[Qgis-user] contour plugin

[Alister Hood]

Hi again,

> Date: Thu, 12 Jul 2012 10:23:13 -0400
> From: "David M. Lawrence" <dave at fuzzo.com>
> To: qgis-user at lists.osgeo.org
> Subject: Re: [Qgis-user] contour plugin
> Message-ID: <4FFEDDD1.6060408 at fuzzo.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> Dear all:
> 
> I tried all of this -- contour does work on my machine with a dataset I
> downloaded from an online tutorial, so I think the problem may be with
> my data instead.  The tutorial's data were obtained from an evenly
> spaced grid.
> 
> My data are not evenly spaced. They are soundings (depth measurements)
> from a quarry linked to geographic data obtained via a GPS receiver.
> The measurements are irregularly spaced.

For what it's worth, I've only ever tried the contour plugin with irregular points.
But it may be very slow for datasets that size.  Try what David suggested; it should be fast:

Raster/Analysis/Grid
Raster/Extraction/Contour

> The latest version of the data set has more than 900 points covering 75
> percent of the aerial extent of a quarry.  I've been running the contour
> procedure for 12 hours now.  The contour dialog box indicates it is "not
> responding," but I know from other software that that may mean the
> contour procedure is actually thinking.
> 
> Every time I check Task Manager (I'm running QGIS Wroclaw on Windows
> 7-64) the number of processes seems to fluctuate around 50, with an
> observed range of 42 to 52.  The amount of memory consumed ranges from
> 68,000 Kb to a little more than 100,000 Kb.

It's worth looking at the CPU column (you might need to use View>Select columns).  If it is actually working, QGIS should almost fully use 1 processor/core i.e. it should show up as 50% on a dual core machine.  If it shows up not using any CPU then something has gone wrong.  (It wouldn't be the case here, but sometimes you can identify that a task has a bottleneck reading data across a network or something, because it is only using a small amount of CPU.)
 
>  From this, I consider the hypothesis that contour is working more
> likely than the hypothesis that it has frozen.  Comments, anyone?

Yes, I think so.
 
Regards,
Alister