[GRASS-dev] speed of r.regression.multi

Hi,

Is it correct that the time it takes to run r.regression.multi depends on the region settings (i.e., total number of raster cells in the region), and is not influenced by the number of NULL values in the raster?

I am asking because I though I could speed up calculations when masking out part area (using a mask based on x random points), but perhaps that doesn’t make sense?

Cheers

Paulo