I just checked the PrimeNet status page out of curiousity, to find
that only two 10M-digit numbers are available. When I looked at the
work completed figures, it became rather obvious that one particular
user is running machines that are severely misconfigured. I don't
think this is deliberate abuse, since the user does have several
machines running a few dozen exponents with nothing apparently wrong
with them.
However, the sheer number of assignments involved speaks for itself.
yeager {~} > cat status.txt | grep netconx | wc -l
2292
What I can't help wondering is whether GIMPS should have some
restriction on how many assignments can be checked out by a given
machine per unit time - in this case, the assignments in question are
being run by only two machines, which appear to be repeatedly losing
track of the work assigned to them. One likely possibility is some
sort of automated program that is repeatedly deleting or blanking the
worktodo.ini file.
Perhaps there should be hard limit, after which the user is given an
error or sent an email telling them of the situation?
In this case, there are enough exponents involved to take a top-end
system multiple centuries to complete, and there's no reason why that
should happen without someone contacting PrimeNet to make special
arrangments, if for no other reason than to always have work available
for everyone.
Nathan
_________________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
