[Tile-serving] [osm2pgsql] Clustering strategies (#208)

[Paul Norman]

> I've limited maxchar to 6 which seems enough to get close to higher zoom
> metatiles.

But do you have any data on either clustering time or rendering speed?

The one test I did showed no change in index creation speed (1328s vs 1330s) for 10 vs 8. Clustering was slightly faster, with 46404s vs 45811s, which is 1.3%.

I'm pretty sure that less disk pages will need to be retrieved with 8 than 6, and there are even gains past that - remember, WGS84 splitting (geohash) doesn't line up with either meta-tiles or buffered meta-tiles.

I should have results tomorrow for Measure option 3 and 4 creation times.

---
Reply to this email directly or view it on GitHub:
https://github.com/openstreetmap/osm2pgsql/issues/208#issuecomment-62333456
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tile-serving/attachments/20141109/3c841d37/attachment.html>