mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > Prime Sierpinski Project

Reply
 
Thread Tools
Old 2007-04-10, 00:07   #1
Gomeler
 
Dec 2006

23 Posts
Default JJSieve

Sorry if I'm asking a question already answered elsewhere

I just put one of my boxes onto SOB:sieve and their download included JJSieve rather than proth_sieve that I have been using in the past for PSP:Sieve. I've noticed that JJSieve is consistently sieving at 635kp/sec compared to 615kp/sec
with proth_sieve, both using the SSE2 versions and using identical SOB.dat files. Is there any reason why I shouldn't use JJSieve? It uses slightly more ram (50mb as opposed to 35mb) but that isn't an issue with 4 gigs

Let me know if there is a reason why we can't use JJSieve, otherwise I'll be swapping over all my rigs over to JJSieve as the sobistrator is a great interface to use, even if just for PSP:Sieve.
Gomeler is offline   Reply With Quote
Old 2007-04-10, 01:48   #2
Citrix
 
Citrix's Avatar
 
Jun 2003

1,579 Posts
Default

Go ahead and use jjsieve.
Citrix is online now   Reply With Quote
Old 2007-04-10, 02:03   #3
Gomeler
 
Dec 2006

23 Posts
Default

Excellent, will be using it from now on, once I finish the few hundred G's I have left that is :)
Gomeler is offline   Reply With Quote
Old 2007-04-10, 23:42   #4
Gomeler
 
Dec 2006

23 Posts
Default

One last question regarding JJSieve/SOB. I'm sieving ranges in the 1212xx G range, should I submit the factors that JJSieve pulls up to PSP:Sieve or are these ranges above/beyond what you guys will be looking at? see a few people have completed ranges in the 1000000 area but everything else is at/below 650000. Makes no sense to have someone else sieve a range that's already been sieved if the results are the same.
Gomeler is offline   Reply With Quote
Old 2007-04-11, 02:31   #5
Joe O
 
Joe O's Avatar
 
Aug 2002

3·52·7 Posts
Default

Quote:
Originally Posted by Gomeler View Post
One last question regarding JJSieve/SOB. I'm sieving ranges in the 1212xx G range, should I submit the factors that JJSieve pulls up to PSP:Sieve or are these ranges above/beyond what you guys will be looking at? see a few people have completed ranges in the 1000000 area but everything else is at/below 650000. Makes no sense to have someone else sieve a range that's already been sieved if the results are the same.
The factors you find are dependent on the dat file you use, not the program.
If you want to benefit both SoB and PSP you would need to use the combined dat file. Now while you could use the combined dat file on the 1212xxx G range, this would cause a lot of extra work for HHH. The easiest thing would be to reserve the ranges here that are next in line in the sieve reservation thread.
Joe O is offline   Reply With Quote
Old 2007-04-11, 19:23   #6
Gomeler
 
Dec 2006

23 Posts
Default

Yup, using the combined DAT file, didn't think it was wise to just sieve for one and not the other. The reason why I ask is that the SOB automated reserveration system is distributing in the 1212xxx range while PSP is in the 500xxx range, I can't sieve SOB in the current PSP region but I'm still finding PSP factors. Should I just place the reservations in the reservation thread and submit the results when PSP gets down to 1212xxx? I'm just trying to milk my clockcycles for all they are worth and would hate to waste clockcycles :)
Gomeler is offline   Reply With Quote
Old 2007-04-12, 02:42   #7
Joe O
 
Joe O's Avatar
 
Aug 2002

3×52×7 Posts
Default

Quote:
Originally Posted by Gomeler View Post
Should I just place the reservations in the reservation thread and submit the results when PSP gets down to 1212xxx? I'm just trying to milk my clockcycles for all they are worth and would hate to waste clockcycles :)
I would urge you to submit all your factors as soon as you find them. You could eliminate an LLR/PRP test or two and save a lot of clockcycles. HHH is going to have to chime in with what he wants to do about the reservations in the high range.

As far as "I can't sieve SOB in the current PSP region", I think that you are labouring under a misunderstanding. If you reserve in the current PSP region you will also be finding factors for SB. Most of these will be for n > 20M and so will not score immediately, but they will be accepted and eventually score.
You may also find factors for n < 20M. Take a look at the attached graph, and also the one in "Pretty Progress Pictures". The light blue points in the dark blue were found by the combined effort. For various reasons, they were missed in the original sieving of these ranges. Quite a few have even eliminated PRP tests for SB.
Attached Thumbnails
Click image for larger version

Name:	GP_RRESULTS2_500T-600T.png
Views:	139
Size:	11.6 KB
ID:	1643  

Last fiddled with by Joe O on 2007-04-12 at 02:45
Joe O is offline   Reply With Quote
Old 2007-04-13, 18:10   #8
Gomeler
 
Dec 2006

23 Posts
Default

Thanks for the reply, I understand that w/ PSP combined sieving I'm finding SOB factors and I've been submitting my results to SOB:S but I needed slightly faster results, making a rush on the DC-Vault :) I'll inform Citrix of the ranges I have/will complete for SoB and submit the factors to PSP also.
Gomeler is offline   Reply With Quote
Old 2007-04-13, 18:26   #9
Citrix
 
Citrix's Avatar
 
Jun 2003

1,579 Posts
Default

It will be better if you inform 'hhh' and 'ltd' of the ranges.
Citrix is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
jjsieve NeoGen Prime Sierpinski Project 8 2006-08-13 01:13
Possible JJsieve bug? Citrix Prime Sierpinski Project 15 2006-08-11 03:27

All times are UTC. The time now is 01:46.

Wed May 19 01:46:58 UTC 2021 up 40 days, 20:27, 0 users, load averages: 2.15, 2.30, 2.22

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, 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.