search

UMD-ARLIS / Graph-Benchmarking-Project

This repository contains work towards creating a comprehensive cross-platform graph processing benchmark for Community Detection and Subgraph Isomorphism problems. The repository is NOT YET COMPLETE and should not be used outside of a preliminary development perspective.
2 stars 0 forks source link

Benchmark Dataset for Community Detection #11

Open osullik opened 9 months ago

osullik commented 9 months ago

As a researcher developing cross-hardware benchmarks, I require a suite of graphs of varying 'difficulty' to use as input to community detection algorithms so that I can evaluate the impact that different hardware has on the Execution Time, Memory Usage, Power Consumption, Graph Traversal order and Processor Utilization.

osullik commented 9 months ago

This is done when:

  1. I have a suite of at least 10 input graphs that vary from 'trivial' to 'extremely difficult' for a community detection algorithm to resolve. (a) There should be a range of 'difficulty' in "Accuracy" - that is, some should be hard for a CD algorithm to resolve to good 'communities. (b) There should be a range of 'difficulty' in "Performance" - that is, some should require longer to compute because of the shape of the input data relative to the complexity.
  2. The Data files are stored in a compressed manner on GIT LFS or some other version-controlled Repository
  3. Instructions (and code) for generating the data files is available on the benchmark website.