Whilst I'm debating declaring 2004 "worst year ever", in the small
field of prime-finding, it's been pretty good. I was expecting one
apple by the end of the year, and Bob's pile of legacy machines
found 4 - which was nice.

We've been pipped into 3rd place by count by the rather amorphous
15k-1 "project" (since when were 23 or 7 multiples of 15?), and
into 5th place by the Riesel search, which I know has about 10
times the CPU capability of us combined, so there's no point even
trying to compete/compare. The above sounds like a cloud, but there's
a silver lining - the #1/count project (GFN14) will be losing primes
nearly as fast as we did (a couple of hundred in a few months) by
the end of the year, and the #2/count project will be leaking primes
slowly and steadily likewise. Our main bulk is at the ~100000-digit
range, which are pretty safe for quite a while, and may well creep
back up to the #1 slot if we're lucky. The #4/score project will
at some point see a slow-down in their prime finding, and aren't
guaranteed to find a huge number more primes. However, the ones that
they do find will be big, so they've almost certainly got an
unassailable lead. However, it'll be nice to see the apples come in
and chip away slowly at their lead.

Steven and Predrag are now hopefully moved over to the apple range,
again using manual reservation of ranges rather than the client/server
setup. Is everything working, guys?

The peaches may be approaching the danger zone maxerr-wise. I say
"may" because I don't know, because no-one's turned on "info"
reporting in their client config files. Can someone with an
Athlon and someone with a P4 please turn that on. The reason I
(gigaspaz and nonospaz) am not doing it is because I'm using my
two machines to test the maxerr of the new program - MaximEis,
rather than ForEis which is what everyone else is using for peaches.

The 'drop dead' error is sin(pi/6)/2 = 0.433. The reason for this
is that rounding needs to take place in a hexagonal grid rather than
the usual rectangular one.

MaximEis is presently bubbling around 0.315-0.333 which is quite
manageable. It would be really nice if I could get some ForEis
comparisons, pretty please? I'm prepared to load up the server
with candidates (I sieved extra-wide, just in case) for as long
as people want to "risk" testing them. I'd not go much beyond
0.4 personally.

Phil

=====
When inserting a CD, hold down shift to stop the AutoRun feature
In the Device Manager, disable the SbcpHid device.
http://www.cs.princeton.edu/~jhalderm/cd3/



__________________________________
Do you Yahoo!?
Send a seasonal email greeting and help others. Do good.
http://celebrity.mail.yahoo.com