(I’m sorry for the previous one going out without subject)
Hello,
I’ve tried to use r.surf.idw on a large raster. It works well in 6.4.4.
I’ve moved to 7.1 SVN, where everything is marvellous, at least what I use, except r.surf.idw which creates the same output as the impot was.
I’ve also svn-ed and compiled 7.0.1, where r.surf.idw did the same.
Namely:
r.surf.idw --overwrite --verbose input=output_temp1 output=output_temp2 npoints=7
when finished, output_temp2 is the same as output_temp1 and no interpolation seems to be done.
One more big difference:
6.4.4 uses less than 2G ram and finishes in a few minutes. In 7.0 and 7.1 it eats all 6GB RAM, 4 GB SWAP and runs much-much longer, understandably…
Should I stay with 6.4.4?
thanks
On Thu, May 7, 2015 at 12:28 PM, Robert Kuszinger <kuszinger@giscom.hu> wrote:
(I'm sorry for the previous one going out without subject)
Hello,
I've tried to use r.surf.idw on a large raster. It works well in 6.4.4.
I've moved to 7.1 SVN, where everything is marvellous, at least what I use,
except r.surf.idw which creates the same output as the impot was.
I've also svn-ed and compiled 7.0.1, where r.surf.idw did the same.
There are two issues open:
https://trac.osgeo.org/grass/query?status=!closed&keywords=~r.surf.idw
Namely:
r.surf.idw --overwrite --verbose input=output_temp1 output=output_temp2
npoints=7
when finished, output_temp2 is the same as output_temp1 and no interpolation
seems to be done.
One more big difference:
6.4.4 uses less than 2G ram and finishes in a few minutes. In 7.0 and 7.1 it
eats all 6GB RAM, 4 GB SWAP and runs much-much longer, understandably...
This is kind of surprising. Are you sure that the computation region
was really the same (extent + resolution)?
Could you provide us with a reproducible example?
Should I stay with 6.4.4?
Likely not 
Markus
Dear Markus,
thanks for the comments, I’m already taking part in the ticket handling mentioned above with testcase creation and testing. It seems that 7.0 problems solved at r65214, I’ll test it further.
Very good community, anyway
Robert
···
2015-05-17 19:28 GMT+02:00 Markus Neteler <neteler@osgeo.org>:
On Thu, May 7, 2015 at 12:28 PM, Robert Kuszinger <kuszinger@giscom.hu> wrote:
(I’m sorry for the previous one going out without subject)
Hello,
I’ve tried to use r.surf.idw on a large raster. It works well in 6.4.4.
I’ve moved to 7.1 SVN, where everything is marvellous, at least what I use,
except r.surf.idw which creates the same output as the impot was.
I’ve also svn-ed and compiled 7.0.1, where r.surf.idw did the same.
There are two issues open:
https://trac.osgeo.org/grass/query?status=!closed&keywords=~r.surf.idw
Namely:
r.surf.idw --overwrite --verbose input=output_temp1 output=output_temp2
npoints=7
when finished, output_temp2 is the same as output_temp1 and no interpolation
seems to be done.
One more big difference:
6.4.4 uses less than 2G ram and finishes in a few minutes. In 7.0 and 7.1 it
eats all 6GB RAM, 4 GB SWAP and runs much-much longer, understandably…
This is kind of surprising. Are you sure that the computation region
was really the same (extent + resolution)?
Could you provide us with a reproducible example?
Should I stay with 6.4.4?
Likely not
Markus