IntroductionThis platform aims at accelerating the frequently used but time-consuming algorithms in neuroscience research. We have accelerated the process of brain network analysis (BNA) with NVIDIA GPUs (Graphic processing unit) and multi-core CPUs. The toolbox provides functions for the construction and the analysis of large networks. Network construction is intended for fMRI data using Pearson’s correlation. Network analysis is general purposed and includes the calculation of clustering coefficient, characteristic path length, network efficiency, and betweenness centrality (and comparisons to Maslov random networks). We accelerate Pearson’s correlation calculation, APSP and betweenness with GPUs. Other functions are implemented on multi-core CPUs.
Release Notesversion 4.5, Sep. 09, 2014
version 4.3, Jun. 20, 2014 New Released Information: The new win64 version integrates a folder with functions for the construction and the analysis of weighted networks. Original or normalized participant coefficient is optional in the new version. More details can be seen in the manuscript. A linux version is coming soon. |
Downloads To get started with CUDA, please follow NVIDIA CUDA Getting Started Guide for Microsoft Windows or NVIDIA CUDA Getting Started Guide for Linux.
Requirements
OS: 64-bit Windows or Linux
GPU: Nvidia with CUDA support version of CPU-only for Linux coming soon... Software: visual studio 2010 or above version for compiling the win64 version; g++ for compiling the Linux version; NVIDIA CUDA toolkit v5.5. If you use another version of CUDA toolkit, you need to reconstruct a Project for win64 version or modify the makefile for Linux version. |
Platform Overview
Some Results
Related Publications
Paper Slides Technical report (based on the previous data, we are now updating the data, but the platform is still good)
If you found the platform is useful, please cite the first paper (Plos one).
Wang, Y., Du, H., Xia, M., Ren, L., Xu, M., Xie, T., ... & He, Y. (2013). A Hybrid CPU-GPU Accelerated Framework for Fast Mapping of High-Resolution Human Brain Connectome. PloS one, 8(5), e62789. [HTML]
Xu, M., Zhang, X., Wang, Y., Ren, L., Wen, Z., Xu, Y., ... & Yang, H. (2012, May). Probabilistic Brain Fiber Tractography on GPUs. In Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW), 2012 IEEE 26th International (pp. 742-751). IEEE. [PDF]
Wang, Y., Xu, M., Ren, L., Zhang, X., Wu, D., He, Y., ... & Yang, H. (2011, November). A heterogeneous accelerator platform for multi-subject voxel-based brain network analysis. In Proceedings of the International Conference on Computer-Aided Design (pp. 339-344). IEEE Press. [PDF]
Wu, D., Wu, T., Shan, Y., Wang, Y., He, Y., Xu, N., & Yang, H. (2010, December). Making human connectome faster: GPU acceleration of brain network analysis. In Parallel and Distributed Systems (ICPADS), 2010 IEEE 16th International Conference on (pp. 593-600). IEEE. [PDF]
Wang, Y., He, Y., Shan, Y., Wu, T., Wu, D., & Yang, H. (2010, August). Hardware computing for brain network analysis. In Quality Electronic Design (ASQED), 2010 2nd Asia Symposium on (pp. 219-222). IEEE. [PDF]
If you found the platform is useful, please cite the first paper (Plos one).
Wang, Y., Du, H., Xia, M., Ren, L., Xu, M., Xie, T., ... & He, Y. (2013). A Hybrid CPU-GPU Accelerated Framework for Fast Mapping of High-Resolution Human Brain Connectome. PloS one, 8(5), e62789. [HTML]
Xu, M., Zhang, X., Wang, Y., Ren, L., Wen, Z., Xu, Y., ... & Yang, H. (2012, May). Probabilistic Brain Fiber Tractography on GPUs. In Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW), 2012 IEEE 26th International (pp. 742-751). IEEE. [PDF]
Wang, Y., Xu, M., Ren, L., Zhang, X., Wu, D., He, Y., ... & Yang, H. (2011, November). A heterogeneous accelerator platform for multi-subject voxel-based brain network analysis. In Proceedings of the International Conference on Computer-Aided Design (pp. 339-344). IEEE Press. [PDF]
Wu, D., Wu, T., Shan, Y., Wang, Y., He, Y., Xu, N., & Yang, H. (2010, December). Making human connectome faster: GPU acceleration of brain network analysis. In Parallel and Distributed Systems (ICPADS), 2010 IEEE 16th International Conference on (pp. 593-600). IEEE. [PDF]
Wang, Y., He, Y., Shan, Y., Wu, T., Wu, D., & Yang, H. (2010, August). Hardware computing for brain network analysis. In Quality Electronic Design (ASQED), 2010 2nd Asia Symposium on (pp. 219-222). IEEE. [PDF]
Related Work
Human Connectome Project, http://humanconnectome.org/
Fast computation of functional networks from fMRI activity: a multi-platform comparison. A.R. Rao, R.Bordawekar, G.A. Cecchi, SPIE Conference on Medical Imaging, SPIE Press, 2011.
Fast computation of functional networks from fMRI activity: a multi-platform comparison. A.R. Rao, R.Bordawekar, G.A. Cecchi, SPIE Conference on Medical Imaging, SPIE Press, 2011.
Developers
Haixiao Du, Tsinghua University, China
Mingrui Xia, Beijing Normal University, China
Ling REN, Tsinghua University, China
Mo XU, Tsinghua University, China
Xiaorui, ZHANG, Tsinghua University, China
Di WU, Tsinghua University, China
Gushu Li, Tsinghua University, China
Jiantao Qiu, Tsinghua University, China
Yuze Chi, Tsinghua University, China
Yu WANG, Tsinghua University, China
Yong HE, Beijing Normal University, China
Mingrui Xia, Beijing Normal University, China
Ling REN, Tsinghua University, China
Mo XU, Tsinghua University, China
Xiaorui, ZHANG, Tsinghua University, China
Di WU, Tsinghua University, China
Gushu Li, Tsinghua University, China
Jiantao Qiu, Tsinghua University, China
Yuze Chi, Tsinghua University, China
Yu WANG, Tsinghua University, China
Yong HE, Beijing Normal University, China
NICS, Department of Electronic Engineering, Tsinghua University
Last updated Mar. 2014
Last updated Mar. 2014