Scalable Parallel Feature Extraction and Tracking for Large Time-Varying 3D Volume Data

Our implementation of the paper Scalable Parallel Feature Extraction and Tracking for Large Time-varying 3D Volume Data. We built this as a part of our course CS677: Topics in Large Data Analysis and Visualization

To run the code the following libraries need to be present:

• A CPP compiler
• MPI library installation: We have used the mpich library to run our code

The script run_script.sh automates the running on the kd lab systems where the hosts can be configured.

Installation and Running

Make a directory named build inside the project folder

mkdir build
cd build

We have used cmake to build the Makefile which needs to be installed. Inside the build directory, run

cmake ..

After this, the Makefile will be formed which can be run directly using make.

make

Once the binary is built, we can directly run the run_script.sh in the parent directory. Be sure to make the PATH variables point correctly.

Now run the following command

../run_script.sh

helper python files information

• binary_merger.py merges all the binary outputs of each of the processors
• binary_to_vti.py for converting binary files to .vti file
• vti_to_binary.py for converting .vti to binary
• tfe_writer.py for writing the .tfe file from the colormap.json file
• normalize_data.py for normalizing the data