<div dir="ltr">I just found <a href="http://aces.snu.ac.kr/software/snu-npb/">http://aces.snu.ac.kr/software/snu-npb/</a> today, but have not tried it.<div><br></div><div>Jeff</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Oct 10, 2017 at 7:37 AM, Jeff Hammond <span dir="ltr"><<a href="mailto:jeff.science@gmail.com" target="_blank">jeff.science@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="auto"><a href="https://github.com/ParRes/Kernels/tree/master/Cxx11" target="_blank">https://github.com/ParRes/<wbr>Kernels/tree/master/Cxx11</a> <wbr>implements at least two kernels in OpenCL, GPU-oriented OpenMP4 (#pragma omp <span style="background-color:rgba(255,255,255,0)">teams distribute parallel <span class="m_-771006556969918392pl-k" style="box-sizing:border-box">for</span> simd <span class="m_-771006556969918392pl-smi" style="box-sizing:border-box">collapse</span>(<span class="m_-771006556969918392pl-c1" style="box-sizing:border-box">2</span>) <span class="m_-771006556969918392pl-c1" style="box-sizing:border-box">schedule</span>(<span class="m_-771006556969918392pl-k" style="box-sizing:border-box">static</span>,<span class="m_-771006556969918392pl-c1" style="box-sizing:border-box">1</span>) as recommended by NVIDIA), RAJA, Kokkos, and numerous CPU implementations.</span><div><span style="background-color:rgba(255,255,255,0)"><br></span></div><div><span style="background-color:rgba(255,255,255,0)">You’ll need to (1) tune the OpenCL code for your GPU as it is currently not optimized for any architecture and (2) tweak the RAJA and KOKKOS implementations to target GPU models in addition to CPU ones. Neither should be very difficult and the PRK team is happy to provide assistance.</span></div><div><span style="background-color:rgba(255,255,255,0)"><br></span></div><div>That project contains one C++ CUDA and two Fortran OpenACC implementations but these are unfinished and need work. However, I’m very happy to see others contribute better versions.  It may not be the easiest route to fame and fortune, but it’s not the worst either 😂</div><div><div><div><div><br></div><div>Write me privately or create GitHub issues if you have questions. I am currently on leave from my day job but will be responsive via GMail and GitHub.</div><div><br></div><div>Sorry if this is a repost. I intended to reply early but didn’t and can’t find it if I did, but mistakes have been known to occur.</div><div><br></div><div>Jeff<br><br><div id="m_-771006556969918392AppleMailSignature">Sent from my iPhone</div><div><div class="h5"><div><br>On Oct 10, 2017, at 6:56 AM, Jonas Hahnfeld via Openmp-dev <<a href="mailto:openmp-dev@lists.llvm.org" target="_blank">openmp-dev@lists.llvm.org</a>> wrote:<br><br></div><blockquote type="cite"><div><span>Hi,</span><br><span></span><br><span>some (public) benchmarks that come to my mind:</span><br><span> * LULESH: <a href="https://codesign.llnl.gov/lulesh.php" target="_blank">https://codesign.llnl.gov/<wbr>lulesh.php</a></span><br><span> * CloverLeaf: <a href="http://uk-mac.github.io/CloverLeaf/" target="_blank">http://uk-mac.github.io/<wbr>CloverLeaf/</a></span><br><span> * my own, a Conjugate Gradient solver: <a href="https://github.com/hahnjo/CGxx" target="_blank">https://github.com/hahnjo/CGxx</a></span><br><span></span><br><span>As James wrote, SPEC ACCEL also has an OpenMP suite, but you need a license.</span><br><span></span><br><span>Regards</span><br><span>Jonas</span><br><span></span><br><span>Am 2017-10-10 09:43, schrieb Daniel Schürmann via Openmp-dev:</span><br><blockquote type="cite"><span>Hello together,</span><br></blockquote><blockquote type="cite"><span>as the title states, I am looking for benchmarks which are</span><br></blockquote><blockquote type="cite"><span>particularly suitable for GPU accelerators (or at least make use of</span><br></blockquote><blockquote type="cite"><span>the #teams pragma).</span><br></blockquote><blockquote type="cite"><span>I already tried rodinia benchmark suite, but they seem to be written</span><br></blockquote><blockquote type="cite"><span>for CPU acceleration only.</span><br></blockquote><blockquote type="cite"><span>I would be very pleased if someone could provide me with one or more</span><br></blockquote><blockquote type="cite"><span>kernels which can be used with the NVPTX backend.</span><br></blockquote><blockquote type="cite"><span>The purpose is a master thesis about using OpenCL and SPIR-V as OpenMP backend.</span><br></blockquote><blockquote type="cite"><span>Thank you in advance and</span><br></blockquote><blockquote type="cite"><span>kind regards,</span><br></blockquote><blockquote type="cite"><span>Daniel</span><br></blockquote><blockquote type="cite"><span>______________________________<wbr>_________________</span><br></blockquote><blockquote type="cite"><span>Openmp-dev mailing list</span><br></blockquote><blockquote type="cite"><span><a href="mailto:Openmp-dev@lists.llvm.org" target="_blank">Openmp-dev@lists.llvm.org</a></span><br></blockquote><blockquote type="cite"><span><a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/openmp-dev" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/openmp-dev</a></span><br></blockquote><span>______________________________<wbr>_________________</span><br><span>Openmp-dev mailing list</span><br><span><a href="mailto:Openmp-dev@lists.llvm.org" target="_blank">Openmp-dev@lists.llvm.org</a></span><br><span><a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/openmp-dev" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/openmp-dev</a></span><br></div></blockquote></div></div></div></div></div></div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br><div class="gmail_signature" data-smartmail="gmail_signature">Jeff Hammond<br><a href="mailto:jeff.science@gmail.com" target="_blank">jeff.science@gmail.com</a><br><a href="http://jeffhammond.github.io/" target="_blank">http://jeffhammond.github.io/</a></div>
</div>