Resource Menu


posted by Alessandro Vicini at Jun 21, 2016 9:54 AM
Quote
Ok, it seems that the first step goes well and that it freezes shortly after. You have secondary electrons in your simulation. Could you try to decrease the electronSecondaryDensification and protonSecondaryDensification values?

Yes, this problem was solved by decreasing the densification factors…

posted by Sébastien Hess at Jun 21, 2016 9:51 AM
Quote
Ok, it seems that the first step goes well and that it freezes shortly after. You have secondary electrons in your simulation. Could you try to decrease the electronSecondaryDensification and protonSecondaryDensification values?
posted by Alessandro Vicini at Jun 9, 2016 10:56 AM
Quote
Hello Alessandro,

there may be many reasons for such a problem and it is hard to diagnose it with so few information. One thing may be a memory problem. In the SPIS launching script there are commands to start SPIS that contain an option -Xmx0000M, with some number instead of 0000. By default the allocated memory is low (1Gb), you may want to increase it. If you already did so or if it does not work, try to increase the level of verbosity (in the global parameters) and post the last lines that were displayed in the console. It may provide some clues.

Regards,

Sébastien

Hullo Sebastien, thanks for your reply. I doubt it can be a memory problem, the mesh has approximately 70000 tetrahedra, and all plasma parameters are left to default values… But I will also give a look to that.

This is what I see in the console:

20000 Thu Jun 09 10:45:42 CEST 2016 Last local volume potential change: 0.0V maximum change, and 0.0V average change 20000 Thu Jun 09 10:45:42 CEST 2016 (last plasma time step = 0.0s) 20000 Thu Jun 09 10:45:42 CEST 2016 20000 Thu Jun 09 10:45:42 CEST 2016 Time: 0.0, Dt = 0.01 s 20000 Thu Jun 09 10:45:42 CEST 2016 20000 Thu Jun 09 10:47:04 CEST 2016 Warning: sampling requested from PICSurfDistrib with zero accumulation duration (normal at first simulation step) 20000 Thu Jun 09 10:47:04 CEST 2016 Warning: sampling requested from PICSurfDistrib with zero accumulation duration (normal at first simulation step) 20000 Thu Jun 09 10:47:04 CEST 2016 after SC integration TIME: 0.01, dt = 0.01 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 In this time interval: 20000 Thu Jun 09 10:47:06 CEST 2016 COLLECTED current: 20000 Thu Jun 09 10:47:06 CEST 2016 -4.910694E-7 A total, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 -4.910694E-7 A on node0, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 EMITTED current: 20000 Thu Jun 09 10:47:06 CEST 2016 -9.810205E-6 A total, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 -9.810205E-6 A on node0, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 Time 0.01 s => POTENTIALS: 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 0.093166724 V on SC ground (node0), 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 0.09316669 V on top of node0, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 0.093166724 Vmin onTopOfNode0, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 0.093166724 Vmax onTopOfNode0, 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 20000 Thu Jun 09 10:47:06 CEST 2016 Monitoring numerics of simulation at time 0.01 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib ions1: 174955 particles, last numerical time step used = 1.0E-5s (max allowed = 1.0E-4s) 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib ions2: 258843 particles, last numerical time step used = 1.0E-5s (max allowed = 1.0E-4s) 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib elec2: 444507 particles, last numerical time step used = 1.0E-6s (max allowed = 1.0E-6s) 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib photoElec: 305719 particles, last numerical time step used = 1.0E-5s (max allowed = 1.0E-5s) 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib secondElec True from ambiant electrons: 0 particles, last numerical time step used = 1.0E-5s (max allowed = 1.0E-5s) 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib secondElec BS from ambiant electrons: 0 particles, last numerical time step used = 1.0E-5s (max allowed = 1.0E-5s) 20000 Thu Jun 09 10:47:06 CEST 2016 PICVolDistrib secondElec from ambiant protons: 0 particles, last numerical time step used = 1.0E-5s (max allowed = 1.0E-5s) 20000 Thu Jun 09 10:47:06 CEST 2016 Last local volume potential change: 4.6083302E-4V maximum change, and 7.8692065E-5V average change 20000 Thu Jun 09 10:47:06 CEST 2016 (last plasma time step = 1.0E-5s) 20000 Thu Jun 09 10:47:06 CEST 2016 SC local surface potential change: 0.09316673V maximum change, and 0.09316657V average change 20000 Thu Jun 09 10:47:06 CEST 2016 SC capacitance = 1.0E-6F, last time step = 0.01s 20000 Thu Jun 09 10:47:07 CEST 2016 Simulation time step control : dt is 0.02 s 20000 Thu Jun 09 10:47:07 CEST 2016 20000 Thu Jun 09 10:47:07 CEST 2016 Time: 0.01, Dt = 0.02 s 20000 Thu Jun 09 10:47:07 CEST 2016

posted by Sébastien Hess at Jun 9, 2016 9:39 AM
Quote
Hello Alessandro,

there may be many reasons for such a problem and it is hard to diagnose it with so few information. One thing may be a memory problem. In the SPIS launching script there are commands to start SPIS that contain an option -Xmx0000M, with some number instead of 0000. By default the allocated memory is low (1Gb), you may want to increase it. If you already did so or if it does not work, try to increase the level of verbosity (in the global parameters) and post the last lines that were displayed in the console. It may provide some clues.

Regards,

Sébastien

posted by Alessandro Vicini at Jun 9, 2016 9:29 AM
Quote
Hullo, I'm just beginning to use SPIS, and I have the following problem:

1) I have made a simple test case for a charging analysis with a sphere, and the code runs; 2) same test case with a parallelepiped, and the code runs; 3) same test case with a parallelepiped plus a cilinder on one side. When I launch the run, the code starts doing something, but after one minute or so the cpu usage goes down to 0%, and the run hangs there forever without any progress. No error messages or whatever.

All the surfaces of the geometry are grouped in a single physical group. what can be the problem.

I try to post a picture, don't know if this will work:

image:key1=http://i64.tinypic.com/k0sdc1.jpg