Home Forums FDTD Ability to run Optifdtd on HPC and other queries

Ability to run Optifdtd on HPC and other queries

Profile Photo
(joined June 2021)
Participant

Hi,

I would like to know if there is a way to run Optifdtd on High-performance computing clusters. For the simulation that I am trying (electric field enhancement in the vicinity of nanoparticles when shine by nanosecond laser), I need to run it till a couple of nanoseconds with mesh size ~ 0.5 nm, which shows very long estimated times(2000 hrs) which is not helpful at all. I would appreciate it if you could let me know some way to optimize the system to run in a decent time.

I have a couple of more questions:
1. when we record the simulation frames, how can we include a legend in the video?
2. The output that we get at the end of the simulation in the analyzer, is it the time-averaged output or the results at the last time step? It is not very clear to me.

Thank You,
Suman

Responses (1):

    • #69525
      Profile Photo
      Suman Kumari
      Participant

      Hi,

      I would like to know if there is a way to run Optifdtd on High-performance computing clusters. For the simulation that I am trying (electric field enhancement in the vicinity of nanoparticles when shine by nanosecond laser), I need to run it till a couple of nanoseconds with mesh size ~ 0.5 nm, which shows very long estimated times(2000 hrs) which is not helpful at all. I would appreciate it if you could let me know some way to optimize the system to run in a decent time.

      I have a couple of more questions:
      1. when we record the simulation frames, how can we include a legend in the video?
      2. The output that we get at the end of the simulation in the analyzer, is it the time-averaged output or the results at the last time step? It is not very clear to me.

      Thank You,
      Suman

    • #69537
      Profile Photo
      Scott Newman
      Moderator

      Hello Suman,

      Please allow me to address your points one at a time.

      1. We do support HPC systems, however this is done through the Linux build of the product. What is done is the design is done on the windows machine and then a Linux design file (FDU) can be generated which can be run on the Linux simulator. You would need the add on license for the Linux simulator and would recommend you contact support (support@optiwave.com) to look at your options.
      2. There is not a legend available for videos but within the recording tab you can specify the upper and lower limits for your videos, assuming you are using one of the more recent releases of the product. You should also be aware that recording video significantly impacts how long your simulation will take. If you can I would recommend not recording videos for such a large simulation.
      3. If by output you are referring to the observation areas or observation lines then what you are seeing is the cumulative DFT of the field during a portion of the simulation. If you are running a pulsed simulation the timeframe is over the entire simulation and if you are running a CW there is a finalization section that is sized based on your central wavelength to be 10 periods that occurs once your requested time steps are completed.

Login You must be logged in to reply to this topic.