<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <title>RE: How much memory per CPU?</title>
  <link rel="alternate" href="https://conferences.xsede.org/c/message_boards/find_recent_posts?p_l_id=" />
  <subtitle>RE: How much memory per CPU?</subtitle>
  <entry>
    <title>RE: How much memory per CPU?</title>
    <link rel="alternate" href="https://conferences.xsede.org/c/message_boards/find_message?p_l_id=&amp;messageId=610772" />
    <author>
      <name>William L Barth</name>
    </author>
    <id>https://conferences.xsede.org/c/message_boards/find_message?p_l_id=&amp;messageId=610772</id>
    <updated>2013-10-18T18:55:41Z</updated>
    <published>2013-10-18T18:55:19Z</published>
    <summary type="html">The --mem=XX option to SLURM is designed to help you select among nodes in the same partition (queue) that have different amounts of memory. Stampede partitions are all homogenous, so there&amp;#039;s no need to use this option for your jobs. &lt;br /&gt;&lt;br /&gt;In the serial partition (queue), every node (which has 16 cores) has 32GB of memory. You should be able to use about 31 GB of that with a typical job. &lt;br /&gt;&lt;br /&gt;If you need more than that, we have 16 nodes in the largemem queue which each have 1TB of RAM. These are pretty heavily requested, but you can run there if you need more than 32GB/node.</summary>
    <dc:creator>William L Barth</dc:creator>
    <dc:date>2013-10-18T18:55:19Z</dc:date>
  </entry>
</feed>

