[GRASS-dev] r.sim.water at debug level 5

[Soeren Gebbert]

Hello,

2010/9/8 Hanlie Pretorius <hanlie.pretorius at gmail.com>:
> Thanks for the reply, Soeren.
>
> I have had similar results on rasters with fewer than 1 million cells.
> The process runs for a litte longer but then also gets killed.
>
> I made the resolution very fine on purpose because I was trying to
> crash r.sim.water to get to the debugging messages. Does that make
> sense?

Unfortunately not really. The process get killed because it tries to
allocate much to much memory and this may be not related to the
problems you face when computing with a fewer number of cells. The
debug messages will always be printed if the environmental variable
was set correctly.

Maybe using breakpoints while debugging will point to the problem?

Best
Soeren

>
> 2010/9/8, Soeren Gebbert <soerengebbert at googlemail.com>:
>> Hello,
>> IMHO your region resolution is much to large for r.sim.water:
>> 33008x33443 cells -> 1.1 billion cells.
>> You can use a coarse resolution or you can split your dataset in
>> smaller pieces for computation.
>>
>> r.sim.water loads all raster maps into your memory and creates
>> additionally a maximum of 7000000 walkers. I would suggest to use
>> about 1000000 cells for computation.
>>
>> Best
>> Soeren
>>
>> 2010/9/8 Hanlie Pretorius <hanlie.pretorius at gmail.com>:
>>> Hi,
>>>
>>> I've been posting about problems with r.sim.water. I have compiled
>>> GRASS64 SVN with debugging symbols and set the GRASS DEBUG environment
>>> variable to level 5. I'm also running GRASS from gdb.
>>>
>>> I ran r.sim.water on a region with 1103886544 cells and it bombed out
>>> quickly. The output was:
>>> -----
>>> GRASS 6.4.0svn (c83a_tm_29deg_e):~ > r.sim.water --overwrite
>>> elevin=c83a_dem_25m at PERMANENT
>>> dxin=c83a_dem_25m_clipped_dx at rsimwater_toetse
>>> dyin=c83a_dem_25m_clipped_dy at rsimwater_toetse
>>> rain=rain_50 at rsimwater_toetse infil=infilt_0 at rsimwater_toetse
>>> manin=manning_0.05 at rsimwater_toetse
>>> depth=depth_1mres_rain50_infil0_niter10 at rsimwater_toetse
>>> disch=disch_1mres_rain50_infil0_niter10 at rsimwater_toetse
>>> D2/5: G__read_Cell_head
>>> D2/5: G__read_Cell_head_array
>>> D3/5: region item: proj:       99
>>> D3/5: region item: zone:       0
>>> D3/5: region item: north:      -3122229.69982072
>>> D3/5: region item: south:      -3155672.30728643
>>> D3/5: region item: east:       -33474.90929218
>>> D3/5: region item: west:       -66483.27873573
>>> D3/5: region item: cols:       33008
>>> D3/5: region item: rows:       33443
>>> D3/5: region item: e-w resol:  1.00001119
>>> D3/5: region item: n-s resol:  0.99998826
>>> D3/5: region item: top:        0.5
>>> D3/5: region item: bottom:     -0.5
>>> D3/5: region item: cols3:      33008
>>> D3/5: region item: rows3:      33443
>>> D3/5: region item: depths:     1
>>> D3/5: region item: e-w resol3: 1.00001119
>>> D3/5: region item: n-s resol3: 0.99998826
>>> D3/5: region item: t-b resol:  1
>>> D3/5: rain_val is set to: -999.990000
>>>
>>> D3/5: manin_val is set to: -999.990000
>>>
>>> D3/5: infil_val is set to: -999.990000
>>>
>>> default nwalk=-2087194208, rwalk=2207773088.000000
>>> Killed
>>> -----
>>>
>>> Since GRASS itself doesn't exit abnormally, gdb doesn't seem to notice
>>> that anything is wrong.
>>>
>>> Is there anywhere else (a log file perhaps) where I can try and find
>>> out what is going on?
>>>
>>> Thanks
>>> Hanlie
>>> _______________________________________________
>>> grass-dev mailing list
>>> grass-dev at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/grass-dev
>>>
>>
>