All things doublecheck!! - Page 4

axn · 2006-09-04, 17:22

Quote:

Originally Posted by ltd

Is it possible that we split up the double check things into another thread?

Your wish is my command

michaf · 2006-09-17, 18:36

posted september 2:

Quote:

Originally Posted by axn1

Meh! I wanted to do that myself. As michaf noted, the "min n untested" column has been bugging me to no end. In fact I would suggest that you go ahead and do the same for Riesels also. After one week, we can release 20k < n < 100k for all k's (both S & R).

I'd reckon we can release those now?

ltd · 2006-09-17, 20:42

I would like to wait until the end of the month.

This has the reason that the llrnet server has some minor problems when
the test are very small and there is some adjustement needed with my internal queue management. So i would like to be able to react when there
is a problem but my job forces me to be out of town for the rest of the month (not on the weekends).
In oktober i should be available the whole month. (I hope

)

Cheers,

Lars

michaf · 2006-09-17, 20:56

Would be perfectly fine;
In the meantime, we'll just have to find three or four primes to make your life easier :)

jaat · 2006-09-20, 19:57

why is the min-untested at 20,000 for each of the k's?

jaat

michaf · 2006-09-20, 20:04

Quote:

Originally Posted by jaat

why is the min-untested at 20,000 for each of the k's?

jaat

There are a lot of numbers of which we have no residues (yet)
We are waiting for them to be submitted, if any are left,
but the candidates are released for (re)testing at the beginning of next month, since some minor glitches in the server require manual inspection I believe.

So in short, they will be dealt with in short time :>

(In fact, I can't wait to see those quick results again... at n=20k, the tests take less time then the communication with the server...)

jaat · 2006-09-20, 20:12

What looked particularly weird was that DC min-untested is higher than min-untested.

michaf · 2006-09-20, 20:14

I think that was because masser? had some check-ins for dc, and the first time tests still don't have residues. Ah well.. all will be sorted in a month or so.

ltd · 2006-09-22, 20:40

I have started to insert the low n without proper residue into the queues.

michaf · 2006-09-23, 18:46

That makes another nice spike in the stats;
Just in time before the other huge one disappears :>

We'll be catching up quickly :) thanks

ltd · 2006-09-24, 13:09

Some information from the database we have 103577 numbers in the DB for which no factor is know and a doublecheck test is performed.

We have no errors at the moment. :surprised

So axn1 and i decided to make the next steps for recovering a little bit different.

We will run small ranges ( n difference of 5000 (like in n between 60000 and 65000) every one or two weeks.

We will at first focus on the ranges where there are no knows dc results as the chance to find a prime there is bigger.
That means we will start with ranges>60000 next weekend.

The stats show at the moment for first pass only tests as minimum untested when the test is higher then 100000. That gives a cleaner look and there is no question "why are dc tests larger the first pass tests" anymore.

Lars

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Doublecheck always have shifted S0 value?	ATH	PrimeNet	11	2010-06-03 06:38
BIG Doublecheck finished OK this morning!	NBtarheel_33	Lounge	0	2008-11-16 20:58
Help needed for doublecheck sieving!	mdettweiler	No Prime Left Behind	188	2008-03-25 16:31
DoubleCheck vs LL assignments	Unregistered	PrimeNet	9	2006-03-26 05:48
doublecheck - results	TheJudger	Data	4	2005-04-04 08:54

2006-09-17, 20:42	#36
ltd Apr 2003 1100000100₂ Posts	I would like to wait until the end of the month. This has the reason that the llrnet server has some minor problems when the test are very small and there is some adjustement needed with my internal queue management. So i would like to be able to react when there is a problem but my job forces me to be out of town for the rest of the month (not on the weekends). In oktober i should be available the whole month. (I hope ) Cheers, Lars

2006-09-17, 20:56	#37
michaf Jan 2005 479 Posts	Would be perfectly fine; In the meantime, we'll just have to find three or four primes to make your life easier :)

2006-09-20, 19:57	#38
jaat Jul 2005 2·3² Posts	why is the min-untested at 20,000 for each of the k's? jaat

2006-09-20, 20:12	#40
jaat Jul 2005 2×3² Posts	What looked particularly weird was that DC min-untested is higher than min-untested.

2006-09-20, 20:14	#41
michaf Jan 2005 737₈ Posts	I think that was because masser? had some check-ins for dc, and the first time tests still don't have residues. Ah well.. all will be sorted in a month or so.

2006-09-22, 20:40	#42
ltd Apr 2003 2²·193 Posts	I have started to insert the low n without proper residue into the queues.

2006-09-23, 18:46	#43
michaf Jan 2005 479 Posts	That makes another nice spike in the stats; Just in time before the other huge one disappears :> We'll be catching up quickly :) thanks

2006-09-24, 13:09	#44
ltd Apr 2003 2²·193 Posts	Some information from the database we have 103577 numbers in the DB for which no factor is know and a doublecheck test is performed. We have no errors at the moment. :surprised So axn1 and i decided to make the next steps for recovering a little bit different. We will run small ranges ( n difference of 5000 (like in n between 60000 and 65000) every one or two weeks. We will at first focus on the ranges where there are no knows dc results as the chance to find a prime there is bigger. That means we will start with ranges>60000 next weekend. The stats show at the moment for first pass only tests as minimum untested when the test is higher then 100000. That gives a cleaner look and there is no question "why are dc tests larger the first pass tests" anymore. Lars