<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">Dear Antoine,<div class=""><br class=""></div><div class="">yes this is expected. I have seen the same on my macbook, GATE being slightly faster in the vGATE compared to the host (MacOS) with the standard compiling options.</div><div class=""><br class=""></div><div class="">Virtual machines are quite good at using the full CPU power, it is a different story for GPU…</div><div class=""><br class=""></div><div class="">So in the end, the only thing that matters (VM or not) is the clock speed of your CPU. You can slightly increase the performance with compiler options related to your CPU architecture. </div><div class=""><br class=""></div><div class="">For example, we have seen an increase of performance of ~20% on an HPC (<a href="https://www.calmip.univ-toulouse.fr/" class="">https://www.calmip.univ-toulouse.fr/</a>) with the compiler option "-march=skylake-avx512” for Intel Skylake processors. <b class="">For this you need to compile both Geant4 and GATE with the compiler optimisation option.</b><br class=""><div><br class=""></div><div>Of course we are talking about performance on a single CPU. You can speed up your simulation by running several sub-simulations on many CPUs.</div><div><br class=""></div><div>Best regards,</div><div> Maxime Chauvin</div><div><br class=""><blockquote type="cite" class=""><div class="">On 10 Jan 2020, at 10:38, Antoine Merlet <<a href="mailto:ant.merlet@gmail.com" class="">ant.merlet@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div dir="ltr" class="">Dear Gate users,<div class=""><br class=""></div><div class="">I have compiled the latest develop version of Gate with Geant4-10.05.p01 under <span class="" id=":c7.28" tabindex="-1" style="">macOS</span> Mojave. Previously using <span class="" id=":c7.29" tabindex="-1" style="">vGate</span> on the same <span class="" id=":c7.30" tabindex="-1" style="">MacBook</span>, I was hoping for decreased simulation time after removing the virtual machine intermediary. However, tests shows nearly no differences - <span class="" id=":c7.31" tabindex="-1" style="">vGate</span> is even a bit faster - regardless of the simulation duration (minutes / dozens of hours). </div><div class=""><br class=""></div><div class="">Therefore, I would like to know if anyone had similar experiences, and if this kind of results are to be expected. Also, is there any way of increasing the performances of Gate on <span class="" id=":c7.32" tabindex="-1" style="">macOS</span> by optimizing the compilation parameters?</div><div class=""><br class=""></div><div class="">You can find enclosed a simple test file which has been adapted from <span class="" id=":c7.33" tabindex="-1" style="">GateContrib</span> Cylindrical PET example (<a href="https://github.com/OpenGATE/GateContrib/tree/master/imaging/PET" target="_blank" class="">https://<span class="" id=":c7.34" tabindex="-1" style="">github</span>.com/<span class="" id=":c7.35" tabindex="-1" style="">OpenGATE</span>/<span class="" id=":c7.36" tabindex="-1" style="">GateContrib</span>/tree/master/imaging/PET</a>). The resulting simulation time is 8min44s when using <span class="" id=":c7.37" tabindex="-1" style="">vGate</span> and 8min50sec when using Gate under <span class="" id=":c7.38" tabindex="-1" style="">macOS</span>.</div><div class=""><br class=""></div><div class="">Best regards,</div><div class="">Antoine</div></div>
<span id="cid:f_k57yqolm0"><main.mac></span><span id="cid:f_k57yqrud1"><GateMaterials.db></span>_______________________________________________<br class="">Gate-users mailing list<br class=""><a href="mailto:Gate-users@lists.opengatecollaboration.org" class="">Gate-users@lists.opengatecollaboration.org</a><br class="">http://lists.opengatecollaboration.org/mailman/listinfo/gate-users</div></blockquote></div><br class=""></div></body></html>