Prev: halting problem proof, via diagonalization?
Next: 3SAT - A Counting Solver
From: Александр on 8 Aug 2008 13:01
Hi theory guru! I am interested in compress large (spare/complete)
(static/dynamic) graph model representation. What data structures
helps to represent graph with 10.000.000 and more vertexes as compress
as possible. And what data structures provide high performance factor
for such graphs?
Maybe, it ll be coded adjacency matrix or map/multimap, sets or hash
table..?
From: Le Chaud Lapin on 12 Aug 2008 00:22
On Aug 8, 12:01špm, áÌÅËÓÁÎÄÒ <palgol...(a)gmail.com> wrote:
> Hi theory guru! I am interested in compress large (spare/complete)
> (static/dynamic) graph model representation. What data structures
> helps to represent graph with 10.000.000 and more vertexes as compress
> as possible. And what data structures provide high performance factor
> for such graphs?
> Maybe, it ll be coded adjacency matrix or map/multimap, sets or hash
> table..?

Hmmm...

I just replied to a very similar post in comp.lang.c++, only the
number was 6.000.000 instead of 10.000.000, so I am curious if there
is any correlation here. :)

http://groups.google.com/group/comp.lang.c++.moderated/browse_frm/thread/bc24afa0bc96f402/

It depends on what you plan to do with the graph. If you only want to
determine adjacency, then that's one matter, but if you need something
to execute, say, Dijkstra's algorithm, you will find that the suitable
data structure suprisingly fat.

What are you doing this for?

-Le Chaud Lapin-
 | 
Pages: 1
Prev: halting problem proof, via diagonalization?
Next: 3SAT - A Counting Solver