[GRASS-user] NVIZ Seg Faults

   The raster map directory holds 2.4G of data. Trying to run nviz causes a
seg fault after processing ~3%:

GRASS 6.5.svn (Washington):~/grassdata > nviz elev=dem_basin
Loading raster map <dem_basin@buckets>...
Segmentation fault

   Anything I can do to prevent this?

Rich

Hello Rich,
with Your provided data it's not possible to tell if You can do something.
Things to check:
1) the size of Your computational region. On 32bit system You might
hit a memory limits.
2) check Your OpenGL system. Does glxinfo and glxgears work fine?
3) run nviz under gdb and generate backtrace. Still You might need to
recompile GRASS with --enable-debug to get also debug symbols for all
GRASS parts. --enable-debug is harmless and wouldn't interfere with
normal working with GRASS, it will only help to track down bugs.

Here are some notes I have made for myself on running nviz under gdb
and valgrind:
gdb $GISBASE/etc/nviz2.2/nviz
run -f "$GISBASE/etc/nviz2.2/scripts/nviz2.2_script"
elevation=elevation.dem@PERMANENT

valgrind $GISBASE/etc/nviz2.2/nviz -f
"$GISBASE/etc/nviz2.2/scripts/nviz2.2_script"
elevation=elevation.dem@PERMANENT

Maris.

2012. gada 23. marts 02:10 Rich Shepard <rshepard@appl-ecosys.com> rakstīja:

The raster map directory holds 2.4G of data. Trying to run nviz causes a
seg fault after processing ~3%:

GRASS 6.5.svn (Washington):~/grassdata > nviz elev=dem_basin
Loading raster map <dem_basin@buckets>...
Segmentation fault

Anything I can do to prevent this?

Rich

_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user

On Fri, 23 Mar 2012, Maris Nartiss wrote:

with Your provided data it's not possible to tell if You can do something.
Things to check:
1) the size of Your computational region. On 32bit system You might
hit a memory limits.

Maris,

   Yep, this is the reason, and it occurred to me while eating dinner
last night that the 3m map was too big. 80,090 rows * 96444 columns =
7.724.199.960 cells. I'll thin it to 10m and try it again.

   I'm building a new server/workstation but my existing one is the only
32-bit Slackware remaining here.

   I need only one presentation map of all four drainage basins, and the cell
resolution can be much lower for that. A few analyses will be conducted on
all four basins, but most will be on a single basin so I can probably use
the higher resolution DEM.

Rich