<?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 Re: Cache Coherence Granularity in Mobile and Desktop Processors</title>
    <link>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598616#M33040</link>
    <description>&lt;P&gt;Hi benlong,&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;The LLC operates on 64 byte cache lines.  It can tag 2 lines together for efficiency in rare circumstances, but the basic architecture is 64 byte lines.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Thanks&lt;/P&gt;&lt;P&gt;Allan&lt;/P&gt;</description>
    <pubDate>Thu, 05 Dec 2013 17:21:43 GMT</pubDate>
    <dc:creator>Allan_J_Intel1</dc:creator>
    <dc:date>2013-12-05T17:21:43Z</dc:date>
    <item>
      <title>Cache Coherence Granularity</title>
      <link>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598615#M33039</link>
      <description>&lt;P&gt;Hi all,&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Empirically the cache line size of most processors is believed to be 64 Bytes. And it is indeed confirmed&lt;/P&gt;&lt;P&gt;in my Linux machine by checking out the file /proc/cpuinfo. &lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;But my question is, what's the basic size of the cache coherence protocol inside the processor? &lt;/P&gt;&lt;P&gt;Supposing a MSI protocol, is it true that a single cache line is tagged as, say, "M" state, or, they&lt;/P&gt;&lt;P&gt;are tagged in a bigger granularity, such as 2 or 4 lines are packed together to be marked in the "M" &lt;/P&gt;&lt;P&gt;state?  Because according to my experience, it seems that the &lt;B&gt;Xeon(R) CPU E5-2650&lt;/B&gt; processor&lt;/P&gt;&lt;P&gt;tags two cache lines together.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Thank you very much.&lt;/P&gt;</description>
      <pubDate>Wed, 04 Dec 2013 14:53:44 GMT</pubDate>
      <guid>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598615#M33039</guid>
      <dc:creator>bzhan16</dc:creator>
      <dc:date>2013-12-04T14:53:44Z</dc:date>
    </item>
    <item>
      <title>Re: Cache Coherence Granularity</title>
      <link>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598616#M33040</link>
      <description>&lt;P&gt;Hi benlong,&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;The LLC operates on 64 byte cache lines.  It can tag 2 lines together for efficiency in rare circumstances, but the basic architecture is 64 byte lines.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Thanks&lt;/P&gt;&lt;P&gt;Allan&lt;/P&gt;</description>
      <pubDate>Thu, 05 Dec 2013 17:21:43 GMT</pubDate>
      <guid>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598616#M33040</guid>
      <dc:creator>Allan_J_Intel1</dc:creator>
      <dc:date>2013-12-05T17:21:43Z</dc:date>
    </item>
    <item>
      <title>Re: Cache Coherence Granularity</title>
      <link>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598617#M33041</link>
      <description>&lt;P&gt;Hi Allan,&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Then what's the implication by "in rare circumstances", does it mean that the LLC can work in two modes, or that &lt;/P&gt;&lt;P&gt;most processors tag 64 Bytes but rare processors do 128 Bytes?&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;And what about the L1 and L2 cache ?&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Thank you very much.&lt;/P&gt;&lt;P&gt;Benlong&lt;/P&gt;</description>
      <pubDate>Fri, 06 Dec 2013 08:11:13 GMT</pubDate>
      <guid>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598617#M33041</guid>
      <dc:creator>bzhan16</dc:creator>
      <dc:date>2013-12-06T08:11:13Z</dc:date>
    </item>
    <item>
      <title>Re: Cache Coherence Granularity</title>
      <link>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598618#M33042</link>
      <description>&lt;P&gt;Hi benlong,&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;It means that in some circumstances, the processor will tag 2 lines at once for efficiency.  It is not CPU model specific and is not common.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;L1 and L2 always work at 64 byte.&lt;/P&gt;&lt;P&gt;&lt;/P&gt;&lt;P&gt;Allan&lt;/P&gt;</description>
      <pubDate>Fri, 06 Dec 2013 18:23:11 GMT</pubDate>
      <guid>https://community.intel.com/t5/Mobile-and-Desktop-Processors/Cache-Coherence-Granularity/m-p/598618#M33042</guid>
      <dc:creator>Allan_J_Intel1</dc:creator>
      <dc:date>2013-12-06T18:23:11Z</dc:date>
    </item>
  </channel>
</rss>

