Sieving

paulunderwood · 2004-08-03, 02:35

Okay we have the new LLRs from Jean and George and we will reach n=2 million all the sooner... but then what? I have started to sieve 2-7million not with aim of completely doing it but with that of have something to LLR after 2 million. Thomas Ritschel took 1.7 Ghz year to sieve 1-2m... so with new LLR's 2-3m would take 6-7 GHz years to sieve. To get 2-7 million to that level is going to take 20+GHz years. I have only 4 GHz of sieving computers at the moment. It would take me 5 years by myself but it should be available with in a year? So is there enough good will out there to DC the sieving? Or should I restrict myself to 2-3m and then take it from there?

ET_ · 2004-08-03, 07:21

Quote:

Originally Posted by paulunderwood

Okay we have the new LLRs from Jean and George and we will reach n=2 million all the sooner... but then what? I have started to sieve 2-7million not with aim of completely doing it but with that of have something to LLR after 2 million. Thomas Ritschel took 1.7 Ghz year to sieve 1-2m... so with new LLR's 2-3m would take 6-7 GHz years to sieve. To get 2-7 million to that level is going to take 20+GHz years. I have only 4 GHz of sieving computers at the moment. It would take me 5 years by myself but it should be available with in a year? So is there enough good will out there to DC the sieving? Or should I restrict myself to 2-3m and then take it from there?

I just bought a new Pentium IV 3.2 GHz

I'm devoted to this project, just choose what work best fits to me

Luigi

paulunderwood · 2004-08-03, 11:17

Your P4 is best suited to LLR work.

I have decided to sieve only 2-3 million. Anybody wanting to help out please email me. You will probably have an Athlon/PIII and be prepared to run it for a few months.

Thomas11 · 2004-08-03, 11:20

Quote:

Originally Posted by paulunderwood

Okay we have the new LLRs from Jean and George and we will reach n=2 million all the sooner... but then what? I have started to sieve 2-7million not with aim of completely doing it but with that of have something to LLR after 2 million. Thomas Ritschel took 1.7 Ghz year to sieve 1-2m... so with new LLR's 2-3m would take 6-7 GHz years to sieve. To get 2-7 million to that level is going to take 20+GHz years. I have only 4 GHz of sieving computers at the moment. It would take me 5 years by myself but it should be available with in a year? So is there enough good will out there to DC the sieving? Or should I restrict myself to 2-3m and then take it from there?

I'm willing to join the sieving and could dedicate a few GHz on it.

Since even with the new (faster) LLR the n=1-2M range wouldn't be completed in relativelly short time, there is plenty of time (well, about one year, or so) to do some extended sieving. Therefore we shouldn't concentrate on 2-3M only but do an even larger range, e.g. the 2-7M you mentioned, Paul. Even in case that we run short of time (or data blocks for LLR), we could split of the 2-3M range from the presieved 2-7M and do the remaining 3-7M later.

But besides all the euphorics about the new LLR we shouldn't forget that the sieving is only about 5% of the whole work. Currently we are only 5 or 6 active members (LLRlers) with 12 or 15 machines. To reach a million digits (not "binary" digits) prime, it will take us a few years. Hopefully the new LLR with it's working stop/restart behaviour will attract some more coworkers.

-- Thomas

Thomas11 · 2004-08-03, 11:24

Quote:

Originally Posted by ET_

I just bought a new Pentium IV 3.2 GHz

I'm devoted to this project, just choose what work best fits to me

Luigi

By the new LLR/P4 your new P4 should be able to do one of the current tests in about 2000 seconds!
So go and grab a new range, Luigi!

paulunderwood · 2004-08-03, 11:33

Well, I have started sieving n=2-3 million with the view of doing any sieving for higher values when we are LLR'ing 2-3m. I have created 20 10trillion blocks which is more than plenty. After the first 38.5 trillion has been done we can incorporate the tail end of n=1-2 million. This will be enough work for a year. 3+million can be started in a year's time -- this will keep us buzy for years...

Thomas email me letting me know how many machines you have free.

Lupine1647 · 2004-08-03, 13:44

I could put my old Celeron 700 MhZ to work on sieving, don't know how much help it would be because it would do it part time on that machine.

paulunderwood · 2004-08-03, 18:38

Probabaly not much help if it's part-time. How many hours per week is it on? Also the the Windoze version of NewPGen has to be modified to make it suitably automatic -- I am waiting for Paul Jobling to "fix" it.

Lupine1647 · 2004-08-03, 22:03

It's on from atleast 8:00 Am - 9:00 PM, It doesn't run during the night.

paulunderwood · 2004-08-03, 22:09

Okay. When Paul Jobling has modified NewPGen I will contact you. Your range will have to be, I suppose, no more that 5 trillion...

Lupine1647 · 2004-08-03, 22:21

k, sounds good.

2004-08-03, 02:35	#1
paulunderwood Sep 2002 Database er0rr 10617₈ Posts	Sieving Okay we have the new LLRs from Jean and George and we will reach n=2 million all the sooner... but then what? I have started to sieve 2-7million not with aim of completely doing it but with that of have something to LLR after 2 million. Thomas Ritschel took 1.7 Ghz year to sieve 1-2m... so with new LLR's 2-3m would take 6-7 GHz years to sieve. To get 2-7 million to that level is going to take 20+GHz years. I have only 4 GHz of sieving computers at the moment. It would take me 5 years by myself but it should be available with in a year? So is there enough good will out there to DC the sieving? Or should I restrict myself to 2-3m and then take it from there?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
NFS sieving?	Dubslow	Factoring	8	2012-09-28 06:47
Line sieving vs. lattice sieving	JHansen	NFSNET Discussion	9	2010-06-09 19:25
10^420 + 1 sieving	juno1369	Factoring	20	2010-04-28 01:11
Sieving	OmbooHankvald	Prime Sierpinski Project	4	2005-06-30 07:51
Sieving	robert44444uk	Sierpinski/Riesel Base 5	8	2005-04-02 22:30

2004-08-03, 11:17	#3
paulunderwood Sep 2002 Database er0rr 5×29×31 Posts	Your P4 is best suited to LLR work. I have decided to sieve only 2-3 million. Anybody wanting to help out please email me. You will probably have an Athlon/PIII and be prepared to run it for a few months.

2004-08-03, 11:33	#6
paulunderwood Sep 2002 Database er0rr 10617₈ Posts	Well, I have started sieving n=2-3 million with the view of doing any sieving for higher values when we are LLR'ing 2-3m. I have created 20 10trillion blocks which is more than plenty. After the first 38.5 trillion has been done we can incorporate the tail end of n=1-2 million. This will be enough work for a year. 3+million can be started in a year's time -- this will keep us buzy for years... Thomas email me letting me know how many machines you have free.

2004-08-03, 13:44	#7
Lupine1647 Jan 2004 Kansas City, Missouri 2×5 Posts	I could put my old Celeron 700 MhZ to work on sieving, don't know how much help it would be because it would do it part time on that machine.

2004-08-03, 18:38	#8
paulunderwood Sep 2002 Database er0rr 4495₁₀ Posts	Probabaly not much help if it's part-time. How many hours per week is it on? Also the the Windoze version of NewPGen has to be modified to make it suitably automatic -- I am waiting for Paul Jobling to "fix" it.

2004-08-03, 22:03	#9
Lupine1647 Jan 2004 Kansas City, Missouri 2·5 Posts	It's on from atleast 8:00 Am - 9:00 PM, It doesn't run during the night.

2004-08-03, 22:09	#10
paulunderwood Sep 2002 Database er0rr 5×29×31 Posts	Okay. When Paul Jobling has modified NewPGen I will contact you. Your range will have to be, I suppose, no more that 5 trillion...

2004-08-03, 22:21	#11
Lupine1647 Jan 2004 Kansas City, Missouri 12₈ Posts	k, sounds good.