FDTD Audio Simulation using CUDA with a visualizer written in C + GTK
This code requires the dev versions of GTK+-3.0 (sudo apt-get install libgtk-3-dev) and lbsndfile (sudo apt-get install libsndfile1-dev) to compile.
Compile the CUDA simulation code with nvcc -o sim sim.cu and the GUI visualizer code with gcc ``pkg-config gtk+-3.0 --cflags`` visualizer.c -lm -o visualizer ``pkg-config gtk+-3.0 --libs`` .
Just run the GUI visualizer with ./visualizer.
You can also optionally choose to run the CUDA simulator from the command line, which allows a simple interface for scripts/programs.
The simulator can be passed options in the format ./sim [-i infolder] [-o outfolder] [-t timesteps] [-g gridsize] [-b blockdimensions].
The simulation will run with no audio inputs by default, unless the input folder includes an audio_ledger.txt file. This will list all audio input files, which must be in the same directory as it. It should contain one line, in the format xpos ypos zpos audiofilename, per audio file.
The simulation will run on an empty grid by default, unless the input folder includes a sim_state.bin file. This contains all values in the simulation at a particular time. By default, the simulator outputs the last frame of each simulation to a sim_state.bin file in the output folder. If you want to continue a simulation from where it stopped, you'll need to put this file in your input folder.
This code is licensed under the MIT license, included in the LICENSE file
Libsndfile is released under the GNU LGPL, a copy of which is included in this repo