Computational tools

Self-written simulation tools

Licence information

My codes are not public. They can only be used under research collaboration.
Don't hesitate to contact me in case of interest.

Femtosecond laser irradiation of silicon
"FLIRT" 1D/2D [repository]

Femtosecond Laser Interaction for the investigation of Ripples in Silicon.

Financial supports:
  • French Ministry of Research, Grant for priority topics on photovoltaics (36 months)
  • BAM Federal Institute for Material Research and Testing (Adolf-Martens-Fund), Germany (15 months)
  • Physics
    • Plasmonic surface wave module : Surface Plasmon Polariton excitation.
    • Sipe/Young surface wave module : "Generation of roughness scattered surface wave"
    • Fourier / Ballistic-Diffusive energy balance equations. 
    • Melting, cooling and resolidification.
    • Single and double-laser pulse with delay sweeping

  • ITs
    • Coded in C/C++
    • Spatial splitting for 1D/2D solving
    • Distributed calculation using MPI (for double-laser pulses, laser fluence, angle of incidence...)
    • Scripted post-treatment using bash+GNUplot
    • Implicit time-integration
    • Finite volume method uses an orthogonal irregular mesh. 
    • Versionning managed with Git and Bitbucket

  • Papers published using this code
    • TJY Derrien et al, Applied Physics A (2013) [DOI],
    • TJY Derrien et al, Optics Express (2013) [PDF]
    • TJY Derrien et al, Journal of Applied Physics (2013) [PDF]
    • TJY Derrien et al, Applied Surface Science (2012) [PDF]
    • TJY Derrien et al, Journal of Optoelectronics and Advanced Materials (2010) [PDF]

Femtosecond laser irradiation of nanostructured materials
"FLAPS" 2D [repository]

Femtosecond Laser Assisted Probe tomography of Silicon.
[official financial support: ANR "Ultrasonde" (11 months)]
  • Physics
    • Inhomogeneous laser absorption using Mie scattering theory (TM, TE)
    • Fermi-Dirac statistics for electrons and holes thermal properties
    • Free-carrier generation and recombination (one-, two- photon transitions, ...)
    • Diffusive transport of the free-carriers
    • Thermal diffusion and heating of the lattice
    • Cooling by the backside

  • ITs
    • Finite volume method
    • Explicit time integration
    • Generation of non-orthogonal structured mesh from a user-defined contour
    • Transport is discretized on direct and interpolated dual mesh
    • Coded in Fortran 90 with OpenMP

Locally available computers

  1. CPU 2 x 8 cores, Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz, 32 GB RAM.
    GPU NVIDIA GeForce GT 630(D3), GPU 96 cores @ 700 MHz, 2 GB RAM.
    Local HD 886 GB.
    Portable HD 466 GB.

  2. CPU 2 x 6 cores, Intel(R) Xeon(R) CPU E5-2630 v2 @ 2.60GHz, 16 GB RAM.
    Radeon Tahiti PRO HD 7950/8950 OEM / R9 280.
    Local HD: 848 GB.

  3. CPU 2 x 4 cores, Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz, 8 GB RAM.
    Local HD: 385 GB + 1.5 TB.

  4. CPU 2 x 2 cores, Intel(R) Core(TM) i5-5200U CPU @ 2.20 GHz, 4 GB RAM.
    NVIDIA GeForce GT940M, GPU 384 cores @ 1072 MHz, 8 GB RAM.
    Local SSD: 256 GB.

Access to Computational clusters

IT4 Innovation (Czech Republic).
    1. ANSELM, Praha (CZ) :
      - CPU-nodes: 2800 x E5-2665 (8 cores w/o Hyperthreading) @ 2.4 GHz, RAM 64 GB/node, HD 500 GB per node.
      - GPU-nodes: 368 x E5-2470 (8 cores w/o Hyperthreading) @ 2.3 GHz, RAM 96 GB/node, GPU Nvidia Tesla Kepler K20, 8 GB DRAM
      - "Many Integrated Cores": 4 nodes of 8 x CPU E5-2470 (8 cores w/o HyperThreading) @ 2.4 GHz, RAM 96 GB / node, Accelerator Intel Phi 5110P/node
      - "Fat nodes": 2 nodes of 2x16 cores, Intel E5-2665 @ 2.4 GHz, RAM 512 GB.

    2. Salomon, Ostrava (CZ) :
      - CPU-nodes: 1000 nodes of 2 x Haswell E5-2680 v3 (12 cores w/o Hyperthreading) @ 2.5 GHz, RAM 128 GB / node.

We successfully obtained 1'100'000 core-hours under the project Quantum-1st-LaP until Jan 2016.

6th call for computational time at IT4I is not yet announced. It's expected round about Novembre 2015.

Interfacing MS Office and Latex users

Since scientific communities do not use the same tools, it is important to allow them to communicate.
Here are some helpful links to convert Latex/LyX documents to DOC/DOCX (MS Office) format.

  1. LyX --> HTML (MS Office version) --> Open with MS Office works very well (!!)
  2. A summary (Nov. 2013) comparing several techniques for academic publication purposes:
  3. Nitro Pro 9 PDF converter (~ USD$ 100) worked well with MS Office 2013, but very badly with MS Office 2010.

FDTD softwares

I have experience now with several FDTD softwares. Here is a summary. 
  • Lumerical: expensive, usually limited to 1 machine, not executable on clusters, native graphs are not publishable. Tips: use the interface with Matlab (to pay also...). But results are good and using it is very easy. 
  • EMexplorer: excellent software. Commercial. Limited to Windows, and 1 core, but scriptable (directly). 2D/3D FDTD. 
  • GsVit: documentation is extensive, but unprecise and not up to date. I lost many days to find out things which were not said. Open source, free, available on GPU/CPU, Windows/Linux. Interface with Python is possible. 3D calculations only. 
  • MEEP: beautiful, fast, 2D/3D interface with Python.