Hi Henning, it seems I still don't understand. The max-nodes values influences the number of tiles created. I don't see a big difference between a map for US consisting of 100 or 250 tiles, esp. if the map doesn't contain routing info. I would try to find a value that allows to utilize all CPUs without reaching heap limits or img limits. Why do you think that you have to find an optimal value? Gerd Henning Scholland wrote
Hi Gerd,
finding a proper max-nodes value wont solve the problem globaly. Eg. then you compare the needed value from europe/north america to the value needed in the rest of the world.
Henning
Am 25.04.2014 19:35, schrieb Gerd Petermann:
Hi Henning,
well, I don't know how well the number of nodes in a tile correlates with the size of the img file, but it seems to work for most users. My understanding is that this should work as well for a style which only processes a few details, only the ratio between number of nodes and tile size will be different, in other words, you have to find out how much higher the max-nodes value can be.
Gerd
Date: Fri, 25 Apr 2014 19:27:09 +0200 From:
osm@
To:
mkgmap-dev@.org
Subject: Re: [mkgmap-dev] mkgmap ToDo list
Hi Gerd, splitter is able to write a density file, which stores the density of OSM-nodes in the specific area. Afterwards the file is splitted and processed by mkgmap. mkgmap will use only the objects, which are addressed in style-file.
Example: In a rectangle of 100x100m are 5 nodes belonging to a highway and 10 nodes belonging to POI, and polygones and style-file only contains highways, mkgmap writes only 5 nodes (and one line) to img-file.
For an more effictive splitting of tiles I would imagine, that a density-file created by mkgmap based on the written data will be much better then the file written by splitter.
Henning
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
-- View this message in context: http://gis.19327.n5.nabble.com/mkgmap-ToDo-list-tp5803388p5804188.html Sent from the Mkgmap Development mailing list archive at Nabble.com.