This software is an optimized implementation of the compute-intensive matrix operations of the LiFE algorithm for GPUs.
The original LiFE [1,2] code can be found using the Github link.
LiFE-GPU-opt software is available under the BSD 3-Clause license.
Copyright (2019), Karan Aggarwal, [email protected]
This work was supported in part by a grant (EMR/2016/008015) from the Science and Engineering Research Board (SERB), India through its Extramural Research funding program.
-
Download LiFE software
git clone https://github.com/brain-life/encode
-
Change directory
cd encode
-
Download vistasoft software
git clone https://github.com/vistalab/vistasoft
-
Download MBA software
git clone https://github.com/francopestilli/mba
-
Download and install CUDA
https://developer.nvidia.com/cuda-downloads
Also, include the CUDA path in bashrc file (use link for help). -
Download demo datasets from the repository doi:10.5967/K8X63JTX
https://scholarworks.iu.edu/cgi-bin/mdssRequest.pl?file=2022/2099/Demo_Data_for_Multidimensional_Encoding_of_Brain_Connectomes.tar.gz
-
Unzip the downloaded .tar.gz file
tar -xvzf Demo_Data_for_Multidimensional_Encoding_of_Brain_Connectomes.tar.gz
-
Download LiFE-GPU-opt software
git clone https://github.com/karanaggarwal1994/life-gpu-opt
-
Run MATLAB
-
Add the encode folder path to MATLAB search path
>>> addpath(genpath('/my/path/to/the/encode/folder/'))
-
Run the script
>>> life_gpu_opt_demo
[3] Kumar, S., Sreenivasan V., Talukdar P., Pestilli F., and Sridharan D. (2019, January) "ReAl-LiFE: Accelerating the discovery of individualized brain connectomes on GPUs." Accepted to AAAI 2019 (proceedings in press).