README
¶
DRONE-EBV
Overview
DRONE is a Go implementation of distributed subgraph-centric graph computing framework. EBV is a parallel graph partition algorithm implement by mpich.
For detailed information of EBV, you can reference here
We support cuda optimization and gcn in master branch now!
Dependencies
DRONE and EBV build, run and test on GUN/Linux. It's depends on:
- A modern c++ compiler
- Open MPI or MPICH
- gflags
Install and run
You should first partition the input graph. For build and run EBV:
mpic++ EBV_mpi.cpp -o EBV_mpi -lgflags
mpirun -np processnum ./EBV_mpi -filename edge_list_file -vertices vertex_num -edges edge_num
This program takes edge-list directed graph (with vertex id start from 0) as input. You should also specify the number of vertices and edges. If you want to dump the results, you should also add two parameter like:
mpirun -np 2 EBV_mpi -filename demo_graph.txt -vertices 6 -edges 5 --is_dump -output output_path
The number of subgraphs is equal to the processnum.
Further, if you want DRONE supports fault tolerance, you should add parameter -two_replica.
For running DRONE, you should also build the DRONE_start.c file at first:
cd DRONE/sbin
mpic++ DRONE_start.cpp -o DRONE_start -lgflags
To run DRONE without fault tolerance:
mpirun -np 3 DRONE_start -jobname cc -graph_path path
// jobname: sssp, cc, pr
If you want to test the fault tolerance performance, you can run with:
mpirun -np 3 DRONE_start -jobname cc -graph_path path -is_rep
Cite
You can cite our work with
@inproceedings{zhang2021efficient,
title={An Efficient and Balanced Graph Partition Algorithm for the Subgraph-Centric Programming Model on Large-scale Power-law Graphs},
author={Zhang, Shuai and Jiang, Zite and Hou, Xingzhong and Guan, Zhen and Yuan, Mengting and You, Haihang},
booktitle={2021 IEEE 41st International Conference on Distributed Computing Systems (ICDCS)},
pages={68--78},
year={2021},
organization={IEEE}
}
Directories