# WISE

**Repository Path**: libaoke/WISE

## Basic Information

- **Project Name**: WISE
- **Description**: No description available
- **Primary Language**: Java
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-11-17
- **Last Updated**: 2021-11-17

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# WISE: Fast and Efficient Vertex-cut Graph Partitioning

This repository contains the code for our novel algorithm to partition large-scale graphs with billions of edges with linear runtime complexity. Users can specify their runtime preference to control the trade-off between partitioning quality and runtime. For example, if you partition the graph as a pre-processing step for complex distributed graph algorithms such as PageRank or Deep Learning algorithms, you want to achieve high partitioning quality at the cost of higher partitioning latency -- as this reduces overall graph processing latency significantly:

For example: we executed PageRank on a web graph partitioned by different state-of-the-art algorithms. Our algorithm (AD)WISE reduces graph processing latency by 50% and network traffic by 66%!

![Traffic Savings](https://github.com/GraphPartitioning/WISE/blob/master/graphTraffic.jpg)
![Latency Savings](https://github.com/GraphPartitioning/WISE/blob/master/graphLatency.jpg)