On Sat, 4 Sep 2010, ef wrote:
Yeah sadly it is a very huge graph (atleast 300 million nodes). I have also written aglorithms that do the analysis while they are created, otherwise it would just be to complex. I will go ahead and look into your suggestions.
Does that large of a graph need to be in Graphviz format? The parser is complicated because the language itself is complicated (such as attributes for subgraphs). If you are writing the graph yourself, you might want to use a simpler format (like DIMACS); even if someone else is producing it in Graphviz format, you may want to consider writing a custom parser that only handles the subset of the language that your file uses and creates your graph type directly. -- Jeremiah Willcock