jan/opencl_fdtd
2
2
Fork 1
Code Issues Pull Requests Releases Wiki Activity
OpenCL FDTD electromagnetic simulation in 3 dimensions
19 Commits 6 Branches 0 Tags 60 MiB
Python 73.8%
C 26.2%
d34c478f1d
Go to file
jan d34c478f1d Rewrite, with the following features: 
- Move to jinja2 templates for the opencl code
- Combine PML code into the E, H updates for speed
- Add Poynting vector calculation code, including precalculation during H update
- Use arrays for PML parameters (p0, p1)
- Switch to linearized, C-ordered fields (~50% performance boost??)

- Added jinja2 and fdfd_tools dependencies
2017-03-28 21:53:51 -07:00
fdtd Rewrite, with the following features: 2017-03-28 21:53:51 -07:00
.gitignore gitignore changes 2016-03-30 15:06:20 -07:00
fdtd.py Rewrite, with the following features: 2017-03-28 21:53:51 -07:00
LICENSE.md Use markdown for license 2016-04-13 03:57:09 -07:00
pcgen.py Fix triangular lattice code in pgcgen 2016-06-21 18:25:23 -07:00
README.md update module name 2016-05-30 22:43:14 -07:00
requirements.txt Rewrite, with the following features: 2017-03-28 21:53:51 -07:00

README.md

opencl_fdtd

opencl_fdtd is a python application for running 3D time-domain electromagnetic simulations on parallel compute hardware (mainly GPUs).

Performance highly depends on what hardware you have available:

  • A 395x345x73 cell simulation (~10 million points, 8-cell absorbing boundaries) runs at around 42 iterations/sec. on my Nvidia GTX 580.
  • On my laptop (Nvidia 940M) the same simulation achieves ~8 iterations/sec.
  • An L3 photonic crystal cavity ringdown simulation (1550nm source, 40nm discretization, 8000 steps) takes about 5 minutes on my laptop.

Capabilities are currently pretty minimal:

  • Absorbing boundaries (CPML)
  • Conducting boundaries (PMC)
  • Anisotropic media (eps_xx, eps_yy, eps_zz, mu_xx, ...)
  • Direct access to fields (eg., you can trivially add a soft or hard current source with just sim.E[1] += sin(f0 * t), or save any portion of a field to a file)

Installation

Requirements:

  • python 3 (written and tested with 3.5)
  • numpy
  • pyopencl
  • h5py (for file output)
  • gridlock
  • masque

To get the code, just clone this repository:

git clone https://mpxd.net/gogs/jan/opencl_fdtd.git

You can install the requirements and their dependencies easily with

pip install -r requirements.txt

Running

The root directory contains fdtd.py, which sets up and runs a sample simulation (cavity ringdown).

python3 fdtd.py