[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
## Re: [igraph] memory igraph

**From**: |
Tamas Nepusz |

**Subject**: |
Re: [igraph] memory igraph |

**Date**: |
Thu, 4 Feb 2010 12:33:51 +0000 |

**User-agent**: |
Mutt/1.5.20 (2009-06-14) |

Dear Raghav,
>* I want to create a large random graph, using erdos_renyi_game GNP:*
>* *
>* N ~ 262144*
>* P ~ 0.5*
This means that the resulting graph will have approximately 17 billion
edges. (262144^2 / 4). There's no way you can fit a graph of that size
into the memory of your computer.
>* In any case I would like to generate a graph that is a group of erdos-renyi *
>* graphs, with the groups connected randomly. There would be of the order *
>* of 1000 nodes in each group and 256 groups. *
In that case, you are probably better off with igraph_preference_game.
This lets you specify a probability distribution of vertex types and a
preference matrix which defines the connection probability between
vertices of different types. If you want 256 groups, define 256 vertex
types with equal probabilities, and use a preference matrix which uses a
large connection probability between vertices of the same type and a
small probability between vertices of different types. That would
probably fit into the memory.
Or, alternatively, generate 256 Erdos-Renyi graphs using
igraph_erdos_renyi_game and then take the disjoint union.
--
Tamas