[OSM-dev] osm2pgsql-64 and benchmarks
Frederik Ramm
frederik at remote.org
Wed Jun 22 08:58:12 BST 2011
Martijn,
On 06/22/11 08:31, Martijn van Oosterhout wrote:
> The reason I went for arrays of integers is diskspace (4 bytes per
> node rather than 40 for a whole row). Did you compare that?
When I initially made osm2pgsql-64 I looked at the database size on disk
and found it to be about 20% bigger in total when using 64-bit IDs than
when using 32-bit IDs. Sadly I forgot to record the disk usage in the
experiment from this thread.
Another thing to consider is of course memory usage; osm2pgsql will
require more RAM (or achieve less caching performance with the same
amount of RAM) when you do 64-bit IDs. I felt that the node caching
could probably be cleverly modified so that it would only require more
memory for those IDs that really were outside the 32bit range but
haven't ventured into that.
Bye
Frederik
More information about the dev
mailing list