Hi, it turned out that the index in the SeaGenerator should NOT be deleted, because it us reused for each job (which also explains what ThreadLocal is good for ;-) @WanMil: It would be quite easy to reduce the memory needs for this index. I assume ~ 1MB would be enough if we just store the information whether the grid element contains land, see, or both, and calculate the file name instead of reading it from the index. Are you still working on the mp_cut branch or do you wait for something from me? I assume that the precomp-sea handling will change when the mp cut algo changes? Gerd
Date: Thu, 10 Jan 2013 09:31:57 -0800 From: gpetermann_muenchen@hotmail.com To: mkgmap-dev@lists.mkgmap.org.uk Subject: [mkgmap-dev] reduce_peak_mem_v1.patch
Hi,
attached is another patch to reduce peak memory usage in mkgmap. - reduce initial size of typically ArrayList instances which typically contain few entries - free way after it was converted in ElementSaver - free the precomp-sea index in SeaGenerator after usage (the related HashMap requires ~17Mb) for each job. I do not yet understand why this is index is a ThreadLocal. I guess it was planned to have a single thread just for the index? I think that would be good, but I don't want to change that part now because it will probably be changed in the mp_cut branch.
Ciao, Gerd
-- View this message in context: http://gis.19327.n5.nabble.com/reduce-peak-mem-v1-patch-tp5743803.html Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://lists.mkgmap.org.uk/mailman/listinfo/mkgmap-dev