From: ijw11@phx.cam.ac.uk (Ian Wells) Subject: Re: Abysmal disk performance Date: 16 May 1992 14:44:04 GMT
Thankyou everybody for telling me the term is 'thrashing' (I got one
or two mail messages as well). I always thought that that referred
solely to when process memory was being paged/swapped to disk, rather
than general hard-disk seeking.
The reason I asked about whether 2 separately-loaded processes share
their data space is that then, when a process ends, its memory need
only be freed when it is next needed. In that way, when a make is
done and gcc is called sequentially many times, the areas of the
executable in memory that haven't been changed since loading could
then be kept until the next time it runs (since there is no
significant activity between calls to GCC in a make, mostly). It
should save quite a bit of disk access.
Ian.