<html><head></head><body><div class="gmail_quote">Am 18. September 2017 16:46:55 MESZ schrieb Roland Haas <rhaas@illinois.edu>:<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<pre class="k9mail">Present: Bill, Chris, Etik, Gabrielle, Sam, Steve, Yosef, Roland, Qian,<br />Ian, Vassili, Peter, <br /><br />Tutorial machine at NCSA:<br />* have final confirmation, about to deploy instance, should happen<br /> this week<br />* contact current outstanding requests about current status<br />* Roland suggests to take down tutorial page using QueenBee and instead<br /> point to Simplified tutorial instead for the time being<br /><br />Working groups<br />* need to pick this up, last action was sending around working group<br /> document<br />* Ian suggests working group for lower level infrastructure in the ET<br />* add to next week's agenda<br /><br />ET workshop in Mallorca:<br />* creating video of early days of Cactus, request old photos<br /><br />Supercomputing:<br />* Steve is going, will have poster about Cactus<br />* not ET related demo planned, no planned use of Cactus<br />* NCSA has demo (Wolfram and NVIDIA) about deep learning with Daniel<br /> George and Eliu Huerta<br /><br />KNL status update<br />* Peter has his cosmology code running on KNL, gets fairly decent<br /> performance<br />* this code is quite different from eg McLachlan<br /><br />Yours,<br />Roland<br /></pre></blockquote></div><br clear="all">Hi<br>
<br>
Concerning supercomputing and an ET demo: in the past we used a neutron star simulation on a Raspberry pi cluster and instant jpeg slices for visualization. It was slow, but worked. With the new pis we could try either the same, or a very low resolution binary neutron star simulation (but that might still be too much, even with the new pis), just for demonstration. Would there be interest? The code still worked on the pis last time I tried, earlier this year. <br>
<br>
I would volunteer in trying some parameter files, but would need a connection to the (full) new cluster.<br>
<br>
Frank<br>
</body></html>