<div dir="ltr"><div><div><div><div><div><div><div><div>Dear Ezzat,<br><br></div>Regarding your two questions:<br></div>1. How to make GATE work with GPU: <a href="http://wiki.opengatecollaboration.org/index.php/Compilation_Instructions_V7.2#GPU_.26_CUDA_tools" target="_blank">http://wiki.opengatecollaborat<wbr>ion.org/index.php/Compilation_<wbr>Instructions_V7.2#GPU_.26_CUDA<wbr>_tools</a><br>In the french link, I go to the 14.04 or 16.04 version: <a href="http://doc.ubuntu-fr.org/cuda" target="_blank">http://doc.ubuntu-fr.org/cuda</a> (not 11.4, but maybe you are in a different page).<br>Before you use GATE for GPU, you need to download 2 packages which are the CUDA toolkit provided by NVIDIA and the GPU computing SDK. <br></div><div>However, as far as I understood SDK comes together with the toolkit now (<a href="https://developer.nvidia.com/cuda-toolkit" target="_blank">https://developer.nvidia.com/<wbr>cuda-toolkit</a>), therefore in cmake they should point to the same folder. <br>Cuda toolkit instructions here: <a href="https://developer.nvidia.com/cuda-downloads" target="_blank">https://developer.nvidia.com/c<wbr>uda-downloads</a><br><br></div>2. There are many ways to increase the simulation performance besides GPUs:<br>- output, e.g. ROOT output is usually faster than ASCII output. Only save the events that you are interested in (Hits, Singles, Coincidences).<br></div>- using cuts <br></div>- not using voxelized phantoms unless necessary<br></div>- using cluster tools, if you have access to several CPUs.<br><br></div><div>I hope this helps a little bit.<br></div><div><br></div>Cheers,<br></div>Liliana<br><div><div><div><br></div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On 15 January 2017 at 10:44, Ezzat Elmo <span dir="ltr"><<a href="mailto:ezzat.elmo@gmail.com" target="_blank">ezzat.elmo@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="ltr">Good day everyone,<div><br></div><div>so I've been using v-Gate for over a year now and few weeks ago I decided to start using the full version on a Linux based system hoping it will improve the simulation speed.</div><div><br></div><div>it did work, but the improvement was about just 10% and i was hoping it would have been more.</div><div>so i was interested to run it on the GPU to see how much it will improve the speed.</div><div><br></div><div>so i have two questions:</div><div>1- how to make gate work using GPU? i tried to work it out following the wiki page, but the cuda section is in France and when I translate it and followed it turns out its for Ubuntu 11.4 while I'm using Ubuntu 14.04.</div><div><br></div><div>2- is there any other way than the GPU to accelerate the simulation performance ?</div><div><br></div><div>Thanks for your time and looking forward to here back from you.</div><span class="m_1111685304666873539HOEnZb"><font color="#888888"><div><br></div><div>Ezzat</div></font></span></div>
<br>______________________________<wbr>_________________<br>
Gate-users mailing list<br>
<a href="mailto:Gate-users@lists.opengatecollaboration.org" target="_blank">Gate-users@lists.opengatecolla<wbr>boration.org</a><br>
<a href="http://lists.opengatecollaboration.org/mailman/listinfo/gate-users" rel="noreferrer" target="_blank">http://lists.opengatecollabora<wbr>tion.org/mailman/listinfo/<wbr>gate-users</a><br></blockquote></div><br></div></div>