[GRASSLIST:4120] i.tm.dehaze eats all my RAM

[tatel at euskalnet.net]

Hi,

When using i.tm.dehaze on a landsat scene, I get this error:

GRASS:~ > i.tm.dehaze tasscap_proba.4 p200r030.nn10 p200r030.nn20
p200r030.nn30 p200r030.nn40 p200r030.nn50 p200r030.nn70
r.stats:  100%
1 variables
68756072 observations
ERROR: Out of Memory
An error occurred. Stop.

Watching with top I can see how r.linear.regression takes more and more
memory to over 95% of RAM when error message appears

My box is an Athlon 1800XP with 1 GB DDR RAM, but kernel is not
recompiled with high memory support, so approx 900MB are recognized. It
is a Debian stable box and GRASS is compiled from CVS

Is this normal? 

I would to know if recompiling would be a valid solution

Best wishes,

Roman