Who are Marin's Mersenne-aries?

Xyzzy · 2003-09-27, 00:58

I hope this is a good idea!

Basically, Marin's Mersenne-aries are people who work for hire, except we don't get paid...

I envision this project as a place where specialized work can be requested to be done...

For example, I like to do P-1 tests... There is really no credit to be had for these, so I may as well help other people...

The idea here is to maximise the output that GIMPS has as a whole... I know people who skip P-1 testing entirely, either because they don't have enough memory or because they run a client, like Mlucas, that doesn't P-1 test...

Later on we might start doing specialized trial factoring too...

Membership in this elite group is open to anyone who cares to join... You don't even have to have a particularly fast computer, just the willingness to put requested work at the top of your worktodo.ini file... That said, this project is perfect for those of you with gobs of memory...

If, by chance, this is a dumb idea, I know you all will let me know! Please be gentle!

Comments are always welcome...

dave_0273 · 2004-03-04, 06:56

In the absence of GP2, I am going to attempt to take over where he left off.

The idea is pretty simple. A small group of people go just ahead of the leading edge of double checking, trying to clear out as many P-1 tests as possible so that when double checks are assigned, most of them are already P-1 tested.

I will post lists of exponents to be P-1 tested just ahead of the leading edge of double checking. This will hopefully accomplish 2 things.

1) Save the computers that are doing double checks from having to do P-1 tests. This way they can start straight on the LL test.
2) We will hopefully find a fair few factors, which will save the double, and in some cases a triple check if the first LL test was in-correct.

Entry into this group is open to anyone who wants to join. However, it would be best that those that did join had a decent amount of RAM allocated so that decent bounds in stage 2 can be run. To join, just post in a thread to claim the exponents and add the list of exponents to the top of your worktodo.ini file. When finished, report your results through the primenet manual test pages and also post your results in the original thread.

Happy P-1 testing....

garo · 2004-03-04, 21:34

Hi dave,
While GP2 is AWOL, thanks for stepping in. I have some experience handling the Primenet server data files and some scripts lying around. Nothing as cool and comprehensive as GP2 has but if you need some help just holler.

dave_0273 · 2004-03-27, 08:42

Thanks for you support garo. It will be interesting to see how this all goes.

I was wondering if you had that little program that GP2(??) wrote that checked that all the factors that have been found so far are correct??? I would just hate for a false factor to be submitted and miss out on a prime.

Xyzzy · 2004-03-27, 12:16

Checking all the factors is easy... You can do it with Excel (the hard way) or with a special Linux program...

http://www.mersenneforum.org/showthr...&threadid=1346

nfortino · 2004-03-27, 14:21

Quote:

Originally Posted by dave_0273

Thanks for you support garo. It will be interesting to see how this all goes.

I was wondering if you had that little program that GP2(??) wrote that checked that all the factors that have been found so far are correct??? I would just hate for a false factor to be submitted and miss out on a prime.

Every time the "weekly" status files come out, I check all the factors. Also, every time a factor is submitted, primenet checks it. I have found a false factor in the data files.

dave_0273 · 2004-03-27, 16:37

Quote:

Originally Posted by nfortino

I have found a false factor in the data files.

Out of interest what was the exponent and the incorrect factor??

nfortino · 2004-03-27, 16:42

Quote:

Originally Posted by dave_0273

Out of interest what was the exponent and the incorrect factor??

Wow! That was an absolutely horrendous typo on my part. That should have read: I have never found a false factor in the data files. Sorry about that.

dave_0273 · 2004-03-27, 16:47

Ah, that is good to hear. Well as long as someone is checking the factor file, I am not going to worry about it. More cycles for p-1 testing...

jinydu · 2004-03-28, 02:00

Quote:

Originally Posted by dave_0273

In the absence of GP2,

Just wondering, what did happen to GP2?

dave_0273 · 2004-04-22, 15:11

Quote:

Originally Posted by jinydu

Just wondering, what did happen to GP2?

I was hoping that someone would answer this so that I would know as well, but after 3+ weeks, I don't think that anyone knows or is going to respond. jinydu, I hope you don't think that I was avoiding your question, I just hoped that someone had more information than what I had and would be able to fill us all in.

My wild stab in the dark, he left as the forum/datamining was just taking up too much time. While he was around, there was always about 2/3 blocks of p-1 work to do. However, just before he left he put a whole lot of blocks of work (i think about 6-8 if I remember correctly) stickied in the mersenne-aries forum. This makes me think that it was a concious decision on his behalf to leave.

I was talking to Xyzzy earlier about it and he has emailed GP2 a couple of times with no reply. I have done the same, and well as private messages through the forum and haven't gotten any reply.

Although GP2 isn't going to see this, I would like to wish him well in whatever he is doing at the moment. I think that everyone appreciated the amount of work that he did. The mersenne-aries and data forums just aren't the same without him.

GP2, we miss you....

/end of sentimental moment - back to crunching

2003-09-27, 00:58	#1
Xyzzy Aug 2002 3²·23·41 Posts	Who are Marin's Mersenne-aries? I hope this is a good idea! Basically, Marin's Mersenne-aries are people who work for hire, except we don't get paid... I envision this project as a place where specialized work can be requested to be done... For example, I like to do P-1 tests... There is really no credit to be had for these, so I may as well help other people... The idea here is to maximise the output that GIMPS has as a whole... I know people who skip P-1 testing entirely, either because they don't have enough memory or because they run a client, like Mlucas, that doesn't P-1 test... Later on we might start doing specialized trial factoring too... Membership in this elite group is open to anyone who cares to join... You don't even have to have a particularly fast computer, just the willingness to put requested work at the top of your worktodo.ini file... That said, this project is perfect for those of you with gobs of memory... If, by chance, this is a dumb idea, I know you all will let me know! Please be gentle! Comments are always welcome...

2004-03-04, 06:56	#2
dave_0273 Oct 2003 Australia, Brisbane 2·5·47 Posts	The Vision.... In the absence of GP2, I am going to attempt to take over where he left off. The idea is pretty simple. A small group of people go just ahead of the leading edge of double checking, trying to clear out as many P-1 tests as possible so that when double checks are assigned, most of them are already P-1 tested. I will post lists of exponents to be P-1 tested just ahead of the leading edge of double checking. This will hopefully accomplish 2 things. 1) Save the computers that are doing double checks from having to do P-1 tests. This way they can start straight on the LL test. 2) We will hopefully find a fair few factors, which will save the double, and in some cases a triple check if the first LL test was in-correct. Entry into this group is open to anyone who wants to join. However, it would be best that those that did join had a decent amount of RAM allocated so that decent bounds in stage 2 can be run. To join, just post in a thread to claim the exponents and add the list of exponents to the top of your worktodo.ini file. When finished, report your results through the primenet manual test pages and also post your results in the original thread. Happy P-1 testing....

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
http://mersenne-aries.sili.net (stats site)	James Heinrich	Marin's Mersenne-aries	51	2015-08-25 02:31
What is "Marin's Mersenne-aries"?	James Heinrich	Marin's Mersenne-aries	1	2015-07-22 21:50
mersenne.ca (ex mersenne-aries.sili.net)	LaurV	mersenne.ca	8	2013-11-25 21:01
Marin Mersenne Documentary	M29	Lounge	4	2013-09-11 02:33
James H's http://mersenne-aries.sili.net utility site.	Uncwilly	Lounge	5	2012-08-11 01:20

2004-03-04, 21:34	#3
garo Aug 2002 Termonfeckin, IE AD0₁₆ Posts	Hi dave, While GP2 is AWOL, thanks for stepping in. I have some experience handling the Primenet server data files and some scripts lying around. Nothing as cool and comprehensive as GP2 has but if you need some help just holler.

2004-03-27, 08:42	#4
dave_0273 Oct 2003 Australia, Brisbane 2·5·47 Posts	Thanks for you support garo. It will be interesting to see how this all goes. I was wondering if you had that little program that GP2(??) wrote that checked that all the factors that have been found so far are correct??? I would just hate for a false factor to be submitted and miss out on a prime.

2004-03-27, 12:16	#5
Xyzzy Aug 2002 3²×23×41 Posts	Checking all the factors is easy... You can do it with Excel (the hard way) or with a special Linux program... http://www.mersenneforum.org/showthr...&threadid=1346

2004-03-27, 16:47	#9
dave_0273 Oct 2003 Australia, Brisbane 470₁₀ Posts	Ah, that is good to hear. Well as long as someone is checking the factor file, I am not going to worry about it. More cycles for p-1 testing...