mersenneforum.org Sieving for CRUS
 Register FAQ Search Today's Posts Mark Forums Read

 2019-09-26, 08:05 #551 pepi37     Dec 2011 After milion nines:) 2×5×131 Posts R463and S463 are sieved right now: from 10000 WU , 1364 workunits is done, finding over 13000 factors. S 896 will be processed after S/R 463 is over
 2019-09-27, 22:16 #552 pepi37     Dec 2011 After milion nines:) 2·5·131 Posts It was not S896 ( it is proven) so I was little confused S892 is over, sent to Gary via mail
2019-09-28, 01:16   #553
gd_barnes

May 2007
Kansas; USA

26×3×53 Posts

Quote:
 Originally Posted by pepi37 It was not S896 ( it is proven) so I was little confused S892 is over, sent to Gary via mail

 2019-09-28, 01:20 #554 gd_barnes     May 2007 Kansas; USA 1017610 Posts Pepi, For a base with 2 k's remaining, what sieve depth would Yoyo need as a starting point for sieving n=200K to 1M ? I will likely do a large amount of this initial sieving for such bases and then forward them on to you for sending to Yoyo. Thanks, Gary
2019-09-30, 10:07   #555
pepi37

Dec 2011
After milion nines:)

2×5×131 Posts

Quote:
 Originally Posted by gd_barnes Pepi, For a base with 2 k's remaining, what sieve depth would Yoyo need as a starting point for sieving n=200K to 1M ? I will likely do a large amount of this initial sieving for such bases and then forward them on to you for sending to Yoyo. Thanks, Gary

Gary you dont need to do initial sieving: I wrote small script that do most of job: so it is pretty automated thing.
So I will put focus on 2 K bases, R - S side and send to YOYO.
Maybe important question is "optimal" sieve depth: if those sequences will be processed with REB , then lower sieve depth is not problem: but "average " user can get benefit from deeper sieving.

It is huge "distance" in candidate that have 200K as exponent and 1M as exponent :)

2019-09-30, 19:32   #556
gd_barnes

May 2007
Kansas; USA

100111110000002 Posts

Quote:
 Originally Posted by pepi37 Gary you dont need to do initial sieving: I wrote small script that do most of job: so it is pretty automated thing. So I will put focus on 2 K bases, R - S side and send to YOYO. Maybe important question is "optimal" sieve depth: if those sequences will be processed with REB , then lower sieve depth is not problem: but "average " user can get benefit from deeper sieving. It is huge "distance" in candidate that have 200K as exponent and 1M as exponent :)

OK great. The focus would be on 2 k bases that are presently only tested to n=200K or 250K.

If you think that n=200K-1M is too much sieve range, you can choose n=200K-500K instead. My thinking in choosing n=200K-1M is that if a prime is found, the base would be 1 k...and it is preferable to have 1 k bases sieved up to n=1M.

2019-09-30, 20:10   #557
pepi37

Dec 2011
After milion nines:)

2·5·131 Posts

Quote:
 Originally Posted by gd_barnes OK great. The focus would be on 2 k bases that are presently only tested to n=200K or 250K. If you think that n=200K-1M is too much sieve range, you can choose n=200K-500K instead. My thinking in choosing n=200K-1M is that if a prime is found, the base would be 1 k...and it is preferable to have 1 k bases sieved up to n=1M.

So be it: from now on focus in on two K bases (200-250k ) at start.
I also share you idea to sieve up to 1M but then I will set sieve depth to 5e14. That sieve depth is ok as far for BOINC as for individuals

2019-09-30, 20:20   #558
MisterBitcoin

"Nuri, the dragon :P"
Jul 2016
Good old Germany

10110101112 Posts

Quote:
 Originally Posted by pepi37 So be it: from now on focus in on two K bases (200-250k ) at start. I also share you idea to sieve up to 1M but then I will set sieve depth to 5e14. That sieve depth is ok as far for BOINC as for individuals

I´d like to add some other suggestion here. Focus on the bases with the lowest "difficulty" as shown in the unproven table.

I will focus my work on those with less than 20K from now on.

Edit: BTW, how is this value calculated anyway?

Last fiddled with by MisterBitcoin on 2019-09-30 at 20:22

2019-10-01, 00:49   #559
gd_barnes

May 2007
Kansas; USA

26·3·53 Posts

Quote:
 Originally Posted by MisterBitcoin I´d like to add some other suggestion here. Focus on the bases with the lowest "difficulty" as shown in the unproven table. I will focus my work on those with less than 20K from now on. Edit: BTW, how is this value calculated anyway?
Good focus area.

Mark came up with the formula. It may be something like the candidates remaining on a sieve to P=1T multiplied by the log of (the base multiplied by search depth squared) and then all divided by some fixed amount to make it reasonable. It effectively gives an exact mathematical difficulty of how long it would take to increase the testing of a base by some small n amount. If you double the test depth you 4x the difficulty (if no primes found).

Last fiddled with by gd_barnes on 2019-10-01 at 00:57

2019-10-10, 19:04   #560
rogue

"Mark"
Apr 2003
Between here and the

3·1,951 Posts

Quote:
 Originally Posted by pepi37 R463and S463 are sieved right now: from 10000 WU , 1364 workunits is done, finding over 13000 factors.
What is the status of R463/S463?

2019-10-10, 20:07   #561
gd_barnes

May 2007
Kansas; USA

1017610 Posts

Quote:
 Originally Posted by rogue What is the status of R463/S463?

https://www.rechenkraft.net/yoyo/y_status_sieve.php

It looks like there is 1 work unit left as of this posting. So it will be done anytime now.

Pepi, can you send this file containing both bases to me soon after it is complete? I will reserve R463 and Mark will reserve S463. Thanks.

Last fiddled with by gd_barnes on 2019-10-10 at 20:08

 Similar Threads Thread Thread Starter Forum Replies Last Post gd_barnes Conjectures 'R Us 143 2014-10-21 23:55 vmod Conjectures 'R Us 213 2014-02-28 21:23 rogue Conjectures 'R Us 35 2013-11-09 09:03 CRGreathouse Puzzles 24 2011-10-28 18:30 Mini-Geek Conjectures 'R Us 1 2010-11-08 20:50

All times are UTC. The time now is 00:51.

Thu Aug 13 00:51:49 UTC 2020 up 26 days, 20:38, 0 users, load averages: 1.70, 1.69, 1.59