Project instructions

MooooMoo · 2006-04-25, 06:26

Gribozavr has sieved range 100M-25G, so I've revised this thread.

Instructions:

1.) Download LLR from

http://www.mersenne.org/gimps/

(choose llr370.zip)

2.) The n is fixed at 195000, and the base is 2. Choose any range of k. However, post your chosen range in this thread before starting on it, and choose the lowest range of k available (see "pre sieved range reservation thread). The ranges can be as large or as small as you want.

3.) Download your range (on the pre sieved range reservation thread). Move that file to the folder where LLR is in. Run the program, which will do the rest.

4.) After your range is complete, check LLR's output file to see whether you've found any primes. Post any primes in the "primes found" thread, even if they are not twin primes.
If you need additional help, feel free to post here.

5.) If you find a twin prime, PM me and e-mail Chris Caldwell (He's the creator of the prime pages. See http://primes.utm.edu/ and scroll to the bottom for his e-mail). Do NOT inform anyone else yet.

The reason for this is that I don't want someone else seeing that ?*2^195000-1 and ?*2^195000+1 are twin primes. That person may just submit it by himself to the prime pages, which means that he/she gets the credit. The actual person who found the prime and the project won't get any credit
Good luck!

MooooMoo · 2006-04-25, 06:46

Oops, I forgot to mention that you should make LLR's output file different from its input file. Also, if the output file is larger than 30 bytes, there's a good chance that you found a prime or two. However, it is possible (and has happened before) for the output file to be smaller than 30 bytes, and still contain at least one prime.

jasong · 2006-05-03, 01:17

100,000 to 100,000,000,000 reserved

I'm just sieving and I'm not sure how fast the twin option will sieve them out. I may change my reservation in a couple hours.

Sorry if I'm annoying anyone.

MooooMoo · 2006-05-03, 03:12

Quote:

Originally Posted by jasong

100,000 to 100,000,000,000 reserved

I'm just sieving and I'm not sure how fast the twin option will sieve them out. I may change my reservation in a couple hours.

Sorry if I'm annoying anyone.

Unfortunately, only ranges above 100,000,000 are available (for LLR testing, see range reservations thread). Gribozavr has sieved everything from 100,000,000 until 5,000,000,000, but you're welcome to sieve after 5 billion or LLR test after 100 million. If you choose the sieving option, PM me, and sieve only to 20 billion (there's a really high chance we'll find a twin prime before 20 billion).

It's not entirely your fault for choosing that range, though. My first post in this thread said "currently, all ranges above 100,000 are available". I wanted to edit this many times before, but the site doesn't allow me to edit after 15 minutes. Therefore, you'll have to check the "range reservations thread" for updates.

edorajh · 2006-05-03, 09:20

Quote:

I wanted to edit this many times before, but the site doesn't allow me to edit after 15 minutes.

MooooMoo, I think you should ask Xyzzy to give you Moderator status for this subforum. That way you will be able to change your posts after 15 minutes. After all, you started this project and as project admin you deserve Moderator status.

Edo

biwema · 2006-05-03, 10:44

Sieving does not need primetest reserve.

If we sieve a large range, I recommen also to include the low n part (for example (3 to 20G). The sieving of a 20G range speed is almost the same as 5G.
Without knowing your efforts I tried also a bit sieving on 3 - 10G up to 10T with a speed of 80M/s or 6.5T/day (P4 3.4GHz).
At 10T, there are ~4.6M candidates left which is about 85M text file or 16.5MByte Zipped. Removing 3 candidtes /s at 10T.
when sieving 20G reching 50T (1 Week sieving), such a file (twice as big) could be distributed, sieved more and then merged again.

I people run prime tests on a range of 100M or less, it already makes more sense to take the presieved ranges insted of own sieving.

pacionet · 2006-05-03, 14:49

MooMoooo can you store somewhere the primes found ? In the site I trace only the NUMBER OF PRIMES FOUND.

Of course if we find any twins, I publish them.

MooooMoo · 2006-05-03, 23:48

Quote:

Originally Posted by pacionet

MooMoooo can you store somewhere the primes found ? In the site I trace only the NUMBER OF PRIMES FOUND.

Of course if we find any twins, I publish them.

I'll store the primes found in a separate thread.

----------------------------------------------------------
edorajh,

I've requested to be a moderator on the "forum feedback" forum, so all threads will be updated. Hopefully, there won't be any more range confusion.

antiroach · 2006-06-06, 12:28

i downloaded a pre-sieved range and started running it in LLR, however I noticed that it is only testing k*2^n - 1 candidates, it isnt doing the +1. am i misunderstanding the intentions of this project or am i just doing something wrong?

lsoule · 2006-06-06, 13:30

It only needs to check the +1 candidate when it finds a prime in
the -1 candidate. That will miss primes of the form k2^n+1, but is
faster towards the goal of a twin prime (both +1 and -1 prime).

2006-04-25, 06:26	#1
MooooMoo Apprentice Crank Mar 2006 2·227 Posts	Project instructions Gribozavr has sieved range 100M-25G, so I've revised this thread. Instructions: 1.) Download LLR from http://www.mersenne.org/gimps/ (choose llr370.zip) 2.) The n is fixed at 195000, and the base is 2. Choose any range of k. However, post your chosen range in this thread before starting on it, and choose the lowest range of k available (see "pre sieved range reservation thread). The ranges can be as large or as small as you want. 3.) Download your range (on the pre sieved range reservation thread). Move that file to the folder where LLR is in. Run the program, which will do the rest. 4.) After your range is complete, check LLR's output file to see whether you've found any primes. Post any primes in the "primes found" thread, even if they are not twin primes. If you need additional help, feel free to post here. 5.) If you find a twin prime, PM me and e-mail Chris Caldwell (He's the creator of the prime pages. See http://primes.utm.edu/ and scroll to the bottom for his e-mail). Do NOT inform anyone else yet. The reason for this is that I don't want someone else seeing that ?2^195000-1 and ?2^195000+1 are twin primes. That person may just submit it by himself to the prime pages, which means that he/she gets the credit. The actual person who found the prime and the project won't get any credit Good luck! Last fiddled with by MooooMoo on 2006-12-17 at 00:28

2006-04-25, 06:46	#2
MooooMoo Apprentice Crank Mar 2006 2×227 Posts	Oops, I forgot to mention that you should make LLR's output file different from its input file. Also, if the output file is larger than 30 bytes, there's a good chance that you found a prime or two. However, it is possible (and has happened before) for the output file to be smaller than 30 bytes, and still contain at least one prime. Last fiddled with by MooooMoo on 2006-05-14 at 07:13

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Broadwell new instructions	tha	Hardware	6	2014-07-18 00:08
Useless SSE instructions	__HRB__	Programming	41	2012-07-07 17:43
Instructions to manual LLR?	OmbooHankvald	PSearch	3	2005-08-05 20:28
Instructions please?	jasong	Sierpinski/Riesel Base 5	10	2005-03-14 04:03
Instructions for using Mprime under Linux	tha	Software	7	2004-01-12 18:45

2006-05-03, 01:17	#3
jasong "Jason Goatcher" Mar 2005 3×7×167 Posts	100,000 to 100,000,000,000 reserved I'm just sieving and I'm not sure how fast the twin option will sieve them out. I may change my reservation in a couple hours. Sorry if I'm annoying anyone.

2006-05-03, 10:44	#6
biwema Mar 2004 3×127 Posts	Sieving does not need primetest reserve. If we sieve a large range, I recommen also to include the low n part (for example (3 to 20G). The sieving of a 20G range speed is almost the same as 5G. Without knowing your efforts I tried also a bit sieving on 3 - 10G up to 10T with a speed of 80M/s or 6.5T/day (P4 3.4GHz). At 10T, there are ~4.6M candidates left which is about 85M text file or 16.5MByte Zipped. Removing 3 candidtes /s at 10T. when sieving 20G reching 50T (1 Week sieving), such a file (twice as big) could be distributed, sieved more and then merged again. I people run prime tests on a range of 100M or less, it already makes more sense to take the presieved ranges insted of own sieving.

2006-05-03, 14:49	#7
pacionet Oct 2005 Italy 3×113 Posts	MooMoooo can you store somewhere the primes found ? In the site I trace only the NUMBER OF PRIMES FOUND. Of course if we find any twins, I publish them.

2006-06-06, 12:28	#9
antiroach Jun 2003 2²·61 Posts	i downloaded a pre-sieved range and started running it in LLR, however I noticed that it is only testing k*2^n - 1 candidates, it isnt doing the +1. am i misunderstanding the intentions of this project or am i just doing something wrong?

2006-06-06, 13:30	#10
lsoule Nov 2004 California 2³·3·71 Posts	It only needs to check the +1 candidate when it finds a prime in the -1 candidate. That will miss primes of the form k2^n+1, but is faster towards the goal of a twin prime (both +1 and -1 prime).