Mersenne status.txt script?

[jasong]

I've heard the saying,"Those who can't, teach." Well, maybe they should add to that,"Those who can't, beg, and sometimes beg for too much."

I have no idea how difficult this would be to implement. Basically, it would be nice if there were a script available that ran about every 24 hours(at the same time each time) on the active Mersenne number database. It could either run the query and post the results to a web page, or maybe, forgetting about the 24 hour thing, people could download the status.txt and run the script at their leisure.

The script would relay certain things, and track some statistics. If there is a file that covers old statistics of tests, maybe it would be good to correlate(my automatic spell-check feature finally approved the word, WOO-HOO!!!) the stats with the more recent information.

You could track things like first and second-pass, and within those two you could track the leading and trailing edge, 90%, 75%, 50%, 25%, and 10%(this basically means that you had done, say 90% of the first or second-pass prime exponents below that n. Alternately, instead of going by the n-values, you could go by whether that percentage of tests had been for first or second-pass, where tracking second-pass would be mean counting the number twice if it hadn't even gotten through first-pass.)

Maybe this has already been done. If so, it would be nice if the results could be posted to the web. And, yes, I know we have a web page which kind of accomplishes that. I guess I just think the more stats a project has, the more fun the project is. The Odd Perfect Number search doesn't even allow individual users or teams, but you can easily track curves, which makes the project tremendously more enjoyable.

-----------------------------------------------

I've reread my statistics request, and think I may end up confusing people, so I've come up with some examples. In the following examples we'll assume the entire range, 1-100,000 is totally primes(100,000 of them) even though we know that's not true.

Okay, first-pass tests that are JUST NOW starting are at about 99,000, second-pass is at about 50,000. Now let's say every single number below 60,000 has gone through a full test at least once. This would be your trailing edge. Now let's say that only half of the tests between 60,000 and 72,000 had gone through a full test. So 72,000 is the LOWEST number for which 90% or fewer tests have had a fully completed first-pass test. So if I'm asking for a 50% second-pass number, I'm asking for the lowest number where 50% of the tested numbers have two residues. Or, alternately, two matching residues. If any of the results for the percentages have a number higher than the leading edge of first-pass, go ahead and list it, it's not a meaningless statistic.

Even if I'm still not totally understood, it's no big deal. Other people might want other statistics, and I don't like to look like a control freak.

[petrw1]

Originally Posted by jasong

You could track things like first and second-pass, and within those two you could track the leading and trailing edge, ....

If you manually go through the assignment report on: http://www.mersenne.org/primenet/status.shtml you can see what exponents are being assigned "today" for:
Double checking: in the M19.6 Range
First time checking: in the M39.6 Range
Factoring: in the M43 Range

[jasong]

Maybe I'm reading the page wrong. I thought first-pass was around 43 million? Does 'F' mean factoring, instead of first-pass? I remember reading about that second column on the status.txt page.

/me will look for it when he's done reading the other Mersenne Forum stuff.

[davieddy]

Hey folks, we don't need two forums here.
I've just responded to petrw's table in the
Lounge "Tracking GIMPS progress.." thread.

May I suggest that this thread belongs either
there or the Primenet forum. It ain't "programming".

David

Yes. F does mean factoring/factored.

[petrw1]

Originally Posted by jasong

Maybe I'm reading the page wrong. I thought first-pass was around 43 million? Does 'F' mean factoring, instead of first-pass? I remember reading about that second column on the status.txt page.

/me will look for it when he's done reading the other Mersenne Forum stuff.

While "F" - Factoring is the first step (actually there are a few factoring steps), when I refer to "first-pass" I mean the first pass of Lucas-Lehmer testing.