Sieve discussion Meaning of first/second pass, combined

Citrix · 2005-12-27, 19:28

Lars,

For ranges larger than 15T, do we have to use the combined dat? Also what does all the stuff in the red mean? Do we have to tell the people at SOB that we are done till 15T etc.

Could you get this sorted out.

Thanks,
Citrix

ltd · 2005-12-27, 19:35

Hi Citrix,

one quote from the starting post:

Quote:

All unfinished ranges lower then 85000 available as secondpass test NOT combined

This means they run with the normal dat file.

The 15T line only means that the dat file is now small enough to allow for combined ranges to makes sence but the higher we push that lower limit the better.

The other red lines mark gaps in the sieving that should be filled with the normal dat file.
Only the range form 101300 to 125000 should be run with the combined file or you have to reserve a range at the yahoo for yourself.

The SOB people have enough work for another two months at least before they come back to the lower ranges. So there is no need yet to contact them to join in.

Lars

hhh · 2005-12-27, 19:39

I think I can answer. Sob is lot further than PSP, even with double check. So, until 50G, all is done for Sob, and more. If you want to do joint sieve, you need to do the ranges explicitely assigned to this. Because everything is a bit messed up, there are all these red lines.

Hope this helps. H.

EDIT: I see, Lars know 10 finger writing. He was some seconds faster.

Citrix · 2005-12-27, 20:08

So what work wil be the most productive at this point in the project? IN what range should most users pick up work?

Citrix

jaat · 2005-12-27, 20:27

Also, how do I submit factors for the combined ranges? Send the whole fact.txt to old ldausch email address or do I need to "separate" out SoB and PSP factors? (if yes, then how?)

I think it would be nice to come with a simple solution for submitting factors as it would attract SoB sievers whenever their firstpass finishes(another couple of months I guess). I remember attracting them was the main motivation here for merging and "repeating" a lot of work.

Last but not the least

101300-101400 reserved by jaat (combined)

ltd · 2005-12-27, 21:30

When you reserve a range from 101000 to 125000 for combined sieving you send the combined factors to my mail address as before. There is no work needed from your side. The factors for PSP will be counted to your stats here as normal. The factors found for SOB will go to a special SOB account. The name of that account will be PSP_Group. If you want to get points for the SOB factors on your own SOB account please send me a mail and i will describe in detail what you have to do. In short you will have to make reservations here(for PSP) and at the yahoo forum (for SOB) and then return the results to both sites.

There is no plan yet how we should make the reservations for the SOB people but i think i could be a way to make it like we do it now for their ranges.

@Citrix:
There is no sieving type more important then the other. Both are important for the project.

Lars

Citrix · 2005-12-28, 01:50

I made some changes, I think we can remove all the red lines now, this should make it less confusing. Let me know if I made a mistake.

So undone ranges under 101000 have to be done by PSP alone (ie. not combined) and after that combined. So are all SOB sievers using a combined file to sieve beyond 101000?

Citrix

Mystwalker · 2005-12-28, 23:56

I think "firstpass test" and "secondpass test" should be explained.
If I'm not mistaken, it's 1006-50M for the first and 20M-50M (resp. 1006-50M as well) for the latter, right?

ltd · 2005-12-29, 08:07

Quote:

Originally Posted by Mystwalker

I think "firstpass test" and "secondpass test" should be explained.
If I'm not mistaken, it's 1006-50M for the first and 20M-50M (resp. 1006-50M as well) for the latter, right?

you are right that in second pass the most results are expectet in 20M-50M but both use the 1006-50M dat file.

NeoGen · 2005-12-29, 23:10

My idea could be wrong, correct me please, but isn't it like this?

Firstpassing is when the program sieved the range one first time and picked up a bundle of factors and stuff to build the initial dat files.
And secondpass is the same thing again, but this time just to check if anything was missed before due to software/human error.
Lars said in this thread:
http://www.amdusers.com/modules.php?...er=asc&start=0
that second pass sieving uncovered several new factors not found on the first pass, so that's why I have this idea of first pass and second pass being the same, except the second being "just to make sure".

ltd · 2005-12-30, 10:12

Second pass is "just to make sure" for n=1006-20M. And at the same time first pass for n=20M-50M.

2005-12-27, 19:28	#1
Citrix Jun 2003 7×233 Posts	Sieve discussion Meaning of first/second pass, combined Lars, For ranges larger than 15T, do we have to use the combined dat? Also what does all the stuff in the red mean? Do we have to tell the people at SOB that we are done till 15T etc. Could you get this sorted out. Thanks, Citrix

2005-12-27, 19:39	#3
hhh Jun 2005 565₈ Posts	I think I can answer. Sob is lot further than PSP, even with double check. So, until 50G, all is done for Sob, and more. If you want to do joint sieve, you need to do the ranges explicitely assigned to this. Because everything is a bit messed up, there are all these red lines. Hope this helps. H. EDIT: I see, Lars know 10 finger writing. He was some seconds faster. Last fiddled with by hhh on 2005-12-27 at 19:41

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Combined sieving discussion	ltd	Prime Sierpinski Project	76	2008-07-25 11:44
Sieve Discussion	em99010pepe	Prime Cullen Prime	61	2007-06-07 22:23
Combined Sieve Guide Discussion	Joe O	Prime Sierpinski Project	35	2006-09-01 13:44
Combined Sieving?	jaat	Sierpinski/Riesel Base 5	5	2006-04-18 02:26
New Sieve Thread Discussion	Citrix	Prime Sierpinski Project	15	2005-08-29 13:56

2005-12-27, 20:08	#4
Citrix Jun 2003 7·233 Posts	So what work wil be the most productive at this point in the project? IN what range should most users pick up work? Citrix

2005-12-27, 20:27	#5
jaat Jul 2005 2·3² Posts	Also, how do I submit factors for the combined ranges? Send the whole fact.txt to old ldausch email address or do I need to "separate" out SoB and PSP factors? (if yes, then how?) I think it would be nice to come with a simple solution for submitting factors as it would attract SoB sievers whenever their firstpass finishes(another couple of months I guess). I remember attracting them was the main motivation here for merging and "repeating" a lot of work. Last but not the least 101300-101400 reserved by jaat (combined)

2005-12-27, 21:30	#6
ltd Apr 2003 2²·193 Posts	When you reserve a range from 101000 to 125000 for combined sieving you send the combined factors to my mail address as before. There is no work needed from your side. The factors for PSP will be counted to your stats here as normal. The factors found for SOB will go to a special SOB account. The name of that account will be PSP_Group. If you want to get points for the SOB factors on your own SOB account please send me a mail and i will describe in detail what you have to do. In short you will have to make reservations here(for PSP) and at the yahoo forum (for SOB) and then return the results to both sites. There is no plan yet how we should make the reservations for the SOB people but i think i could be a way to make it like we do it now for their ranges. @Citrix: There is no sieving type more important then the other. Both are important for the project. Lars

2005-12-28, 01:50	#7
Citrix Jun 2003 11001011111₂ Posts	I made some changes, I think we can remove all the red lines now, this should make it less confusing. Let me know if I made a mistake. So undone ranges under 101000 have to be done by PSP alone (ie. not combined) and after that combined. So are all SOB sievers using a combined file to sieve beyond 101000? Citrix

2005-12-28, 23:56	#8
Mystwalker Jul 2004 Potsdam, Germany 831₁₀ Posts	I think "firstpass test" and "secondpass test" should be explained. If I'm not mistaken, it's 1006-50M for the first and 20M-50M (resp. 1006-50M as well) for the latter, right?

2005-12-29, 23:10	#10
NeoGen Dec 2005 2×3³ Posts	My idea could be wrong, correct me please, but isn't it like this? Firstpassing is when the program sieved the range one first time and picked up a bundle of factors and stuff to build the initial dat files. And secondpass is the same thing again, but this time just to check if anything was missed before due to software/human error. Lars said in this thread: http://www.amdusers.com/modules.php?...er=asc&start=0 that second pass sieving uncovered several new factors not found on the first pass, so that's why I have this idea of first pass and second pass being the same, except the second being "just to make sure".

2005-12-30, 10:12	#11
ltd Apr 2003 304₁₆ Posts	Second pass is "just to make sure" for n=1006-20M. And at the same time first pass for n=20M-50M.