<div dir="ltr">On Mon, May 21, 2018 at 2:32 PM, Pawel Dziekonski <span dir="ltr"><<a href="mailto:pawel.dziekonski@wcss.pl" target="_blank">pawel.dziekonski@wcss.pl</a>></span> wrote:<br><br><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">OpenMPI 2.1.3, GCC 7.3.0, netlib BLAS/LAPACK:<br>
./configure --enable-openmp --enable-parallel --with-netlib<br>
CP:<br>
<a href="https://pastebin.com/x4dqkxCA" rel="noreferrer" target="_blank">https://pastebin.com/x4dqkxCA</a></blockquote><div><br></div><div>Small numerical differences. Different mathematical libraries and execution modes may lead to slightly different results.<br> <br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
PW:<br>
<a href="https://pastebin.com/hgjMuYCK" rel="noreferrer" target="_blank">https://pastebin.com/hgjMuYCK</a></blockquote><div><br></div><div>The "Langevin" test gives results that depend upon the number of processors used and should be discarded.<br></div><div>The "Berry" and many "vdw" tests fail for reasons that is impossible to know without looking at the outputs<br></div><div>All the remaining cases fall in the "small numerical differences" case.<br></div><div>
<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Intel parallel studio xe 2016.3.067 (I'm aware of MKL bug):<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
./configure --enable-parallel<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
CP:<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<a href="https://pastebin.com/8F49sL6u" rel="noreferrer" target="_blank">https://pastebin.com/8F49sL6u</a><br></blockquote><br>Same as above for CP<br><br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">PW:<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<a href="https://pastebin.com/njFEmt0M" rel="noreferrer" target="_blank">https://pastebin.com/njFEmt0M</a><br></blockquote>
<br>same as above for PW<br><br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">BTW, when I start compilation by<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">make all<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">only some targets are build (mostly cp, nothing from pwall).<br></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Is this expected?<br></blockquote><div><br></div><div>no<br><br></div><div><font color="#888888">Paolo</font><span class="m_7730770486370054269gmail-HOEnZb"></span><br></div><span class="m_7730770486370054269gmail-HOEnZb"></span></div></div>-- <br><div class="m_7730770486370054269gmail_signature"><div dir="ltr"><div><div dir="ltr"><div>Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,<br>Univ. Udine, via delle Scienze 208, 33100 Udine, Italy<br>Phone +39-0432-558216, fax +39-0432-558222<br><br></div></div></div></div></div>
</div></div>