The x86 FPU is unique with its 64 bits of precision, and
is the only architecture capable of accurately testing the
larger candidates in each range, which are the ones I'm
begining to hand out. Can all non-x86 people (probably only
Thomas and I) please turn on doinfo=1 and verbose=1 in their
client config file, and keep a close tab on their maximum
FFT errors. If you ever see even 1 error > 0.3125, then it's
pretty much game over. (Non-x86 seem to fail castrophically,
some errors >0.3125 are not so fatal on x86s.)

Such machines would make great sievers for project 101, for
example, or as ECM clients for my Aurefeuillian factorisation
project:
http://fatphil.org/maths/palindrome/101.html
http://62.236.152.54:8195/query.cgi?port=8192
http://62.236.152.54:8195/cands.cgi?port=8192

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!?
Declare Yourself - Register online to vote today!
http://vote.yahoo.com