I have a simulation that keeps getting killed by Ubuntu. I do not believe it is an error in the simulation or MEEP. How do I diagnose this? #3083
-
|
These are my files, it's a very basic dipole wrapped in a 3D printed dielectric. The issue I am having is once the resolution is 12 or more, it gets killed by Ubuntu (signal 9) with no explanation. I am running this on Windows Subsystem for Linux 2. I am thinking that MAYBE it is running out of RAM, which is odd because Ia m already allocating it 64 GB. I saw there was a calculation for an empty simulation being SxSySz(r^3)*96 gives the total number of bytes required for an empty simulation. Per my calculations, I get 2906212*(12^3)=372,833,280=372 MB Obviously, this is a lower bound, but it seems like something that should be running. Thank you. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
This sounds like an OOM. What are your cell dimensions in meep units? I'm thinking your memory calculation is off. Looking at the code, you're using meeps gds library which isn't the most reliable. I typically use gdspy to parse the gds file and manually build up my simulation from that. |
Beta Was this translation helpful? Give feedback.
-
|
How do I go from nm to meep units? Also, do you have any resources for how to use gdspy instead of MEEP's gds library? |
Beta Was this translation helpful? Give feedback.
This sounds like an OOM. What are your cell dimensions in meep units? I'm thinking your memory calculation is off.
Looking at the code, you're using meeps gds library which isn't the most reliable. I typically use gdspy to parse the gds file and manually build up my simulation from that.