<?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 TSX PEBS profiling is not accurate enough in Intel® Moderncode for Parallel Architectures</title>
    <link>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/TSX-PEBS-profiling-is-not-accurate-enough/m-p/946389#M5105</link>
    <description>&lt;P&gt;Hi everyone,&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp; I am recently working on a project that takes advantage of the intel TSX extension (RTM). Overall, &amp;nbsp;the program works well except that there are occasions that there are some unexpected abortions due to memory conflict.&amp;nbsp;The result is obtained both&amp;nbsp;with the returned EAX register and&amp;nbsp;PEBS profiling.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;At first, we suspect that the conflicts&amp;nbsp;are&amp;nbsp;induced from our programming mistakes. We have studied the Intel manual again and again, but could not find any clue.&amp;nbsp;However, when we only run a single thread, conflict still happens. &amp;nbsp;We now have to&amp;nbsp;suspect PEBS for TSX is somehow not accurate enough.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;Do anyone know the details? If possible, how can I report&amp;nbsp;this issue to the Intel engineer.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;The target CPU in&amp;nbsp;our project is Intel core&amp;nbsp;i7 4770S.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;Thanks all!&lt;/P&gt;</description>
    <pubDate>Fri, 27 Dec 2013 08:11:52 GMT</pubDate>
    <dc:creator>le_g_1</dc:creator>
    <dc:date>2013-12-27T08:11:52Z</dc:date>
    <item>
      <title>TSX PEBS profiling is not accurate enough</title>
      <link>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/TSX-PEBS-profiling-is-not-accurate-enough/m-p/946389#M5105</link>
      <description>&lt;P&gt;Hi everyone,&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp; I am recently working on a project that takes advantage of the intel TSX extension (RTM). Overall, &amp;nbsp;the program works well except that there are occasions that there are some unexpected abortions due to memory conflict.&amp;nbsp;The result is obtained both&amp;nbsp;with the returned EAX register and&amp;nbsp;PEBS profiling.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;At first, we suspect that the conflicts&amp;nbsp;are&amp;nbsp;induced from our programming mistakes. We have studied the Intel manual again and again, but could not find any clue.&amp;nbsp;However, when we only run a single thread, conflict still happens. &amp;nbsp;We now have to&amp;nbsp;suspect PEBS for TSX is somehow not accurate enough.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;Do anyone know the details? If possible, how can I report&amp;nbsp;this issue to the Intel engineer.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;The target CPU in&amp;nbsp;our project is Intel core&amp;nbsp;i7 4770S.&lt;/P&gt;

&lt;P&gt;&amp;nbsp; &amp;nbsp;Thanks all!&lt;/P&gt;</description>
      <pubDate>Fri, 27 Dec 2013 08:11:52 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/TSX-PEBS-profiling-is-not-accurate-enough/m-p/946389#M5105</guid>
      <dc:creator>le_g_1</dc:creator>
      <dc:date>2013-12-27T08:11:52Z</dc:date>
    </item>
  </channel>
</rss>

