Multi-Gi.L.A. is a multi-level distributed layout drawing algorithm implemented in Giraph.

This is a support website for the publication submitted to GD 2016. You can download the graphs used in the experimentation and you can find a link to the source code hosted on GitHub.


  • Java 1.7
  • Maven (needed to compile the code)
  • A Hadoop cluster running Hadoop 2.6.0

Go to next page to find out how to use the software.


Follow the link to connect to the GitHub repository of MultiGila.


The following archives contain the test graphs used in the paper and the startup scripts.
The graphs are undirected and encoded in JSON format: [id,x,y,[(neighbor_id,)*]]

Real Graphs

5 real world graph with sizes up to 1,5M edges.

Big Graphs

5 graphs with sizes up to 12M edges.

Startup Scripts

The startup scripts and instructions to use MultiGila.


  • Web-GoogleGoogle web graph. 800k vertices x 5M edges
  • RoadNet-PAPennsylvania road network. 1M vertices x 1.5M edges
  • Hugetric-207M vertices x 11M edges
  • Hugetric-106,5M vertices x 10M edges
  • Delaunay_n22Triangulation of random points. 7M vertices x 12M edges
  • DBLPDblp co-authors network. 300k vertices x 1M edges.
  • Asic-320Circuit Simulation Problem. 300k vertices x 500k edges
  • Close-up of ASIC-320 graph.
  • Amazon0302Amazon co-purchasing network. 250k vertices x 800k edges
  • com-AmazonAmazon co-purchasing network. 300k vertices x 900k edges
  • Close-up of amazon graph.
  • wikitalkWikpedia Talk Network. 3M vertices x 5M edges
  • 3elt
  • 4970
  • 4elt
  • crack
  • flower_005
  • Grid 40x40 double folded
  • cti
Web-Google1 RoadNet-PA2 Hugetric-203 Hugetric-104 Delaunay_n225 DBLP6 Asic-3207 8 Amazon03029 com-Amazon10 11 wikitalk12 3elt13 497014 4elt15 crack16 flower_00517 Grid 40x40 double folded18 cti19
jquery slider by v8.7.1m


  • Alessio Arleo

  • Walter Didimo

  • Giuseppe Liotta

  • Fabrizio Montecchiani