<?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>Using lfs setstripe on stampede</title>
  <link rel="alternate" href="https://conferences.xsede.org/c/message_boards/find_recent_posts?p_l_id=" />
  <subtitle>Using lfs setstripe on stampede</subtitle>
  <entry>
    <title>Using lfs setstripe on stampede</title>
    <link rel="alternate" href="https://conferences.xsede.org/c/message_boards/find_message?p_l_id=&amp;messageId=545976" />
    <author>
      <name>Francois Gygi</name>
    </author>
    <id>https://conferences.xsede.org/c/message_boards/find_message?p_l_id=&amp;messageId=545976</id>
    <updated>2013-06-25T18:33:54Z</updated>
    <published>2013-06-25T06:12:24Z</published>
    <summary type="html">When writing a 15 GB file on the stampede $SCRATCH file system, our application (Qbox) achieves an aggregate write rate of 85 MB/s (writing from 1024 tasks using MPI-IO). On ranger, we used to improve this considerably by setting the stripe count of the file (before writing it) using lfs setstripe -c 32 filename.&lt;br /&gt;We tried this on stampede, but it causes the MPI-IO call to hang and the resulting file size is zero. &lt;br /&gt;Does anyone know how to use striped files on stampede? &lt;br /&gt;Thanks</summary>
    <dc:creator>Francois Gygi</dc:creator>
    <dc:date>2013-06-25T06:12:24Z</dc:date>
  </entry>
</feed>

