<?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 Auto-vectorization constraints... in OpenCL* for CPU</title>
    <link>https://community.intel.com/t5/OpenCL-for-CPU/Auto-vectorization-constraints/m-p/819886#M1056</link>
    <description>Hi!&lt;BR /&gt;&lt;BR /&gt;Currently auto-vectorization will become enabled only if WorkSize is a multiple of 4. Is it maybe possible to implement optimization which allows vectorization up to (WorkSize/4 * 4) and does only the remainder as scalar? (and/or only the begining thus removing also the constraint for 256 byte alignment). &lt;BR /&gt;&lt;BR /&gt;Thanks!&lt;BR /&gt;Atmapuri&lt;BR /&gt;</description>
    <pubDate>Mon, 03 Oct 2011 13:16:03 GMT</pubDate>
    <dc:creator>janez-makovsek</dc:creator>
    <dc:date>2011-10-03T13:16:03Z</dc:date>
    <item>
      <title>Auto-vectorization constraints...</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Auto-vectorization-constraints/m-p/819886#M1056</link>
      <description>Hi!&lt;BR /&gt;&lt;BR /&gt;Currently auto-vectorization will become enabled only if WorkSize is a multiple of 4. Is it maybe possible to implement optimization which allows vectorization up to (WorkSize/4 * 4) and does only the remainder as scalar? (and/or only the begining thus removing also the constraint for 256 byte alignment). &lt;BR /&gt;&lt;BR /&gt;Thanks!&lt;BR /&gt;Atmapuri&lt;BR /&gt;</description>
      <pubDate>Mon, 03 Oct 2011 13:16:03 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Auto-vectorization-constraints/m-p/819886#M1056</guid>
      <dc:creator>janez-makovsek</dc:creator>
      <dc:date>2011-10-03T13:16:03Z</dc:date>
    </item>
  </channel>
</rss>

