mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Marin's Mersenne-aries (https://www.mersenneforum.org/forumdisplay.php?f=30)
-   -   Looking for a volunteer for a dangerous mission... (https://www.mersenneforum.org/showthread.php?t=2375)

ThomRuley 2004-04-21 15:36

Looking for a volunteer for a dangerous mission...
 
I am looking for one volunteer for a TF/P-1 mission well outside PrimeNet territory. I have been working on a LMH range for some time and was wanting to do some full trial-factoring and some P-1 on some of the leftovers. However, my TF computer is busy doing the trial-factoring. I can keep someone supplied with P-1 for quite a while. Only one catch, the person doing this P-1 would also be doing a bit of TF in the process. If anyone is interested, please let me know.

ThomRuley

dave_0273 2004-04-22 04:22

What sort of range of exponents are we looking at?? Is it part of the LMH>79.3M or is it a range below 79.3M and just not on primenet yet?? Also, why are you interested in p-1 factoring it??? (I'm not objecting to it being done, I am just curious)

ThomRuley 2004-04-22 04:53

Excellent questions, dave_0273. First, it is a range under 79.3M but not on PrimeNet (56.5-57M, to be exact). I have been trial-factoring in this range for about six months or so. Not long ago, the idea occured to me to start testing some of these on a P4 I built last month. However, the computer I have been using for factoring is still tied up with that factoring project and the P4 is tied up testing an exponent. I recently noticed the Mersenne-aries group and thought someone might be interested in helping me get a few of these ready for LL-testing. Just an idea for someone who wants to be a Mersenn-ary but wants to avoid competing with other Mersenne-aries for exponents.

dave_0273 2004-04-24 17:09

I have to say that it is something that I was alreadying pondering over for my LMH range (44-45M). Over the past couple of days since you have posted, I have been giving it even more thought. These are the things that have me nervous.

1. Each exponent would first have to be trial-factored up to 2^70. On my computer (2.4Ghz P4), this would take at least 2 days (very rough estimate) as only 2 exponents are at 70 bit, the remaining 12850 exponents in the 56.5 to 57M range are still only at 63 bit.

2. The p-1 test would take at least a day, probably close to 2 days.

So, at a guestimate, it would take about 3 days per exponent.

In a couple of days when one of my computers finishes the batch of exponents that it is currently working on, I am going to do some proper timings to find out how long it will take per exponent.

geoff 2004-04-25 14:37

[QUOTE=dave_0273]1. Each exponent would first have to be trial-factored up to 2^70. On my computer (2.4Ghz P4), this would take at least 2 days (very rough estimate) as only 2 exponents are at 70 bit, the remaining 12850 exponents in the 56.5 to 57M range are still only at 63 bit.

2. The p-1 test would take at least a day, probably close to 2 days.[/QUOTE]

Since you are doing this manually anyway, it might be better to change the order of the work. Maybe TF to 2^68, then stage one P-1, then TF to 2^69, then stage 2 P-1, then TF to 2^70 for example.

dave_0273 2004-04-26 07:05

The only problem with doing it that way is that you would need to keep each of the save files from stage 1 of p-1 factoring. Unless of course you mean doing it in that order, but one exponent at a time. In which case, you may as well just do it in the normal order ie fully trial-factor, then fully p-1 factor.

geoff 2004-04-26 19:40

I was thinking about doing it one exponent at a time, but you could also do all TF to 2^68, then P-1 stage 1 and 2, then TF to 2^70. That would probably work out a bit faster than doing all the TF first because a lot more factors will be found in P-1 than in the last 2 bits of TF.


All times are UTC. The time now is 11:50.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.