mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > Five or Bust - The Dual Sierpinski Problem

Reply
 
Thread Tools
Old 2010-01-27, 17:37   #56
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT/BST)

2×2,861 Posts
Default

Now you are down to one remaining sequence why haven't you moved to sr1sieve?
henryzz is offline   Reply With Quote
Old 2010-01-27, 18:52   #57
philmoore
 
philmoore's Avatar
 
"Phil"
Sep 2002
Tracktown, U.S.A.

1,117 Posts
Default

Awesome! There are now 148733 entries left in the sieve file, and I am uploading a new sieve file today in celebration. The average number of entries per work file is now just below 31. I think we have eliminated another 2% since the last sieve file was uploaded in November! Thanks everyone, especially the enormous effort by enderak!

We did shrink the sieve file size by about 4% when we quit sieving on the range 0-2M, but this is still an impressive effort. The original 5-sequence sieve file was around 800,000 entries, I think. We have now completed sieving to 50 bits.

Henryzz, sr1sieve is not set up to sieve the dual form yet. Geoff is swamped at the moment and not able to add the code, but he thinks that he might be able to squeeze a few percent greater efficiency when he does.
philmoore is offline   Reply With Quote
Old 2010-01-27, 23:54   #58
enderak
 
enderak's Avatar
 
Feb 2009

3×13 Posts
Default

Well, I didn't start on this project until after it was down to the last k, so I think my sieving numbers are a bit misleading when compared to the amount of actual time and work put in. I am glad to help in any case. Here's to hoping that most of our sieving work goes unneeded! :)
enderak is offline   Reply With Quote
Old 2010-02-07, 08:57   #59
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

237338 Posts
Default

Are you guys still using sr2sieve for this effort? Just checking because sr1sieve will be far faster for the one k that is remaining.
gd_barnes is offline   Reply With Quote
Old 2010-02-07, 11:23   #60
philmoore
 
philmoore's Avatar
 
"Phil"
Sep 2002
Tracktown, U.S.A.

1,117 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
Are you guys still using sr2sieve for this effort? Just checking because sr1sieve will be far faster for the one k that is remaining.
See my previous post in this thread. Geoff thinks that sr2sieve is running very close to the maximum speed possible with sr1sieve, but might be able to get another 3-5% increase when he eventually adds the dual form to sr1sieve.
philmoore is offline   Reply With Quote
Old 2010-02-07, 11:49   #61
philmoore
 
philmoore's Avatar
 
"Phil"
Sep 2002
Tracktown, U.S.A.

1,117 Posts
Default

Here is a suggestion from Rincewind:

Quote:
Originally Posted by Rincewind
Hi,

I think I've found a way to use sr2sieve with less switches.
We can use the standard range-file 'sr2work.txt'. All we have to do is add 3 zeros ('000') to the ranges we want to sieve.
My range-file for my reserved Range (1210 - 1211) has the following content:

"1210000,1211000"

With this file I can start sr2sieve without the -p and the -P switches. All I need to add is the -i and the -Q switches:

"./sr2sieve -i dual_s.abcd -zz -Q720"

Plus the program writes checkpoints and can be restarted with the same command.
I'm not sure if this is some really new news, but imho this line looks much friendlier to new sievers.

Have a nice day
Rincewind
philmoore is offline   Reply With Quote
Old 2010-02-07, 13:20   #62
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

17×251 Posts
Default

Quote:
Originally Posted by philmoore View Post
Here is a suggestion from Rincewind:
If you're looking to minimize or eliminate the number of switches to pass, read this: (from sr2sieve's readme.txt)
Code:
If no command line arguments are given but `sr2sieve-command-line.txt'
exists in the current directory, then the command line will be as if the
first line of this file had been used to invoke sr2sieve. This may be useful
on some GUI machines where the command shell and batch files have been
disabled for security reasons.
Also, if you rename dual_s.abcd to sr2data.txt, it will read it from that file without any -i command.
Mini-Geek is offline   Reply With Quote
Old 2010-02-09, 17:13   #63
Rincewind
 
Rincewind's Avatar
 
Oct 2006

103 Posts
Default

This sounds good, it's like having a little shell-scrip without the script.
Just write the whole command-line with switches an such things and all following starts are much easier.

Can someone explain me why we use the -Q switch or what this switch makes.
Rincewind is offline   Reply With Quote
Old 2010-02-09, 18:05   #64
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

426710 Posts
Default

Quote:
Originally Posted by Rincewind View Post
Can someone explain me why we use the -Q switch or what this switch makes.
Code:
-Q --subseq Q         Force sieving k*b^n+c as subsequences (k*b^d)*(b^Q)^m+c.
I'm guessing it makes the sieving for this project faster for some reason.
Mini-Geek is offline   Reply With Quote
Old 2010-02-10, 04:46   #65
geoff
 
geoff's Avatar
 
Mar 2003
New Zealand

13×89 Posts
Default

Quote:
Originally Posted by Rincewind View Post
Can someone explain me why we use the -Q switch or what this switch makes.
If you are sieving sequences k*b^n+/-1 then internally sr2sieve converts each one into a number of subsequences of the form (k*b^d)*(b^Q)^m+/-1, where 0 <= d < Q, m = n\Q, and Q is some value chosen using a method that involves a rough approximation of the work needed for each different value of Q considered.

(You can get a list of the values considered and the estimated work by using the -vv switch.)

Normally sr2sieve chooses a good value for Q automatically, but sometimes (usually when there are a very small or very large number of sequences being sieved together) it gets it wrong, and so the -Q switch lets you set the value of Q manually in those cases.

It happens that, for the last remaining sequence this project is sieving, the optimal value for Q is 720 but sr2sieve chooses something else (180 I think). -Q720 just sets the optimal value manually.

Last fiddled with by geoff on 2010-02-10 at 04:47 Reason: spelling
geoff is offline   Reply With Quote
Old 2010-02-10, 05:12   #66
geoff
 
geoff's Avatar
 
Mar 2003
New Zealand

13×89 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
Are you guys still using sr2sieve for this effort? Just checking because sr1sieve will be far faster for the one k that is remaining.
As philmoore said, sr1sieve doesn't yet work with the dual sequences b^n+/-k, and I don't think I will have the time to add that capability anytime soon.

But fortunately, the latest sr2sieve version 1.8.11 included some changes that make sieving a single sequence much faster than earlier versions, and so for most seqences the difference in speed between sr1sieve and sr2sieve 1.8.11 is not anywhere near as great as it used to be.
geoff is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
S9 and general sieving discussion Lennart Conjectures 'R Us 31 2014-09-14 15:14
Sieving discussion thread jasong Twin Prime Search 311 2010-10-22 18:41
Combined sieving discussion ltd Prime Sierpinski Project 76 2008-07-25 11:44
Sieving Discussion ltd Prime Sierpinski Project 26 2005-11-01 07:45
Sieving Discussion R.D. Silverman Factoring 7 2005-09-30 12:57

All times are UTC. The time now is 07:02.

Tue Sep 22 07:02:31 UTC 2020 up 12 days, 4:13, 0 users, load averages: 1.80, 1.59, 1.55

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.