Opened 9 years ago

Last modified 5 years ago

#2672 new defect

r.surf.idw should support FCELL and DCELL

Reported by: annakrat Owned by: grass-dev@…
Priority: normal Milestone: 7.0.7
Component: Raster Version: svn-releasebranch70
Keywords: r.surf.idw Cc:
CPU: All Platform: All

Description

Currently r.surf.idw reads and outputs raster as CELL. Float and double precision support is needed.

Change History (12)

comment:1 by neteler, 9 years ago

For now I have added a G_fatal_error(): r65260 + r65261: exit with fatal error in case of FCELL or DCELL input.

comment:2 by neteler, 9 years ago

Once fixed, the commented example (see r65262) can be activated in the manual

comment:3 by kuszinger, 9 years ago

Comment only: Any solution is generated, I've realized that in V7 r.surf.idw eats much more memory than in 6.4. Addition of FCELL and DCELL is handy, but I'm afraid that it may lead to further memory footprint growing. I'm ready for testing, anyway, any new version. I use r.surf.idw extensively on huge sparse rasters where it takes 2 - 10 GB RAM to create the result.

When it is about an FCELL source my workaround is creating a temporal CELL coverage with mapcalc and multiplication and rounding original values and when idw is done I calculate it back to an FCELL and remove the temporal map.

Last edited 9 years ago by kuszinger (previous) (diff)

in reply to:  3 comment:4 by neteler, 9 years ago

Replying to kuszinger:

Comment only: Any solution is generated, I've realized that in V7 r.surf.idw eats much more memory than in 6.4. Addition of FCELL and DCELL is handy, but I'm afraid that it may lead to further memory footprint growing. I'm ready for testing, anyway, any new version.

I made a test with valgrind (http://grasswiki.osgeo.org/wiki/GRASS_Debugging#Using_Valgrind) and did not find any leak:

# current region:
rows:       7490
cols:       6990
cells:      52355100

# valgrind ...
==29843== LEAK SUMMARY:
==29843==    definitely lost: 660 bytes in 24 blocks
==29843==    indirectly lost: 0 bytes in 0 blocks
==29843==      possibly lost: 0 bytes in 0 blocks
==29843==    still reachable: 101,344 bytes in 52 blocks
==29843==         suppressed: 0 bytes in 0 blocks

I use r.surf.idw extensively on huge sparse rasters where it takes 2 - 10 GB RAM to create the result.

Please indicate how many cols/rows or ideally post a test with the NC dataset.

comment:5 by neteler, 9 years ago

Milestone: 7.0.17.0.2

Ticket retargeted after 7.0.1 milestone closed

comment:6 by neteler, 9 years ago

Milestone: 7.0.27.0.3

Ticket retargeted after milestone closed

comment:7 by neteler, 8 years ago

Milestone: 7.0.3

Ticket retargeted after milestone closed

comment:8 by neteler, 8 years ago

Milestone: 7.0.4

Ticket retargeted after 7.0.3 milestone closed

comment:9 by martinl, 8 years ago

Milestone: 7.0.47.0.5

comment:10 by neteler, 8 years ago

Milestone: 7.0.57.0.6

comment:11 by neteler, 6 years ago

Milestone: 7.0.67.0.7

comment:12 by martinl, 5 years ago

Still relevant?

Note: See TracTickets for help on using tickets.