Sophie: hoard-3.7.1-1mdv2009.1 i586

hoard-3.7.1-1mdv2009.1.i586.rpm

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE document PUBLIC "-//APACHE//DTD Documentation V2.0//EN" "http://forrest.apache.org/dtd/document-v20.dtd">
<document> 

  <header> 
    <title>Screenshots</title> 
  </header> 

<body>

  <p>
  Because Hoard is a transparent plug-in replacement, ordinary
  screenshots don't make much sense, since the applications look and
  work the same! However, it's easy to visualize Hoard's impact on
  application performance. Here are a few example speedup
  graphs. Ideally, the speedup as the number of processors increases
  is a straight diagonal line going from the lower left of each graph
  to the upper right. These benchmarks compare Hoard to Ptmalloc (the
  Linux allocator), mtmalloc version 8, and the Solaris default
  allocator, and are run on a 14-processor Sun.
</p>

  <table>
  <tr>
    <td>
    <!-- <center> -->
    <img alt="Threadtest graph" src="images/plot-threadtest-speedup.gif"/>
    <!-- </center> -->
    </td>
  </tr>
  <tr>
    <td>
    <!-- <center> -->
<p>
    This benchmark has multiple threads, each allocating and freeing
    objects. It stresses raw scalability of allocation in individual
    threads. In this case, Hoard achieves a linear speedup. Ptmalloc
    scales to four processors but then runs out of steam.
</p>
    <!-- </center> -->
    </td>
  </tr>
  <tr>
    <td>
    <!-- <center> -->
    <img alt="Larson graph" src="images/plot-larson-speedup.gif"/>
    <!-- </center> -->
    </td>
  </tr>
  <tr>
    <td>
    <!-- <center> -->
<p>
    In this benchmark, multiple threads allocate objects and pass some
    objects to other threads that free them. This stresses the ability
    of the memory manager to handle object sharing. Only Hoard scales
    for this case.
</p>
    <!-- </center> -->
    </td>
  </tr>
  <tr>
    <td>
    <!-- <center> -->
    <img alt="Cache-scratch graph" src="images/plot-cache-scratch-speedup.gif"/>
    <!-- </center> -->
    </td>
  </tr>
  <tr>
    <td>
    <!-- <center> -->
<p>
    A test of whether a memory manager avoids causing false sharing of
    cache lines (an insidious effect that destroys scalability). The
    linear speedup shows that Hoard avoids false sharing completely.
</p>
    <!-- </center> -->
    </td>
  </tr>
  </table>

</body>
</document>