Final deadline and oral exam

Your final project files should be copied to a directory created in your home directory on NIC5 named project-final before 03/01/2024 23h59:

  mkdir ~/project-final
  cp -r directory-were-you-code-is/* ~/project-final/

Don't forget that several datasets are available and that we will test your code on new input datasets.

The oral exam will take place in room 1.21 of the Montefiore Institute (B28) according to the following schedule:

You should prepare a 10 min presentation to present your results. Focus on analyzing your results, not on generic reminders. Structuring your presentation according to the Instructions section of the project description is recommended. Be prepared to answer questions individually, to modify and recompile your code on the fly, and to run your code on the NIC5 cluster during the exam.

The exam is face-to-face but we will ask you to share your screen on Teams using this link, so make sure to have Teams installed on your computer before the exam.

Intermediate deadline

mkdir ~/project-intermediate
cp -r directory-were-you-code-is/* ~/project-intermediate/

Additional instructions for the project

Note: the module load Info0939Tools command appears repeatedly throughout this document before the invocation of a tool used for the project. This command only has to be run once per session.

• Get the serial code and sample input data
• Compile and run the serial code
• Generate audio from pressure output
• Convert an output data file to a Gmsh file
• Generate input with different sizes
• Generate input data file from an audio file
• Intermediate deadline

Get the serial code and sample input data

Once you are connected to NIC5, you can retrieve the serial code by running the following commands:

module load Info0939Tools

get_info0939_project

This command will copy all the relevant files in a directory named project_info0939. If a directory with this name already exists, the files will we be placed in a directory named project_info0939-1.

Compile and run the serial code

module load GCC
cd project_info0939
gcc -O3 -o fdtd fdtd.c -lm

This will generate an executable named fdtd. You can test the executable by running one of the simple (and small) examples:

cd example_inputs/simple3d
../../fdtd param_3d.txt

This simple example should run in approx. 1 minute and will produce 4 files: out_p.dat, out_vx.dat, out_vy.dat and out_vz.dat.

Generate audio from pressure output

The data2audio tool allows you to generate an audio file from your output pressure files. Basic usage is

module load Info0939Tools

data2audio -i input_pressure_data -x xcoord -y ycoord -z zcoord

The command above will extract the pressure from the file input_pressure_data at the point with coordinates x, y and z, and produce a file named output.mp3. The name of the output file can be specified using the -o output_name.mp3 option.

If the duration of the audio file is too short, most players will refuse to play it. In this situation where the simulation time is very short (which is the case for the simple example param_3d.txt), you can use the --repeat numreapeat option to repeat the pressure data numreapeat times. The tool will print a warning suggesting you to use this option if the audio file duration is inferior to 2s.

All available options can be obtained using the --help command line switch:

$ data2audio --help

Usage: data2audio [OPTIONS]

 -x or --x COORD the x coordinates of the point to extract
 -y or --y COORD the y coordinates of the point to extract
 -z or --z COORD the z coordinates of the point to extract

 -i or --input  INPUTFILE  the data file from which the value should be extracted
 -o or --output OUTPUTFILE the path to the output audio file

 -r or --repeat NUMREPEAT reapeat the input data NUMREPEAT times. Use this option if the input is very short.

 --normalize-loudness normalize the (perceived) loudness of the output audio file. Use this option if input the volume is low.

Convert an output data file to a Gmsh file

You can use the data2gmsh tool to convert a data file to a Gmsh file for vizualization. Basic usage of the tool is as follows:

module load Info0939Tools

data2gmsh input_data

The command above will produce a file with name full.msh. This can be opened with Gmsh.

For large files, it is recommended to convert a cut instead of the full data. This can be done by using -cutx coordx, -cuty coordy or -cutz coordz option which will produce a file with name cutx.msh, cuty.msh and cutz.msh respectively.

$ data2gmsh -h
Usage: data2gmsh [option] file(s)
Option can be
  -cutx valx: to extract data on plane x = valx
  -cuty valy: to extract data on plane y = valy
  -cutz valz: to extract data on plane z = valz
If no option is given, extract the full data

The output .msh file can be opened with Gmsh (https://gmsh.info)

You can use scripts to automate all the operations in Gmsh: either using Gmsh's built-in scripting language (".geo" files), or using Python or Julia.

For example, let us assume that you want to open multiple .msh files at the same time, apply the same visualization range to all of them, and only show the scale bar for the first one. You can open all the .msh files using the File->Open menu, then merge the following script with File->Merge:

Mesh.SurfaceEdges = 0;
Mesh.SurfaceFaces = 0;
Mesh.VolumeEdges = 0;
For i In {0:PostProcessing.NbViews-1}
  View[i].RangeType = 2; // custom range
  View[i].CustomMin = -1;
  View[i].CustomMax = 1;
  If (i == 0)
    View[i].Name = "Ma jolie vue : pression en Pa";
  Else
    View[i].ShowScale = 0;
  EndIf
EndFor

On the command line you could achieve the same result with:

  $ gmsh -open file1.msh file2.msh file3.msh myscript.geo

Generate input with different sizes

 $ geninput --help

Usage: geninput [OPTIONS]

 -s or --size SIZE an integer value used as the problem size multiplier (default: 1)
 -t or --type TYPE the problem type which can be simple or slit (default: simple)

 $ mkdir slit4x && cd slit4x
 $ geninput --type slit --size 4

Type of input: slit
Size multiplier: 4
Output parameter file: param_slit_4x.txt

[SUCCES] Input files created

Generate input data file from an audio file

You can generate an input data file from an audio file using the audio2input tool. Basic usage is as follows:

module load Info0939Tools

audio2input -i input_audio_file

The command will produce a file with name in_audio.dat that you can then use as an input parameter for your simulation. For example:

audio in_audio.dat 0.5 0.5 0.5

The name of the output file can be specified using the -o output_name.dat option. A list of the available options can be obtained by invoking the tool with the --help command line switch:

$ audio2input --help
Usage: audio2input [OPTIONS]

 -i or --input  INPUT  the audio file to convert to audio input
 -o or --output OUTPUT the file to which the input audio will be written to