This software is an optimized implementation of the compute-intensive matrix operations of the LiFE algorithm for CPUs and GPUs.

The original LiFE [1,2] code can be found using the Github link.

LiFE-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.

• LiFE
• MatLab
• vistasoft
• Matlab Brain Anatomy (MBA)
• CUDA
• OpenMP
• OpenBLAS

1. Download LiFE software

   git clone https://github.com/brain-life/encode

2. Change directory

   cd encode

3. Download vistasoft software

   git clone https://github.com/vistalab/vistasoft

4. Download MBA software

   git clone https://github.com/francopestilli/mba

5. Download and install CUDA

   https://developer.nvidia.com/cuda-downloads Also, include the CUDA path in bashrc file (use link for help).

6. 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

7. Unzip the downloaded .tar.gz file

   tar -xvzf Demo_Data_for_Multidimensional_Encoding_of_Brain_Connectomes.tar.gz

8. Download LiFE-GPU-opt software

   git clone https://github.com/karanaggarwal1994/life-gpu-opt

1. Run MATLAB

2. Add the encode folder path to MATLAB search path

   >>> addpath(genpath('/my/path/to/the/encode/folder/'))

3. Run the script

   >>> life_opt_demo(sys,BLAS_PATH)

   sys is an integer with options 0 to run original CPU code, 1 to run optimized CPU code and 2 to run optimized GPU code.

   BLAS_PATH is path of the OpenBLAS library in the system.

   Sample run: life_opt_demo(1,'/home/ubuntu/openblas')

Karan Aggarwal, Uday Bondhugula "Optimizing the Linear Fascicle Evaluation Algorithm for Multi-Core Systems" Accepted to International Conference on Supercomputing (ICS) 2019 (to appear).

Karan Aggarwal, Uday Bondhugula "Optimizing the Linear Fascicle Evaluation Algorithm for Multi-Core and Many-Core Systems".

[1] Pestilli, Franco, Jason D. Yeatman, Ariel Rokem, Kendrick N. Kay, and Brian A. Wandell. Evaluation and statistical inference for human connectomes. Nature methods 11, no. 10 (2014): 1058-1063.

[2] Caiafa, C. and Pestilli, F. Multidimensional encoding of brain connectome. Nature Scientific Reports 7, Article number: 11491 (2017)

[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).