<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>3389</id>
    <completedYear>2012</completedYear>
    <publishedYear>2012</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>74</pageFirst>
    <pageLast>83</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">On Improving the Performance of Multi-threaded CUDA Applications with Concurrent Kernel Execution by Kernel Reordering</title>
    <parentTitle language="eng">Application Accelerators in High Performance Computing (SAAHPC), 2012 Symposium on</parentTitle>
    <identifier type="doi">10.1109/SAAHPC.2012.12</identifier>
    <enrichment key="PeerReviewed">yes</enrichment>
    <author>Florian Wende</author>
    <author>Frank Cordes</author>
    <author>Thomas Steinke</author>
    <collection role="institutes" number="vas">Distributed Algorithms and Supercomputing</collection>
    <collection role="persons" number="steinke">Steinke, Thomas</collection>
    <collection role="projects" number="2011-ENHANCE">2011-ENHANCE</collection>
  </doc>
  <doc>
    <id>3577</id>
    <completedYear>2003</completedYear>
    <publishedYear>2003</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>4</pageFirst>
    <pageLast>5</pageLast>
    <pageNumber/>
    <edition/>
    <issue>2</issue>
    <volume>18</volume>
    <type>article</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">How Does a Protein Work?</title>
    <parentTitle language="eng">BioTOPics J. of Biotechnology Berlin-Brandenburg</parentTitle>
    <author>Frank Cordes</author>
    <author>Robert Preissner</author>
    <author>Thomas Steinke</author>
    <collection role="institutes" number="vas">Distributed Algorithms and Supercomputing</collection>
    <collection role="persons" number="steinke">Steinke, Thomas</collection>
  </doc>
  <doc>
    <id>5131</id>
    <completedYear/>
    <publishedYear>2014</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>788</pageFirst>
    <pageLast>799</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume>8632</volume>
    <type>conferenceobject</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Concurrent Kernel Execution on Xeon Phi within Parallel Heterogeneous Workloads</title>
    <parentTitle language="eng">Euro-Par 2014: Parallel Processing.  20th International Conference, Porto, Portugal, August 25-29, 2014, Proceedings</parentTitle>
    <identifier type="url">http://www.springer.com/computer/swe/book/978-3-319-09872-2</identifier>
    <identifier type="doi">10.1007/978-3-319-09873-9_66</identifier>
    <enrichment key="Series">Lecture Notes in Computer Science</enrichment>
    <enrichment key="PeerReviewed">yes</enrichment>
    <author>Florian Wende</author>
    <submitter>Regine Kossick</submitter>
    <author>Frank Cordes</author>
    <author>Thomas Steinke</author>
    <collection role="institutes" number="vas">Distributed Algorithms and Supercomputing</collection>
    <collection role="persons" number="steinke">Steinke, Thomas</collection>
    <collection role="projects" number="2013-Many-Core-HPC">2013-Many-Core-HPC</collection>
  </doc>
  <doc>
    <id>5134</id>
    <completedYear/>
    <publishedYear>2014</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst/>
    <pageLast/>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">A Unified Programming Model for Intra- and Inter-Node Offloading on Xeon Phi Clusters</title>
    <abstract language="eng">Standard offload programming models for the Xeon Phi, e.g. Intel LEO and OpenMP 4.0, are restricted to a single compute node and hence a limited number of coprocessors. Scaling applications across a Xeon Phi cluster/supercomputer thus requires hybrid programming approaches, usually MPI+X. In this work, we present a framework based on heterogeneous active messages (HAM-Offload) that provides the means to offload work to local and remote (co)processors using a unified offload API. Since HAM-Offload provides similar primitives as current local offload frameworks, existing applications can be easily ported to overcome the single-node limitation while keeping the convenient offload programming model. We demonstrate the effectiveness of the framework by using it to enable a real-world application from the field of molecular dynamics to use multiple local and remote Xeon Phis. The evaluation shows good scaling behavior. Compared with LEO, performance is equal for large offloads and significantly better for small offloads.</abstract>
    <parentTitle language="eng">SC '14: Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis. SC14, November 16-21, 2014, New Orleans, Louisiana, USA</parentTitle>
    <identifier type="url">http://dl.acm.org/citation.cfm?id=2683616</identifier>
    <identifier type="doi">10.1109/SC.2014.22</identifier>
    <enrichment key="PeerReviewed">yes</enrichment>
    <author>Matthias Noack</author>
    <submitter>Regine Kossick</submitter>
    <author>Florian Wende</author>
    <author>Thomas Steinke</author>
    <author>Frank Cordes</author>
    <collection role="institutes" number="vas">Distributed Algorithms and Supercomputing</collection>
    <collection role="persons" number="steinke">Steinke, Thomas</collection>
    <collection role="projects" number="2013-Many-Core-HPC">2013-Many-Core-HPC</collection>
  </doc>
  <doc>
    <id>5036</id>
    <completedYear/>
    <publishedYear/>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst/>
    <pageLast/>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>reportzib</type>
    <publisherName/>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>2014-02-06</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">Multi-threaded Kernel Offloading to GPGPU Using Hyper-Q on Kepler Architecture</title>
    <abstract language="eng">Small-scale computations usually cannot fully utilize the compute capabilities of modern GPGPUs. With the Fermi GPU architecture Nvidia introduced the concurrent kernel execution feature allowing up to 16 GPU kernels to execute simultaneously on a shared GPU device for a better utilization of the respective resources. Insufficient scheduling capabilities in this respect, however, can significantly reduce the theoretical concurrency level. With the Kepler GPU architecture Nvidia addresses this issue by introducing the Hyper-Q feature with 32 hardware managed work queues for concurrent kernel execution.&#13;
We investigate the Hyper-Q feature within heterogeneous workloads with multiple concurrent host threads or processes offloading computations to the GPU each.&#13;
By means of a synthetic benchmark kernel and a hybrid parallel CPU-GPU real-world application, we evaluate the performance obtained with Hyper-Q on GPU and compare it against a kernel reordering mechanism introduced by the authors for the Fermi architecture.</abstract>
    <identifier type="issn">1438-0064</identifier>
    <identifier type="urn">urn:nbn:de:0297-zib-50362</identifier>
    <author>Florian Wende</author>
    <submitter>Florian Wende</submitter>
    <author>Thomas Steinke</author>
    <author>Frank Cordes</author>
    <series>
      <title>ZIB-Report</title>
      <number>14-19</number>
    </series>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>GPGPU</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Hyper-Q</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>Concurrent Kernel Execution</value>
    </subject>
    <collection role="ccs" number="B.">Hardware</collection>
    <collection role="pacs" number="80.00.00">INTERDISCIPLINARY PHYSICS AND RELATED AREAS OF SCIENCE AND TECHNOLOGY</collection>
    <collection role="msc" number="00-XX">GENERAL</collection>
    <collection role="institutes" number="vas">Distributed Algorithms and Supercomputing</collection>
    <collection role="persons" number="steinke">Steinke, Thomas</collection>
    <collection role="projects" number="2013-Many-Core-HPC">2013-Many-Core-HPC</collection>
    <file>https://opus4.kobv.de/opus4-zib/files/5036/ZR-14-19.pdf</file>
  </doc>
</export-example>
