Alternative Sieving for 10M digit prime search - Page 3

japelprime · 2007-01-31, 01:14

Thanks VB for good input here.

I sieved my range (+1 to 10T and -1 to 12T ) and have stopped there at the moment.

I have split up the first 100T to 2 x10 files. I am trying to have some page running on old home made server where people can reserve range. I hope I can make it running soon.
As you pointed out (VB) this is not the best choice. At the moment I am just using very old PC´s with limited amount of memory for doing many k/n numbers. But for me this is an opportunity to start somewhere on 10M range and see if there is some interest in this. If some interest is doing the LLR test then it is fine by me. I will not do it on my own. If there will be great interest in some other and bigger10M digit range I would be glad to be a part of it. I don’t have much knowledge about primes....but have some passion for this.....but not much time. So any good idea is most welcome.

Citrix · 2007-01-31, 07:27

Quote:

Originally Posted by CedricVonck

Always interested to help out

If you give me the right instructions

I will stick to the 1-2M range since there is no way that only 2 users together can find a 10M+ prime. Thanks for the reply.

VBCurtis · 2007-02-01, 07:22

Quote:

Originally Posted by japelprime

Thanks VB for good input here.
I sieved my range (+1 to 10T and -1 to 12T ) and have stopped there at the moment.

I find it entertaining to calculate optimal sieve depth for this project. For the k=7, -1 sieve at 12T, can you send me your CPU type/speed in mhz, and how many billion p it completes in 24 hours? I also need the number of candidates left at 12T. I'll estimate how long it would take to LLR a number, and solve for how many T of sieving would be necessary to eliminate one candidate in the time an LLR test would take. I'm sure the number is jaw-droppingly high, which is the fun of the calculation.
-Curtis

japelprime · 2007-02-01, 22:44

I used PIII- 450Mhz and and PIII-750 for this. I don´t remember wich one was doing +1 or -1. This is not the PC’s. I will be using for sieveing. I will probably start with 2x AMD. 600 Mhz Athlon (slotA) and 550Mhz K5-2. I can give you how many billion p it completes later when I start again. If I did some LLR tests on this today I will use 3Ghz P4. This is only what I will use to day.

k*b^n+1
Sieved up to 12*10^12.
Remaining candidates: 1834

k*b^n-1
Sieved up to 10*10^12.
Remaining candidates: 798

Chaos · 2007-02-16, 01:26

Hi Citrix,

I just downloaded the srsieve software and tried to sieve for 3^16. It seems very fast. If you don't mind I would like to help take this k higher. I suggest someone should start a project. Citrix, don't worry about there being only 2 users right now, if the project is worth while, more users will join in.

Let me know what I can do to help?

Citrix · 2007-02-18, 02:31

I can email you the sieve file, if you PM me your email. It is about 2 MB zipped.

I have sieved 1-2M upto 200G and 1M to 50M upto 40G.
We will need to sieve more first, before we PRP. The sieve is very fast so should not take more than a week to reach 10T for 1-50M on a single computer. (Sieve is about 12 times faster than PSP/SOB sieve combined)

If anyone else is also interested, PM me your email.

tha · 2007-02-19, 10:51

Can someone post a message here explaining exactly what you guys are doing in terms that can be understood by a novice? Like what different method and software is used, what kind of results are obtained and how they are merged with the GIMPS project?

Citrix · 2007-02-19, 23:18

We are looking for primes of the form 43046721*2^n+1. This has nothing to do with GIMPS (other than we will be using the same software library as GIMPS).

Uncwilly · 2007-02-20, 04:11

Quote:

Originally Posted by Citrix

We are looking for primes of the form 43046721*2^n+1. This has nothing to do with GIMPS (other than we will be using the same software library as GIMPS).

Why is in the LMH forum then, maybe there is a better place.

Citrix · 2007-02-20, 04:48

I suggest moving this to the lounge or math sub forum. As for the 'why' ask jasong, who started the thread.

ValerieVonck · 2007-02-24, 15:50

Quote:

Originally Posted by Citrix

I suggest moving this to the lounge or math sub forum. As for the 'why' ask jasong, who started the thread.

Citrix,

I crunched 50% of the range,
Currently @ 1053957739937
Found 77000+ factors

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Search of all even-15-digit Aliquot cycles	Drdmitry	Aliquot Sequences	25	2016-12-16 15:26
Polynomial search for 204-digit cofactor of M1009	fivemack	Factoring	45	2012-02-14 08:50
Deep Sieving 10m Digit Candidates	lavalamp	Open Projects	53	2008-12-01 03:59
Help Sieving 10 Million Digit Candidates	lavalamp	Riesel Prime Search	26	2008-05-25 08:24
idea about 10 million digit search(possibly dumb)	jasong	Math	5	2006-06-07 10:39

2007-01-31, 01:14	#23
japelprime "Erling B." Dec 2005 3·5·7 Posts	Thanks VB for good input here. I sieved my range (+1 to 10T and -1 to 12T ) and have stopped there at the moment. I have split up the first 100T to 2 x10 files. I am trying to have some page running on old home made server where people can reserve range. I hope I can make it running soon. As you pointed out (VB) this is not the best choice. At the moment I am just using very old PC´s with limited amount of memory for doing many k/n numbers. But for me this is an opportunity to start somewhere on 10M range and see if there is some interest in this. If some interest is doing the LLR test then it is fine by me. I will not do it on my own. If there will be great interest in some other and bigger10M digit range I would be glad to be a part of it. I don’t have much knowledge about primes....but have some passion for this.....but not much time. So any good idea is most welcome.

2007-02-01, 22:44	#26
japelprime "Erling B." Dec 2005 3×5×7 Posts	I used PIII- 450Mhz and and PIII-750 for this. I don´t remember wich one was doing +1 or -1. This is not the PC’s. I will be using for sieveing. I will probably start with 2x AMD. 600 Mhz Athlon (slotA) and 550Mhz K5-2. I can give you how many billion p it completes later when I start again. If I did some LLR tests on this today I will use 3Ghz P4. This is only what I will use to day. kb^n+1 Sieved up to 1210^12. Remaining candidates: 1834 kb^n-1 Sieved up to 1010^12. Remaining candidates: 798

2007-02-16, 01:26	#27
Chaos Sep 2005 5 Posts	Hi Citrix, I just downloaded the srsieve software and tried to sieve for 3^16. It seems very fast. If you don't mind I would like to help take this k higher. I suggest someone should start a project. Citrix, don't worry about there being only 2 users right now, if the project is worth while, more users will join in. Let me know what I can do to help?

2007-02-18, 02:31	#28
Citrix Jun 2003 2²·11·37 Posts	I can email you the sieve file, if you PM me your email. It is about 2 MB zipped. I have sieved 1-2M upto 200G and 1M to 50M upto 40G. We will need to sieve more first, before we PRP. The sieve is very fast so should not take more than a week to reach 10T for 1-50M on a single computer. (Sieve is about 12 times faster than PSP/SOB sieve combined) If anyone else is also interested, PM me your email.

2007-02-19, 10:51	#29
tha Dec 2002 881 Posts	Can someone post a message here explaining exactly what you guys are doing in terms that can be understood by a novice? Like what different method and software is used, what kind of results are obtained and how they are merged with the GIMPS project?

2007-02-19, 23:18	#30
Citrix Jun 2003 2²·11·37 Posts	We are looking for primes of the form 43046721*2^n+1. This has nothing to do with GIMPS (other than we will be using the same software library as GIMPS).

2007-02-20, 04:48	#32
Citrix Jun 2003 2²×11×37 Posts	I suggest moving this to the lounge or math sub forum. As for the 'why' ask jasong, who started the thread.