Hi Michael,
yes 20sec is a lot - maybe it feels longer at is actually shorter? I
never measured. But clear is it is too long. However, can't do some
stuff with other programs thanks to OJs fantastic geometry and analysis
editing functions (mostly topology related functions I am using to
create (street/sidewalk) polygons out of linework.
so I can consider it a Java memory vs OS memory swap problem...
> For me, typical ratio with a dataset having 99% geometries and 1%
> attribute is
> - shp : 13 MB
> - jml : 37 MB
> - in-memory : 31 MB
> So 8MB (jml) to 40 MB (memory) seems too much
With some panning ops it jumps from 30MB memory used to 40 for the 8 mb
file.
>
> How do you fill 200-400 MB. Do you load many 8MB JML files ?
files of different sizes... lets say 10 layers of 1-30 MB data.. its al
topographic layers: eg. pavement borders, sidewalk lines, building
outlines, trees etc. (basically a complex DWG from the city surveying
department was extracted into many dxf/shp layers)
I need to display most of them to get the object context. But only one
layer is edited.
>
>> My configuration is OJ 1.5.1, MacOSX Snow Leopard, 1GB memory.
>> As I have only 1GB of memory it could be also that this is a system
>> issue as lots of memory disk swapping occurs.
> On windows, I consider that as soon as the jvm has to use swapped
> memory, performance fall down and application often become unusable.
mhm. yes.. that is what I suspect to.
> Have you other application using cpu on the same machine ?
Yes (like email and browser) - but thats a different case here. OSX
seems to put them into a different mode, and when I switch between aps
it gives long but expected switching times to until I can activate it
again. So I have to wait for OJ even if I am not switching - that's the
point.
But actually the best answer I (we) would like to hear (maybe from Jukka
or Arndt?) that with lots of memory there is no unresponsive times.
thanks for your feedback,
stefan