| <?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>clEnqueueNDRangeKernel</keyword> |
| </keywordset> |
| </refentryinfo> |
| |
| <refmeta> |
| <refentrytitle>clEnqueueNDRangeKernel</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="clEnqueueNDRangeKernel"> |
| <refname>clEnqueueNDRangeKernel</refname> |
| |
| <refpurpose> |
| Enqueues a command to execute a kernel on a 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> |
| clEnqueueNDRangeKernel |
| </function> |
| </funcdef> |
| |
| <paramdef><link xlink:href="abstractDataTypes.html">cl_command_queue</link><parameter>command_queue</parameter></paramdef> |
| <paramdef><link xlink:href="abstractDataTypes.html">cl_kernel</link><parameter>kernel</parameter></paramdef> |
| <paramdef><link xlink:href="scalarDataTypes.html">cl_uint</link><parameter>work_dim</parameter></paramdef> |
| <paramdef>const <link xlink:href="scalarDataTypes.html">size_t</link><parameter>*global_work_offset</parameter></paramdef> |
| <paramdef>const <link xlink:href="scalarDataTypes.html">size_t</link><parameter>*global_work_size</parameter></paramdef> |
| <paramdef>const <link xlink:href="scalarDataTypes.html">size_t</link><parameter>*local_work_size</parameter></paramdef> |
| <paramdef><link xlink:href="scalarDataTypes.html">cl_uint</link><parameter>num_events_in_wait_list</parameter></paramdef> |
| <paramdef>const <link xlink:href="abstractDataTypes.html">cl_event</link><parameter>*event_wait_list</parameter></paramdef> |
| <paramdef><link xlink:href="abstractDataTypes.html">cl_event</link><parameter>*event</parameter></paramdef> |
| |
| </funcprototype> |
| </funcsynopsis> |
| </refsynopsisdiv> |
| |
| <!-- ================================ PARAMETERS --> |
| |
| <refsect1 id="parameters"> |
| <title>Parameters</title> |
| <variablelist> |
| |
| <varlistentry> |
| <term> <varname> command_queue </varname> </term> |
| <listitem> |
| <para> |
| A valid host command-queue. The kernel will be queued for execution on |
| the device associated with <varname>command_queue</varname>. |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> <varname> kernel </varname> </term> |
| <listitem> |
| <para> |
| A valid kernel object. The |
| OpenCL context associated with <varname>kernel</varname> and |
| <varname>command_queue</varname> must be the same. |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> <varname> work_dim </varname> </term> |
| <listitem> |
| <para> |
| The number of dimensions |
| used to specify the global work-items and work-items in the |
| work-group. <varname>work_dim</varname> must be greater than zero and less |
| than or equal to <constant>CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS</constant>. |
| If <varname>global_work_size</varname> is NULL, or the value in |
| any passed dimension is 0 then the kernel |
| command will trivially succeed after its event |
| dependencies are satisfied and subsequently |
| update its completion event. The behavior in this |
| situation is similar to that of an enqueued |
| marker, except that unlike a marker, an enqueued |
| kernel with no events passed to |
| <varname>event_wait_list</varname> may run at any time. |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> <varname> global_work_offset </varname> </term> |
| <listitem> |
| <para> |
| <varname>global_work_offset</varname> |
| can be used to specify an array of <varname>work_dim</varname> unsigned |
| values that describe the offset used to calculate the global ID of a |
| work-item. If <varname>global_work_offset</varname> is NULL, the global |
| IDs start at offset (0, 0, ... 0). |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> <varname> global_work_size </varname> </term> |
| <listitem> |
| <para> |
| Points to an array of |
| <varname>work_dim</varname> unsigned values that describe the number |
| of global work-items in <varname>work_dim</varname> dimensions |
| that will execute the kernel function. The total number of global |
| work-items is computed as <varname>global_work_size</varname>[0] *...* |
| <varname>global_work_size</varname>[<varname>work_dim</varname> - 1]. |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> <varname> local_work_size </varname> </term> |
| <listitem> |
| <para> |
| Points to an array of |
| <varname>work_dim</varname> unsigned values that describe the number |
| of work-items that make up a work-group (also referred to as the |
| size of the work-group) that will execute the kernel specified |
| by <varname>kernel</varname>. The total number of work-items in |
| a work-group is computed as <varname>local_work_size</varname>[0] |
| *... * <varname>local_work_size</varname>[<varname>work_dim</varname> - |
| 1]. The total number of work-items in the work-group must be less than |
| or equal to the <constant>CL_KERNEL_WORK_GROUP_SIZE</constant> |
| value specified in table of OpenCL Device Queries for |
| <citerefentry><refentrytitle>clGetDeviceInfo</refentrytitle></citerefentry> |
| and the number of work-items specified |
| in <varname>local_work_size</varname>[0],... |
| <varname>local_work_size</varname>[<varname>work_dim</varname> - 1] |
| must be less than or equal to the corresponding values specified |
| by <constant>CL_DEVICE_MAX_WORK_ITEM_SIZES</constant>[0],.... |
| <constant>CL_DEVICE_MAX_WORK_ITEM_SIZES</constant>[<varname>work_dim</varname> |
| - 1]. The explicitly specified <varname>local_work_size</varname> |
| will be used to determine how to break the global work-items specified |
| by <varname>global_work_size</varname> into appropriate work-group |
| instances. |
| </para> |
| |
| <para> |
| Enabling non-uniform work-groups requires a |
| kernel to be compiled with the -cl-std=CL2.0 flag |
| and without the -cl-uniform-work-group-size |
| flag. If the program was created using |
| clLinkProgram and any of the linked programs |
| were compiled in a way that only supports |
| uniform work-group sizes, the linked program |
| only supports uniform work group sizes. If |
| local_work_size is specified and the OpenCL |
| kernel is compiled without non-uniform work- |
| groups enabled, the values specified in |
| global_work_size[0], ... global_work_size[work_dim - 1] |
| must be evenly divisible by the corresponding |
| values specified in local_work_ size[0], ... |
| local_work_size[work_dim – 1]. |
| </para> |
| |
| <para> |
| If non-uniform work-groups are enabled for the |
| kernel, any single dimension for which the |
| global size is not divisible by the local size |
| will be partitioned into two regions. One region will |
| have work-groups that have the same number of |
| work items as was specified by the local size |
| parameter in that dimension. The other region |
| will have work-groups with less than the number |
| of work items specified by the local size |
| parameter in that dimension. The global IDs and group |
| IDs of the work items in the first region will |
| be numerically lower than those in the second, and |
| the second region will be at most one work-group |
| wide in that dimension. Work-group sizes |
| could be non-uniform in multiple dimensions, |
| potentially producing work-groups of up to 4 |
| different sizes in a 2D range and 8 different |
| sizes in a 3D range. |
| </para> |
| |
| <para> |
| If <varname>local_work_size</varname> is NULL and the |
| kernel is compiled without support for non-uniform work- |
| groups, the OpenCL runtime will implement the |
| ND-range with uniform work-group sizes. If |
| <varname>local_work_size</varname> is NULL and |
| non-uniform-work-groups are enabled, the OpenCL runtime is |
| free to implement the ND-range using uniform or |
| non-uniform work-group sizes, regardless of |
| the divisibility of the global work size. If the |
| ND-range is implemented using non-uniform |
| work-group sizes, the work-group sizes, global |
| IDs and group IDs will follow the same pattern |
| as described in above paragraph. |
| </para> |
| |
| <para> |
| The work-group size to be used for |
| <varname>kernel</varname> can also be |
| specified in the program source or |
| intermediate language. In this case the size |
| of work group specified by <varname>local_work_size</varname> must |
| match the value specified in the program source. |
| </para> |
| |
| <para> |
| These work-group instances are executed in parallel across multiple compute units or |
| concurrently on the same compute unit. |
| </para> |
| |
| <para> |
| Each work-item is uniquely identified by a global identifier. The global ID, |
| which can be read inside the kernel, is computed using the value given by |
| <varname>global_work_size</varname> and <varname>global_work_offset</varname>. |
| In addition, a work-item is also identified within a work-group by a unique local ID. |
| The local ID, which can also be read by the kernel, is computed using the value given |
| by <varname>local_work_size</varname>. The starting local ID is always (0, 0, ... 0). |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> |
| <varname>event_wait_list</varname> and |
| <varname>num_events_in_wait_list</varname> |
| </term> |
| |
| <listitem> |
| <para> |
| Specify events that need |
| to complete before this particular command can be executed. If |
| <varname>event_wait_list</varname> is NULL, then this particular command does |
| not wait on any event to complete. If <varname>event_wait_list</varname> |
| is NULL, <varname>num_events_in_wait_list</varname> must be 0. If |
| <varname>event_wait_list</varname> is not NULL, the list of events |
| pointed to by <varname>event_wait_list</varname> must be valid |
| and <varname>num_events_in_wait_list</varname> must be greater |
| than 0. The events specified in <varname>event_wait_list</varname> |
| act as synchronization points. The context associated with events in |
| <varname>event_wait_list</varname> and <varname>command_queue</varname> must |
| be the same. The memory associated with <varname>event_wait_list</varname> |
| can be reused or freed after the function returns. |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| <varlistentry> |
| <term> <varname> event </varname> </term> |
| <listitem> |
| <para> |
| Returns an event object that identifies |
| this particular kernel-instance. Event objects are unique and |
| can be used to identify a particular kernel execution instance later on. If |
| <varname>event</varname> is NULL, no event will be created for this kernel |
| execution instance and therefore it will not be possible for the application |
| to query or queue a wait for this particular kernel execution instance. |
| If the <varname>event_wait_list</varname> and the |
| <varname>event</varname> arguments are not NULL, the |
| <varname>event</varname> argument |
| should not refer to an element of the |
| <varname>event_wait_list</varname> array. |
| </para> |
| </listitem> |
| </varlistentry> |
| |
| </variablelist> |
| </refsect1> |
| |
| <!-- ================================ NOTES --> |
| |
| <!-- |
| <refsect1 id="notes"><title>Notes</title> |
| </refsect1> |
| --> |
| |
| <!-- ================================ ERRORS --> |
| |
| <refsect1 id="errors"><title>Errors</title> |
| <para> |
| Returns <errorname>CL_SUCCESS</errorname> if the kernel execution was successfully queued. |
| Otherwise, it returns one of the following errors: |
| </para> |
| |
| <itemizedlist mark="disc"> |
| <listitem> |
| <errorname>CL_INVALID_PROGRAM_EXECUTABLE</errorname> |
| if there is no successfully built program |
| executable available for device associated with <varname>command_queue</varname>. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_COMMAND_QUEUE</errorname> if <varname>command_queue</varname> |
| is not a valid host command-queue. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_KERNEL</errorname> if <varname>kernel</varname> is |
| not a valid kernel object. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_CONTEXT</errorname> if context associated with |
| <varname>command_queue</varname> and <varname>kernel</varname> are not the same |
| or if the context associated with <varname>command_queue</varname> and events in |
| <varname>event_wait_list</varname> are not the same. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_KERNEL_ARGS</errorname> if the kernel argument values |
| have not been specified or if |
| a kernel argument declared to be a pointer to a |
| type does not point to a named address space. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_WORK_DIMENSION</errorname> if <varname>work_dim</varname> |
| is not a valid value (i.e. a value between 1 and 3). |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_GLOBAL_WORK_SIZE</errorname> if |
| any of the values specified in |
| <varname>global_work_size</varname>[0], ...<varname>global_work_size</varname> |
| [<varname>work_dim</varname> - 1] exceed the range given by the |
| <code>sizeof(size_t)</code> for the device on which the kernel execution will |
| be enqueued. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_GLOBAL_OFFSET</errorname> if the value specified |
| in <varname>global_work_size</varname> + the corresponding values in |
| <varname>global_work_offset</varname> for any dimensions is greater than the |
| <code>sizeof(size_t)</code> for the device on which the kernel execution will |
| be enqueued. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_WORK_GROUP_SIZE</errorname> if |
| <varname>local_work_size</varname> is specified and does not match |
| the required work-group size for <varname>kernel</varname> in the program source. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_WORK_GROUP_SIZE</errorname> if |
| <varname>local_work_size</varname> is specified and is not |
| consistent with the required number of sub-groups for |
| <varname>kernel</varname> in the program source. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_WORK_GROUP_SIZE</errorname> |
| if <varname>local_work_size</varname> is |
| specified and the total number |
| of work-items in the work-group computed as |
| <varname>local_work_size</varname>[0] * ... |
| <varname>local_work_size</varname>[<varname>work_dim</varname> – 1] |
| is greater than the value specified by |
| <constant>CL_KERNEL_WORK_GROUP_SIZE</constant> in table 5.21. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_WORK_GROUP_SIZE</errorname> if the program was compiled |
| with <code>–cl-uniform-work-group-size</code> and the number of work-items |
| specified by <varname>global_work_size</varname> is not |
| evenly divisible by size of work-group given by <varname>local_work_size</varname> |
| or by the required work- |
| group size specified in the kernel source. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_WORK_ITEM_SIZE</errorname> if the number of |
| work-items specified in any of <varname>local_work_size</varname>[0], |
| ... <varname>local_work_size</varname>[<varname>work_dim</varname> |
| - 1] is greater than the corresponding values specified by |
| <constant>CL_DEVICE_MAX_WORK_ITEM_SIZES</constant>[0], .... |
| <constant>CL_DEVICE_MAX_WORK_ITEM_SIZES</constant>[<varname>work_dim</varname> |
| - 1]. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_MISALIGNED_SUB_BUFFER_OFFSET</errorname> if a sub-buffer object |
| is specified as the value for an argument that is a buffer object and the |
| <varname>offset</varname> specified when the sub-buffer object is created is not |
| aligned to <constant>CL_DEVICE_MEM_BASE_ADDR_ALIGN</constant> value for device |
| associated with <varname>queue</varname>. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_IMAGE_SIZE</errorname> if an image object is specified |
| as an argument value and the image dimensions (image width, height, specified |
| or compute row and/or slice pitch) are not supported by device associated with |
| <varname>queue</varname>. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_IMAGE_FORMAT_NOT_SUPPORTED</errorname> if an image object is specified |
| as an argument value and the image format (image channel order and data type) |
| is not supported by device associated with <varname>queue</varname>. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_OUT_OF_RESOURCES</errorname> if there is a failure to queue the |
| execution instance of <varname>kernel</varname> on the command-queue because of |
| insufficient resources needed to execute the kernel. For example, the explicitly |
| specified <varname>local_work_size</varname> causes a failure to execute the kernel |
| because of insufficient resources such as registers or local memory. Another example |
| would be the number of read-only image args used in <varname>kernel</varname> |
| exceed the <constant>CL_DEVICE_MAX_READ_IMAGE_ARGS</constant> value for |
| device or the number of write-only image args used in <varname>kernel</varname> |
| exceed the <constant>CL_DEVICE_MAX_READ_WRITE_IMAGE_ARGS</constant> value for |
| device or the number of samplers used in <varname>kernel</varname> exceed |
| <constant>CL_DEVICE_MAX_SAMPLERS</constant> for device. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_MEM_OBJECT_ALLOCATION_FAILURE</errorname> if there is a failure to |
| allocate memory for data store associated with image or buffer objects specified |
| as arguments to <varname>kernel</varname>. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_EVENT_WAIT_LIST</errorname> |
| if <varname>event_wait_list</varname> is NULL and |
| <varname>num_events_in_wait_list</varname> > 0, |
| or <varname>event_wait_list</varname> is not NULL and |
| <varname>num_events_in_wait_list</varname> is 0, or if event objects in |
| <varname>event_wait_list</varname> are not valid events. |
| </listitem> |
| |
| <listitem> |
| <errorname>CL_INVALID_OPERATION</errorname> if SVM |
| pointers are passed as arguments to a kernel and the |
| device does not support SVM or if system |
| pointers are passed as arguments to a kernel |
| and/or stored inside SVM allocations passed as |
| kernel arguments and the device does not |
| support fine grain system SVM allocations. |
| </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="clEnqueueNDRangeKernel">OpenCL Specification</olink> |
| </para> |
| </refsect1> |
| |
| <!-- ================================ ALSO SEE --> |
| |
| <refsect1 id="seealso"><title>Also see</title> |
| <para> |
| <citerefentry><refentrytitle>clCreateCommandQueueWithProperties</refentrytitle></citerefentry>, |
| <citerefentry><refentrytitle>clGetDeviceInfo</refentrytitle></citerefentry>, |
| <citerefentry><refentrytitle>clEnqueueNativeKernel</refentrytitle></citerefentry>, |
| <citerefentry href="workItemFunctions"><refentrytitle>Work-Item Functions</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> |
| |
