https://github.com/wheatman/Packed-Compressed-Sparse-Row

Dynamic data structure for sparse graphs.
https://github.com/wheatman/Packed-Compressed-Sparse-Row

Last synced: 30 days ago
JSON representation

Dynamic data structure for sparse graphs.

• Host: GitHub
• URL: https://github.com/wheatman/Packed-Compressed-Sparse-Row
• Owner: wheatman
• License: mit
• Created: 2018-09-28T04:44:41.000Z (about 6 years ago)
• Default Branch: master
• Last Pushed: 2024-04-03T02:37:07.000Z (8 months ago)
• Last Synced: 2024-05-23T00:11:03.828Z (7 months ago)
• Language: C++
• Homepage:
• Size: 14.6 KB
• Stars: 26
• Watchers: 9
• Forks: 6
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-dynamic-graphs - PCSR

README

        
# Packed-Compressed-Sparse-Row

Dynamic data structure for sparse graphs.

compile with `g++ -std=c++11 PCSR.cpp`

```

  // initialize the structure with how many nodes you want it to start with

  PCSR pcsr = PCSR(10);

  // add some edges

  for (int i = 0; i < 5; i++) {

    pcsr.add_edge(i, i, 1);

  }

  // update the values of some edges

  for (int i = 0; i < 5; i++) {

    pcsr.add_edge_update(i, i, 2);

  }

  // print out the graph

  pcsr.print_graph();

```

For more information see https://ieeexplore.ieee.org/abstract/document/8547566

Please cite as:

```

@inproceedings{wheatman2018packed,

  title={Packed Compressed Sparse Row: A Dynamic Graph Representation},

  author={Wheatman, Brian and Xu, Helen},

  booktitle={2018 IEEE High Performance extreme Computing Conference (HPEC)},

  pages={1--7},

  year={2018},

  organization={IEEE}

}

```

# Subsequent Papers and Projects

This work was continued and made parallel in the paper [A Parallel Packed Memory Array to Store Dynamic Graphs](https://epubs.siam.org/doi/abs/10.1137/1.9781611976472.3)

The Parallel PMA was later used in the larger Terrace system.  More details can be found in [Terrace: A Hierarchical Graph Container for Skewed Dynamic Graphs](https://dl.acm.org/doi/abs/10.1145/3448016.3457313). The code for terrace which includes the Parallel PMA can be found at https://github.com/PASSIONLab/terrace.

Many ideas from this work also went into the creation of [Streaming Sparse Graphs using Efficient Dynamic Sets](https://ieeexplore.ieee.org/abstract/document/9671836) for which the code can be found at https://github.com/wheatman/SSTGraph.

All of these systems are parallel and much faster than the original PCSR system, but are more complex.  

PCSR is not being updated, but I will review any pull requests.