Hi,
I am running v.distance on a massive point vector which represents coastline vertices, uploading the attributes of another point vector, in this case, the vectorized point version of a DEM slope raster.
While the program is working as expected, (I have seen the output on smaller files, and it looks correct), it is extremely slow because the ‘from’ vector has ~ 700k points in it, and my search radius is 200 m for each point. I checked htop and v.distance is only using one CPU. The process has been running for about a week, and is at 69%. I was wondering if there were any way to make v.distance use more than one CPU (I have four)? Or perhaps there is an environment variable I’m not aware of?
~ Eric.
On 03/11/17 14:45, Eric Patton wrote:
Hi,
I am running v.distance on a massive point vector which represents coastline vertices, uploading the attributes of another point vector, in this case, the vectorized point version of a DEM slope raster.
While the program is working as expected, (I have seen the output on smaller files, and it looks correct), it is extremely slow because the ‘from’ vector has ~ 700k points in it, and my search radius is 200 m for each point. I checked htop and v.distance is only using one CPU. The process has been running for about a week, and is at 69%.
It would be interesting to find out if this is due to the actual distance algorithm, or more to database access. You might want to try using the -p flag and redirect the output to a file just to see if this is faster.
I was wondering if there were any way to make v.distance use more than one CPU (I have four)? Or perhaps there is an environment variable I’m not aware of?
AFAIK there is no parallelization in v.distance.
Moritz