No subject

[chuck at castor.GEOG.UCSB.EDU]

>Can someone tell me if there is an "ideal" setting for swap space 
>(relative to the amount of RAM) for running GRASS, particularly for 
>r.watershed?  I suspect I don't have mine set optimally.  I'm using 
>the 2 X RAM rule-of-thumb.

You can assume that MOST GRASS commands will only take up a few megs
of memory each. r.watershed is not one of those programs. r.watershed
is a memory hog that grabs an amount of memory proportional to the
size of the elevation map you are analyzing. The amount of memory
depends on the number and type of output maps required.

The "seg" part of r.watershed will allow you solve problems that take
up more memory than your RAM and swap space if: 1) you have enough hard
disk space to hold the segmented data, and 2) you have PLENTY of time
to solve the problem. Seg is ssssssssllllllllllloooooooooowwwwwwwww.

Chuck Ehlschlaeger