Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

HI Massimo, To avoid all the osgdem tasks from having to parse all the source files VPB supports generating a cache of all the geospatial details of the source files. The vpbcache tool can be used to read the original source file and then generate a cache file from this. You then provide this

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

Hi all, i work with OSG for fun, and i have created a DataBase for whole Italy at HiRes 1m/pixel in about 3 weeks. The final DB is about 600G. The Source Images are about 13000 GeoTiff files (700G), and about 10 DEM files (40 meters res, 2.5G) I have used and old notebook dual core 1.6Ghz, and

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

J.P. Delport wrote: The vpbmaster command created 473 tasks and after 50 hours of processing created 1.5 million files with a total size of 487GB. The input data is around 800GB if I remember OK. So the 60 days does sound a bit extreme. In our case, VPB created about 130 000 tasks...

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

zonk wrote: Hi Luc, At first: Are you really really sure you want to use windows? Don't misunderstand me, I'm not a linux-geek, I use myself windows for nearly everything, but VPB is one of the things which are running much faster on linux. We can use Linux, no problem with that,

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

Hi Luc, On Wed, Jan 13, 2010 at 10:33 AM, Luc Claustres luc.claust...@vegatechnologies.fr wrote: That is a possible issue as we use external USB drives with a NTFS file system. This is your problem. USB really isn't a good solution for IO bound applications. NTFS file system isn't a good

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

Hi Luc, Luc Claustres wrote: J.P. Delport wrote: The vpbmaster command created 473 tasks and after 50 hours of processing created 1.5 million files with a total size of 487GB. The input data is around 800GB if I remember OK. So the 60 days does sound a bit extreme. In our case, VPB

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

Hi, J.P. Delport wrote: Hi Luc, Luc Claustres wrote: J.P. Delport wrote: The vpbmaster command created 473 tasks and after 50 hours of processing created 1.5 million files with a total size of 487GB. The input data is around 800GB if I remember OK. So the 60 days does sound a bit extreme.

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

In order to generate less tasks I also played with the --tile-image-size and --tile-terrain-size options. With a value of 1024 VPB created about 8 000 tasks, and 2 000 with a value of 2048. If I/O is the bottleneck I wonder if this would be faster to work with bigger tiles. According to the

[osg-users] [vpb] Are we too optimistic about large database generation ?

Hi, After different positive results using VPB to generate world databases (at 450m/150m imagery resolution and 300m/200m DEM resolution), we have tried something more ambitious. Our input data are now : - a world imagery cover at 0.5s (~15m) : 72 ECW files ~ 94 GB - a world DEM cover at 3s

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

Hi Luc, Luc Claustres wrote: Hi, After different positive results using VPB to generate world databases (at 450m/150m imagery resolution and 300m/200m DEM resolution), we have tried something more ambitious. Our input data are now : - a world imagery cover at 0.5s (~15m) : 72 ECW files ~ 94

Re: [osg-users] [vpb] Are we too optimistic about large database generation ?

Hi Luc, I had problems like you, but I managed to discover a setup which works without problems, and renders large databases. At first: Are you really really sure you want to use windows? Don't misunderstand me, I'm not a linux-geek, I use myself windows for nearly everything, but VPB is one