Known: George announced on November 14 (shortly after noon) that has verification run on M#39 will complete November 24.
Trivial: It will thus take ten days, within six percent either way. Remembered, if possibly incorrectly: He owned, when he was developing the v21 client last winter, a P4-1.4 ghz. Assumed: Since then, he has not purchased another P4, overclocked the current one, or located a trusted friend owning a faster P4. Known: On a P4-1.4, an exponent in the 768 K FFT requires .060 seconds per iteration, and thus can be computed at 1,440,000 iterations per day. Known: to take 9.5 through 10.5 days would require an exponent between 13.68M and 15.12M iterations, where M is a decimal million. Known: Results are only just beginning to be returned to primenet at the extreme bottom of that range. Conjectured: George is not using his P4 for the verification -OR- it is slower than I remembered -OR- the new prime was found by QA -OR- for some bizarre reason, he is not operating the machine 24/7 - merely using it as his primary machine, even for heavy gaming an hour or two per day, wouldn't goof up my assumptions all that badly. -OR- my assumptions are otherwise fundamentally wrong. I suspect that one of the first two statements above will prove to be correct, since the humbers I concluded might contain the new prime are all over 4 million digits, and thus one would expect George to announce that the prime was "over 4 million" rather than "over 3.5 million" digits. Nathan _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
