bdb and mpi selectors for the graph library
Hi all, as part of a project dealing with (very) big graphs, I am currently writing two containers. The first one is aimed at situations when you have huge graphs and want to distribute them among different computers. It uses MPI to communicate between computers. The second is aimed at situations when you can't distribute your graph (e.g. you're partitionning it before distributing it), or speed isn't an issue. It uses a berkeley DB database to store the graph. Is this code of interest for boost, would you consider reviewing it once it's tested and stable ? Pierre DOUCY -- Written slowly for people who can't read fast.
On Monday 03 May 2004 09:41 am, pdoucyml@free.fr wrote:
Hi all,
as part of a project dealing with (very) big graphs, I am currently writing two containers. The first one is aimed at situations when you have huge graphs and want to distribute them among different computers. It uses MPI to communicate between computers.
The second is aimed at situations when you can't distribute your graph (e.g. you're partitionning it before distributing it), or speed isn't an issue. It uses a berkeley DB database to store the graph.
Is this code of interest for boost, would you consider reviewing it once it's tested and stable ?
Yes. We're going to be looking into parallel graph computation in the BGL in the coming months as well. Doug
participants (2)
-
Douglas Gregor -
pdoucyml＠free.fr