<div dir="ltr">On Mon, Mar 6, 2017 at 10:55 AM, Roland Haas <span dir="ltr"><<a href="mailto:rhaas@illinois.edu" target="_blank">rhaas@illinois.edu</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hello all,<br>
<span class=""><br>
> We quickly discussed the failing tests, and given that the option list<br>
> used<br>
> (<a href="https://bitbucket.org/ianhinder/cactusjenkins/raw/d7021a52bd83448db589b2346c43441682eecabb/build.cfg" rel="noreferrer" target="_blank">https://bitbucket.org/<wbr>ianhinder/cactusjenkins/raw/<wbr>d7021a52bd83448db589b2346c4344<wbr>1682eecabb/build.cfg</a>)<br>
> is not using -ffast-math or -mmarch=native the working assumption is<br>
> that the occurrence of the failures is due to to an updated OS and<br>
> newer compiler. Erik asked if we had a docker container for the<br>
> Jenkins test slave, which we were not sure (there is a docker image at<br>
> <a href="https://bitbucket.org/ianhinder/et-jenkins-slave" rel="noreferrer" target="_blank">https://bitbucket.org/<wbr>ianhinder/et-jenkins-slave</a> though it is not<br>
> clear if this is the one used).<br>
<br>
</span>after the call of course I realized what may cause differences even<br>
without -march=native or -ffast-math: the explicit vectorization in<br>
thorn Vectors depends on the machine architecture so different hardware<br>
(eg AVX2) will produce different vector code and thus different<br>
answers. Though given the thorns affected (GRHydro and<br>
SphericalHarmonicRecon) I doubt that it is the thorn's code itself but<br>
rather eg McLachlan or the fftw library (for the PITTNull code). This<br>
*may* cause differences though a different compiler *will* cause<br>
differences (in particular a different Fortran compiler).</blockquote></div><br>That's a good point.</div><div class="gmail_extra"><br></div><div class="gmail_extra">-erik<br clear="all"><div><br></div>-- <br><div class="gmail_signature" data-smartmail="gmail_signature">Erik Schnetter <<a href="mailto:schnetter@cct.lsu.edu" target="_blank">schnetter@cct.lsu.edu</a>><br><a href="http://www.perimeterinstitute.ca/personal/eschnetter/" target="_blank">http://www.perimeterinstitute.ca/personal/eschnetter/</a></div>
</div></div>