Today is a western holiday (labor day) so I don't think this really counts as a normal Monday stat. but I will leave that determination up to the division of statistics and holiday attenuation :-)

I would not presume to argue stats with the Sek master but Boinc stats shows USA contributing 36.7% of WCG points so I think when there is a holiday it may be necesary to consider that as a factor regardless of actual population.

Holiday shmoladay--their all the same to me --i leave ours on 24/7

What's needed to be precise is an afternoon session of 141 after a strong 147 matinée, i.e. room for a 6 year drop.

Last Friday we had a 144 / 140 / 284 i.e. a drop of 4 years in the second half. The range of part 2 seems to lie between +2 and -7, average -3.6.

Curiously the actual is higher than the trailing:

20090904 1955.84 F
20090905 1970.03 S
20090906 1984.56 S
20090907 1989.58 M
20090908 1995.99 T
20090909 1998.28 W
20090910 1997.94 T

But somehow I think tomorrow is better to scale this milestone.

PS Steve WCG, the WCG own data says ~31% few days ago and yesterday 31.2%, but I knew that when bringing up the 4% population labor day celebration point ;>).

I'm not sure how Willy does the averages, rolling or sliding, but 5% is a huge difference. On WCG total TFlop he's more than a rounding factor behind. We show 312 TFL 60 day mean, he shows 307. That's on 60 days a whopping 300TFL cumulative missing... so this is where I ignore BOINCstats on anything beyond individual.

Of course my numbers could be wrong, but then other cross-checks would fail as well. The 60 day points average is 218 million. Divide that by 700,000 and you get 312 TFL again. Who knows does he get the real Fpops from somewhere but beats me where that could be. Never tried to add up the hosts fpops/fiops exported and we know that hosts file is incomplete given all hidden devices are excluded and that's thousands

Fatto, it became 2004 CPU years in a Week (7 day consecutive) period.

Applause rolling around the grid



(chart above updated).