[GRASS-user] v.surf.rst limitations?

[roberto.antolin at polimi.it]

Hi Michael and all

Michael Perdue <michael_perdue at yahoo.ca> ha escrito:

> Hey Roberto,
> Can you tell me more about this bug? I ran a tile through   
> v.surf.bspline that was significantly larger than this and I never   
> ran into any problems. Nor can I find any defects in the output.
...
> This grid was 33mil cells with ~30mil input data points.
>
Good to know! :)  I had some problems with memory allocation weeks ago  
and I was probably wrong about the number of cell v.surf.bspline can  
manage. I'll check again. Thank you for the information.

> However, I do receive a seg fault when I try to define a set of   
> sparce data nodes to be interolated (input_ext=).
>
Yes, this is one of the "many" bugs. But, at least, this is also  
fixed. A new patch will be available tomorrow (I think).  This new  
patch will also include a best stdout info. Well, it will actually  
include stdout info :-P

Just a curiosity, how much time did it take to run? Maybe the new  
version makes it to run faster ;-) Let me know, please.

> Cheers,
>
> Mike
>
> PS Yes, v.surf.bspline is much faster than v.surf.rst. But more   
> importantly, it seems to step through the data to prevent memory   
> allocation errors. Is this true? Nice!!! :-)
>
Yes, that's totally true. If the current region is too large, it  
divides the whole region into smaller regions and it works with them  
one at a time. The module also considered overlapped zones between  
sub-regions to avoid discontinuities in the dtm heights.

Regards,
Roberto.

PS: I think the mailing list didn't receive my last mail, did it?. If  
that's right, I'm sorry. I have tons of problems with my mail server.  
However, you can find it below.


> At this moment, there is actually a limitation with v.surf.bspline.
> There is a bug and I think it is no possible working with 13.5 mill
> cells. I'm working on it, but by the moment It will reach only 4mill
> cells. Yes, I know it is no enough, sorry. I hope that in a near future
> this will be also fixed. Maybe, if you try with a lower resolution
> (about 20m) v.surf.bspline should work.
>
> However, I think v.surf.bspline will be faster than v.surf.rst. ;-)
>
> Regards,
> Roberto.
>