<?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 switching between vector and image formats in Intel® Integrated Performance Primitives</title>
    <link>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826393#M5215</link>
    <description>Yes, but can there be 2 strides for element access patterns, one for an "as vector" access and second for &lt;BR /&gt;"as image" access ? (maybe a third for volume access too). Why I say this is that in Matlab (which also expects vectorized coding style) it is possible, and it simplifies vectorization efforts and yields a good performance impact (as compared to a non-vectorized matlab code). I have personally found it very beneficial to re-interpret the same data block as vector, stack of image frames or volumes within the same application, depending on what is being extracted from (of) the data. &lt;BR /&gt;&lt;BR /&gt;Just a thought at best, maybe an entry in my wish list too...</description>
    <pubDate>Thu, 22 Sep 2011 17:19:58 GMT</pubDate>
    <dc:creator>siddy</dc:creator>
    <dc:date>2011-09-22T17:19:58Z</dc:date>
    <item>
      <title>switching between vector and image formats</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826391#M5213</link>
      <description>Hi experts,&lt;BR /&gt; Any advise on how the application can seamlessly interpret an image (ippi..) as a vector (ipps..)?, of the data type that exists in both representations?</description>
      <pubDate>Tue, 20 Sep 2011 19:17:29 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826391#M5213</guid>
      <dc:creator>siddy</dc:creator>
      <dc:date>2011-09-20T19:17:29Z</dc:date>
    </item>
    <item>
      <title>switching between vector and image formats</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826392#M5214</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;&lt;P&gt;IPPS functions are expected data type is single dimensional (1D), IPPI functions processing data as 2D.&lt;/P&gt;&lt;P&gt;Regards,&lt;/P&gt;&lt;P&gt;Naveen Gv&lt;/P&gt;</description>
      <pubDate>Wed, 21 Sep 2011 04:11:57 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826392#M5214</guid>
      <dc:creator>Naveen_G_Intel</dc:creator>
      <dc:date>2011-09-21T04:11:57Z</dc:date>
    </item>
    <item>
      <title>switching between vector and image formats</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826393#M5215</link>
      <description>Yes, but can there be 2 strides for element access patterns, one for an "as vector" access and second for &lt;BR /&gt;"as image" access ? (maybe a third for volume access too). Why I say this is that in Matlab (which also expects vectorized coding style) it is possible, and it simplifies vectorization efforts and yields a good performance impact (as compared to a non-vectorized matlab code). I have personally found it very beneficial to re-interpret the same data block as vector, stack of image frames or volumes within the same application, depending on what is being extracted from (of) the data. &lt;BR /&gt;&lt;BR /&gt;Just a thought at best, maybe an entry in my wish list too...</description>
      <pubDate>Thu, 22 Sep 2011 17:19:58 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/switching-between-vector-and-image-formats/m-p/826393#M5215</guid>
      <dc:creator>siddy</dc:creator>
      <dc:date>2011-09-22T17:19:58Z</dc:date>
    </item>
  </channel>
</rss>

