logo separator

[mkgmap-dev] splitter that generates problem list

From Gerd Petermann gpetermann_muenchen at hotmail.com on Wed Nov 7 21:43:56 GMT 2012



> Date: Wed, 7 Nov 2012 22:16:50 +0100
> From: osm at aighes.de
> To: mkgmap-dev at lists.mkgmap.org.uk
> Subject: Re: [mkgmap-dev] splitter that generates problem list
> 
> Am 07.11.2012 19:19, schrieb Henning Scholland:
> > Am 07.11.2012 19:06, schrieb WanMil:
> > If I believe in TaskManager, splitter takes 17% CPU-usage. This means
> > one core is completly busy.
> I have to correct this. splitter uses more cores while writing the tiles.

yes, and up to now this is the only part of the program that uses multiple threads 
to reduce runtime.

> > HDD as bottleneck was only meant in general. I'm running splitter and
> > mkgmap (and all input and output) on a SSD. But I think this is not
> > typical. Maybe it would increase performance, if max-area keep smaller
> > and splitter runs multiple times. E.g. I run two splitter with
> > max-area=512. Will try this.
> I tried it and it wasn't better. It tooks 10 min longer (1:30 to 1:20 
> h). Reducing max-area from 1024 to 512 doesn't save enough RAM, that two 
> splitter instances have enough RAM. So I think it would be more 
> effective to use threading also in analysing the polygons e.g one threat 
> for reading pbf and another one for checking the object, don't know if 
> this is possible.

OK. Reg. threading: Maybe some routines in the problem-list branch can be 
devided to use multiple threads, but I don't see a way to process the input 
out of sequence, but that's optimization which could be done later.
My proposed solution for higher speed is to use the o5m format (for the input file), 
but it seems that  wants to jump on this train ?

Gerd


 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mkgmap.org.uk/pipermail/mkgmap-dev/attachments/20121107/61f8ad48/attachment.html 


More information about the mkgmap-dev mailing list