<?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 Asymetrical performance between PCIe peer-to-peer devices in Software Tuning, Performance Optimization &amp; Platform Monitoring</title>
    <link>https://community.intel.com/t5/Software-Tuning-Performance/Asymetrical-performance-between-PCIe-peer-to-peer-devices/m-p/981330#M3140</link>
    <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;@admins: if this is a wrong place, feel free to move it to the relevant topic.&lt;/P&gt;
&lt;P&gt;I have developed a driver for enabling direct communication between two PCIe devices. So, the default path is PCI A device -&amp;gt; Host Memory -&amp;gt; PCI B device. My framework allows the direct path PCI A -&amp;gt; PCI B. The problem is that while, the speedup in one direction (A-&amp;gt;B) yields significant speedup, in the other direction (B-&amp;gt;A) there is an order of magnitude drop down.&lt;/P&gt;
&lt;P&gt;I can't figure the reason of the above behavior (both device firmware are closed-source). The chipset of the setup is the 5520 (IOH-36D). I have checked out the manual (&lt;A href="http://www.intel.com/content/dam/doc/datasheet/5520-5500-chipset-ioh-datasheet.pdf"&gt;http://www.intel.com/content/dam/doc/datasheet/5520-5500-chipset-ioh-datasheet.pdf&lt;/A&gt;), but still no clue. So, I was wondering if there is an obvious reason for that. If you can sched some light for my case I would be glad.&lt;/P&gt;
&lt;P&gt;Thanks in advance!&lt;/P&gt;</description>
    <pubDate>Tue, 26 Mar 2013 16:32:51 GMT</pubDate>
    <dc:creator>S_G_2</dc:creator>
    <dc:date>2013-03-26T16:32:51Z</dc:date>
    <item>
      <title>Asymetrical performance between PCIe peer-to-peer devices</title>
      <link>https://community.intel.com/t5/Software-Tuning-Performance/Asymetrical-performance-between-PCIe-peer-to-peer-devices/m-p/981330#M3140</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;@admins: if this is a wrong place, feel free to move it to the relevant topic.&lt;/P&gt;
&lt;P&gt;I have developed a driver for enabling direct communication between two PCIe devices. So, the default path is PCI A device -&amp;gt; Host Memory -&amp;gt; PCI B device. My framework allows the direct path PCI A -&amp;gt; PCI B. The problem is that while, the speedup in one direction (A-&amp;gt;B) yields significant speedup, in the other direction (B-&amp;gt;A) there is an order of magnitude drop down.&lt;/P&gt;
&lt;P&gt;I can't figure the reason of the above behavior (both device firmware are closed-source). The chipset of the setup is the 5520 (IOH-36D). I have checked out the manual (&lt;A href="http://www.intel.com/content/dam/doc/datasheet/5520-5500-chipset-ioh-datasheet.pdf"&gt;http://www.intel.com/content/dam/doc/datasheet/5520-5500-chipset-ioh-datasheet.pdf&lt;/A&gt;), but still no clue. So, I was wondering if there is an obvious reason for that. If you can sched some light for my case I would be glad.&lt;/P&gt;
&lt;P&gt;Thanks in advance!&lt;/P&gt;</description>
      <pubDate>Tue, 26 Mar 2013 16:32:51 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Tuning-Performance/Asymetrical-performance-between-PCIe-peer-to-peer-devices/m-p/981330#M3140</guid>
      <dc:creator>S_G_2</dc:creator>
      <dc:date>2013-03-26T16:32:51Z</dc:date>
    </item>
  </channel>
</rss>

