<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Intel OpenCL SDK: clGetKernelWorkGroupInfo return value in OpenCL* for CPU</title>
    <link>https://community.intel.com/t5/OpenCL-for-CPU/Intel-OpenCL-SDK-clGetKernelWorkGroupInfo-return-value/m-p/972110#M2393</link>
    <description>&lt;P&gt;Hi&amp;nbsp;&lt;/P&gt;

&lt;P&gt;1. I have a global work size of 1024 by 1024.&lt;/P&gt;

&lt;P&gt;2. I set the local work size to 16 by 16.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;3. My CPU opnecl device has a maximum work-group-size of 8192.&lt;/P&gt;

&lt;P&gt;4. I call&amp;nbsp;clEnqueueNDRangeKernel with the desired local-work-size (along with all other necessary parameters)&lt;/P&gt;

&lt;P&gt;5. I call:&lt;/P&gt;

&lt;P&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; a.&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_WORK_GROUP_SIZE, sizeof(size_t), (void*)&amp;amp;workGroupSizeUsed, NULL);&lt;/SPAN&gt;&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; b. clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE, sizeof(size_t), (void*)&amp;amp;workGroupSizeUsed, NULL);&lt;/P&gt;

&lt;P&gt;6. Both calls return 8192. How is this possible?&lt;BR /&gt;
	&lt;BR /&gt;
	My expectation is 16 - the value that I passed to it.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Any help?&lt;/P&gt;</description>
    <pubDate>Sat, 05 Apr 2014 21:57:37 GMT</pubDate>
    <dc:creator>Dave_O_</dc:creator>
    <dc:date>2014-04-05T21:57:37Z</dc:date>
    <item>
      <title>Intel OpenCL SDK: clGetKernelWorkGroupInfo return value</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Intel-OpenCL-SDK-clGetKernelWorkGroupInfo-return-value/m-p/972110#M2393</link>
      <description>&lt;P&gt;Hi&amp;nbsp;&lt;/P&gt;

&lt;P&gt;1. I have a global work size of 1024 by 1024.&lt;/P&gt;

&lt;P&gt;2. I set the local work size to 16 by 16.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;3. My CPU opnecl device has a maximum work-group-size of 8192.&lt;/P&gt;

&lt;P&gt;4. I call&amp;nbsp;clEnqueueNDRangeKernel with the desired local-work-size (along with all other necessary parameters)&lt;/P&gt;

&lt;P&gt;5. I call:&lt;/P&gt;

&lt;P&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; a.&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_WORK_GROUP_SIZE, sizeof(size_t), (void*)&amp;amp;workGroupSizeUsed, NULL);&lt;/SPAN&gt;&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; b. clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE, sizeof(size_t), (void*)&amp;amp;workGroupSizeUsed, NULL);&lt;/P&gt;

&lt;P&gt;6. Both calls return 8192. How is this possible?&lt;BR /&gt;
	&lt;BR /&gt;
	My expectation is 16 - the value that I passed to it.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Any help?&lt;/P&gt;</description>
      <pubDate>Sat, 05 Apr 2014 21:57:37 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Intel-OpenCL-SDK-clGetKernelWorkGroupInfo-return-value/m-p/972110#M2393</guid>
      <dc:creator>Dave_O_</dc:creator>
      <dc:date>2014-04-05T21:57:37Z</dc:date>
    </item>
    <item>
      <title>Querying</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Intel-OpenCL-SDK-clGetKernelWorkGroupInfo-return-value/m-p/972111#M2394</link>
      <description>&lt;P&gt;Querying clGetKernelWorkGroupInfo with CL_KERNEL_WORK_GROUP_SIZE returns the maximum work group size supported for that kernel as determined by its resource utilization (e.g., private, local memory) or kernel attribute max_work_group_size. &amp;nbsp;On the other hand, CL_KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE is a performance hint often used to map to a device's underlying SIMD architecture. &amp;nbsp;The two may return the same value, and CPUs seem like they would benefit from large work group sizes more than small work group sizes.&lt;/P&gt;</description>
      <pubDate>Sat, 05 Apr 2014 22:29:23 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Intel-OpenCL-SDK-clGetKernelWorkGroupInfo-return-value/m-p/972111#M2394</guid>
      <dc:creator>Sean_S_3</dc:creator>
      <dc:date>2014-04-05T22:29:23Z</dc:date>
    </item>
  </channel>
</rss>

