16384 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 16384 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_BASE_
ADDRESS_ALIGNMENT</constant></entry>
<entry>cl_uint</entry>
<entry>
This query should be used when a 2D
image is created from a buffer which was
created using <constant>CL_MEM_USE_HOST_PTR</constant>.
The value returned must be a power of 2.
This query specifies the minimum
alignment in pixels of the <varname>host_ptr</varname>
specified to <citerefentry><refentrytitle>clCreateBuffer</refentrytitle></citerefentry>.
If the device does not support images, this
value must be 0.
</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_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_PITCH_ALIGNMENT</constant></entry>
<entry>cl_uint</entry>
<entry>
The row pitch alignment size in pixels for
2D images created from a buffer. The
value returned must be a power of 2.
If the device does not support images, this
value must be 0.
</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_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</constant> for the embedded
platform profile only. This must be <constant>CL_TRUE</constant> if
<constant>CL_DEVICE_COMPILER_AVAILABLE</constant> is <constant>CL_TRUE</constant>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_LOCAL_MEM_SIZE</constant></entry>
<entry>cl_ulong</entry>
<entry>
Size of local memory region 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_GLOBAL_
VARIABLE_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
The maximum number of bytes of storage
that may be allocated for any single
variable in program scope or inside a
function in OpenCL C declared in the
global address space.
The minimum value is 64 KB.
</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(min(1024*1024*1024, 1/4th of
<constant>CL_DEVICE_GLOBAL_MEM_SIZE</constant>), 32*1024*1024) for devices
that are not of type <constant>CL_DEVICE_TYPE_CUSTOM</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_NUM_SUB_GROUPS</constant></entry>
<entry>cl_uint</entry>
<entry>
Maximum number of sub-groups in a
work-group that a device is capable of
executing on a single compute unit, for any
given kernel-instance running on the
device. The minimum value is 1. (Refer
also to
<citerefentry><refentrytitle>clGetKernelSubGroupInfo</refentrytitle></citerefentry>.)
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_ON_DEVICE_EVENTS</constant></entry>
<entry>cl_uint</entry>
<entry>
The maximum number of events in use by
a device queue. These refer to events
returned by the <function>enqueue_</function> built-in
functions to a device queue or user events
returned by the
<citerefentry><refentrytitle>create_user_event</refentrytitle></citerefentry>
built-in function that have not been
released.
The minimum value is 1024.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_ON_DEVICE_QUEUES</constant></entry>
<entry>cl_uint</entry>
<entry>
The maximum number of device queues
that can be created per context.
The minimum value is 1.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_PARAMETER_SIZE</constant></entry>
<entry>size_t</entry>
<entry>
Max size in bytes of all 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_PIPE_ARGS</constant></entry>
<entry>cl_uint</entry>
<entry>
The maximum number of pipe objects that
can be passed as arguments to a kernel.
The minimum value is 16.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MAX_READ_IMAGE_ARGS</constant></entry>
<entry>cl_uint</entry>
<entry>
Max number of image objects arguments
of a kernel declared with the <code>read_only</code>
qualifier. 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_READ_
WRITE_IMAGE_ARGS</constant></entry>
<entry>cl_uint</entry>
<entry>
Max number of image objects arguments
of a kernel declared with the <code>write_only</code> or <code>read_write</code>
qualifier. The
minimum value is 64 if <constant>CL_DEVICE_IMAGE_SUPPORT</constant>
is <constant>CL_TRUE</constant>.
NOTE: <constant>CL_DEVICE_MAX_WRITE_IMAGE_ARGS</constant>
is only there for backward compatibility.
<constant>CL_DEVICE_MAX_READ_WRITE_IMAGE_ARGS</constant>
should be used instead.
</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 that a device is capable of
executing on a single compute unit, for any
given kernel-instance running on the device.
(Refer also to
<citerefentry><refentrytitle>clEnqueueNDRangeKernel</refentrytitle></citerefentry>)
and <constant>CL_KERNEL_WORK_GROUP_SIZE</constant>).
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> <type>size_t</type> 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 image objects arguments
of a kernel declared with the
<code>write_only</code> qualifier.
The minimum value is 64 if
<constant>CL_DEVICE_IMAGE_SUPPORT</constant> is <constant>CL_TRUE</constant>.
NOTE: <constant>CL_DEVICE_MAX_WRITE_IMAGE_ARGS</constant>
is only there for backward compatibility.
<constant>CL_DEVICE_MAX_READ_WRITE_IMAGE_ARGS</constant>
should be used instead.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_MEM_BASE_ADDR_ALIGN</constant></entry>
<entry>cl_uint</entry>
<entry>
Alignment requirement (in bits) for sub-buffer offsets.
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_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 2.0 if <constant>CL_DEVICE_VERSION</constant> is OpenCL 2.0.
</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>
<para>
The <varname>major_version.minor_version</varname> value returned can
be 1.0 or 1.1 if <constant>CL_DEVICE_VERSION</constant> is OpenCL 1.0.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PARENT_DEVICE</constant></entry>
<entry>cl_device_id</entry>
<entry>
Returns the <type>cl_device_id</type> of the parent device to which this sub-device
belongs. If <varname>device</varname> is a root-level device, a NULL value is 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 <varname>device</varname> 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_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 <varname>device</varname>. This is an array of
<type>cl_device_partition_property</type> 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 cannot be partitioned (i.e.
there is no partitioning scheme supported
by the device that will return at least two
subdevices), 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 <varname>properties</varname> argument specified in
<citerefentry><refentrytitle>clCreateSubDevices</refentrytitle></citerefentry>
if <varname>device</varname> is a subdevice. In the case where the <varname>properties</varname>
argument to <citerefentry><refentrytitle>clCreateSubDevices</refentrytitle></citerefentry> is
<constant>CL_DEVICE_PARTITION_BY_AFFINITY_DOMAIN</constant>,
<constant>CL_DEVICE_AFFINITY_DOMAIN_NEXT_PARTITIONABLE</constant>, the affinity domain used to
perform the partition will be returned. This
can be one 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>
</itemizedlist>
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_PIPE_MAX_
ACTIVE_RESERVATIONS</constant></entry>
<entry>cl_uint</entry>
<entry>
The maximum number of reservations that
can be active for a pipe per work-item in a
kernel. A work-group reservation is
counted as one reservation per work-item.
The minimum value is 1.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PIPE_MAX_PACKET_SIZE</constant></entry>
<entry>cl_uint</entry>
<entry>The maximum size of pipe packet in bytes.
The minimum value is 1024 bytes.
</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_
GLOBAL_ATOMIC_ALIGNMENT</constant></entry>
<entry>cl_uint</entry>
<entry>
Returns the value representing the
preferred alignment in bytes for OpenCL
2.0 atomic types to global memory. This
query can return 0 which indicates that the
preferred alignment is aligned to the
natural size of the type.
</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_PREFERRED_
LOCAL_ATOMIC_ALIGNMENT</constant></entry>
<entry>cl_uint</entry>
<entry>
Returns the value representing the
preferred alignment in bytes for OpenCL
2.0 atomic types to local memory. This
query can return 0 which indicates that the
preferred alignment is aligned to the
natural size of the type.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_PREFERRED_
PLATFORM_ATOMIC_ALIGNMENT</constant></entry>
<entry>cl_uint</entry>
<entry>
Returns the value representing the
preferred alignment in bytes for OpenCL
2.0 fine-grained SVM atomic types. This
query can return 0 which indicates that the
preferred alignment is aligned to the
natural size of the type.
</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 in bytes 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_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>
<para>
The platform profile returns the profile that is
implemented by the OpenCL framework. If the
platform profile
returned is <constant>FULL_PROFILE</constant>,
the OpenCL framework will support devices that
are <constant>FULL_PROFILE</constant> 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>
</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_ON_
DEVICE_MAX_SIZE</constant></entry>
<entry>cl_uint</entry>
<entry>
The max. size of the device queue in bytes. The minimum value is 256 KB for the full
profile and 64 KB for the embedded profile
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_QUEUE_ON_
DEVICE_PREFERRED_SIZE</constant></entry>
<entry>cl_uint</entry>
<entry>
The size of the device queue in bytes
preferred by the implementation.
Applications should use this size for the
device queue to ensure good performance.
The minimum value is 16 KB.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_QUEUE_
ON_DEVICE_PROPERTIES</constant></entry>
<entry>
<para>cl_command_-</para>
<para>queue_properties</para>
</entry>
<entry>
<para>
Describes the on device 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>clCreateCommandQueueWithProperties</refentrytitle></citerefentry>.
The mandated minimum capability is
<constant>CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE</constant>
| <constant>CL_QUEUE_PROFILING_ENABLE</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_QUEUE_ON_
HOST_PROPERTIES</constant></entry>
<entry>
<para>cl_command_-</para>
<para>queue_properties</para>
</entry>
<entry>
<para>
Describes the on host 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>clCreateCommandQueueWithProperties</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 <varname>device</varname> 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>
For the full profile, 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>.
For the embedded profile, see section 10.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_SPIR_VERSIONS</constant></entry>
<entry>char []</entry>
<entry>
If the <citerefentry><refentrytitle>cl_khr_spir</refentrytitle></citerefentry>
extension is enabled,
a space separated list of SPIR versions
supported by the device.
For example returning “1.2 2.0” in this
query implies that SPIR version 1.2 and
2.0 are supported by the implementation.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_SUBGROUP_
INDEPENDENT_
FORWARD_PROGRESS</constant></entry>
<entry>cl_bool</entry>
<entry>
Is <constant>CL_TRUE</constant> if this device supports
independent forward progress of sub-groups,
<constant>CL_FALSE</constant> otherwise. If
<citerefentry><refentrytitle>cl_khr_subgroups</refentrytitle></citerefentry>
is supported by the
device this must return <constant>CL_TRUE</constant>.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_SVM_CAPABILITIES</constant></entry>
<entry><para>cl_device_svm_-</para><para>capabilities</para></entry>
<entry><para>
Describes the various shared virtual
memory (a.k.a. SVM) memory allocation
types the device supports. Coarse-grain
SVM allocations are required to be
supported by all OpenCL 2.0 devices. This
is a bit-field that describes a combination
of the following values:</para>
<para><constant>CL_DEVICE_SVM_COARSE_GRAIN_BUFFER</constant>
Support for coarse-grain buffer
sharing using <citerefentry><refentrytitle>clSVMAlloc</refentrytitle></citerefentry>. Memory
consistency is guaranteed at
synchronization points and the host must
use calls to
<citerefentry><refentrytitle>clEnqueueMapBuffer</refentrytitle></citerefentry> and
<citerefentry><refentrytitle>clEnqueueUnmapMemObject</refentrytitle></citerefentry>.</para>
<para><constant>CL_DEVICE_SVM_FINE_GRAIN_BUFFER</constant>
– Support for fine-grain buffer sharing
using <citerefentry><refentrytitle>clSVMAlloc</refentrytitle></citerefentry>. Memory consistency
is guaranteed atsynchronization points
without need for <citerefentry><refentrytitle>clEnqueueMapBuffer</refentrytitle></citerefentry>
and <citerefentry><refentrytitle>clEnqueueUnmapMemObject</refentrytitle></citerefentry>.</para>
<para><constant>CL_DEVICE_SVM_FINE_GRAIN_SYSTEM</constant>
– Support for sharing the host’s entire
virtual memory including memory
allocated using malloc. Memory
consistency is guaranteed at
synchronization points.</para>
<para><constant>CL_DEVICE_SVM_ATOMICS</constant> – Support
for the OpenCL 2.0 atomic operations that
provide memory consistency across the
host and all OpenCL devices supporting
fine-grain SVM allocations.</para>
<para>
The mandated minimum capability is
<constant>CL_DEVICE_SVM_COARSE_GRAIN_BUFFER</constant>.
</para>
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_TERMINATE_CAPABILITY_KHR</constant></entry>
<entry>cl_device_terminate_capability_khr</entry>
<entry>
If the
<citerefentry><refentrytitle>cl_khr_terminate_context</refentrytitle></citerefentry>
extension is enabled, describes the termination capability of the OpenCL
device. This is a bitfield where a value of
<constant>CL_DEVICE_TERMINATE_CAPABILITY_CONTEXT_KHR</constant>
indicates that context termination is supported.
</entry>
</row>
<row>
<entry><constant>CL_DEVICE_TYPE</constant></entry>
<entry>cl_device_type</entry>
<entry>
The OpenCL device type. Currently supported values are:
<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 2.1.
</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
<constant>FULL_PROFILE</constant>, the OpenCL framework will support devices
that are <constant>FULL_PROFILE</constant> 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 in 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 <varname>param_name</varname> 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>clCreateCommandQueueWithProperties</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>
<!-- 9-Nov-2015, API ver 2.1 rev 19; Ext ver 2.1 rev 10; C lang ver 2.0 rev 30 -->
</refentry>