<div dir="ltr">Dear Filippo Spiga<br>Thank you very much for you response.<div>1. You write that there is no benefit in using GPU with small systems (less than 40-50 atoms). What about phonon calculations? </div><div>2. We want to do a much more complicated calculation but we do not succeed even with the very simple we write about before.</div><div>3. We try to work with a simple GPU for now, but we intend to work with a better one, but we need to try it first. </div><div>4. What can we do to run a more complicated calculations on the present GPU?<br><div><br></div></div><div>Uri Argaman</div><div>Ben-Gurion University</div><div>Israel</div><div><br></div><blockquote style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex" class="gmail_quote"><span style="color:rgb(0,0,0)">Dear Uri,<br></span>
you are indeed right, the calculation is simple and it should work. There are many think that could have done wrong (e.g FFT grid or the multi-plan CUFFT driver that has not enough space on the card to initialize the FFT plan). However your case is so simple that adding GPU will not bring you any benefit. The QE-GPU is designed and has born to accelerate time-to-solution for mid/large systems. If you have less than 40~50 atoms (unless your case is particularly computational expensive) then QE-GPU will not bring you any benefit.<br>
Save you time, keep using the normal QE that works great.<br>
Cheers,<br>F</blockquote></div>