P-1 Question

[diamonddave]

Originally Posted by Dubslow

My copy has a different title:

Yet the chart is still refers to Stage 2 memory

[axn]

Originally Posted by diamonddave

NONE of these will permit me to get P-1 work! According to the threshold page

You can get P-1 work without asking for it -- if you get a first time LL test which hasn't had P-1 done, then P95 will automatically do it. This sort of bypasses the threshold check.

In fact, that is how these half-assed P-1s get end up done in the first place.

[axn]

Originally Posted by Dubslow

Though, on further thought, an LL test certainly does not need that much memory (around 50x what's necessary) and yet that is nowhere near enough for P-1 (which needs 10-50x more than that). Now I am suddenly confused.

1. the memory setting is for stage 2.
2. the documentation was written much earlier, and has not been kept up-to-date (not a high priority). "For example, my machine is a dual-processor with 512MB of memory." -- really? George uses a 512MB m/c? Poor thing
3. Around 200MB is sufficient for a 50M exponent to do stage 2.

[Dubslow]

Originally Posted by axn

2. the documentation was written much earlier, and has not been kept up-to-date (not a high priority). "For example, my machine is a dual-processor with 512MB of memory." -- really? George uses a 512MB m/c? Poor thing

An excellent reason to update the default. My mom's laptop from 2005 has 512 MB of memory.

[Rodrigo]

Originally Posted by Dubslow

Interesting idea, but that runs right on the "Beggar's can't be choosers" line, in that nobody has ever gotten that email, nobody is expecting that email, and there's an option to sign up for the Mersenne Newsletter. That last bit means: people we need to reach won't sign up for that, which means that might be angry that they got an email even though they didn't sign up. You might risk losing more than 1 participant if we (George) send out an email.

Glad you mentioned that. In general, GIMPS could stand to use more frequent communication "from on high" with its members (as opposed to conversations on this forum). I did check the box for the newsletter when I signed up, and have yet to receive a single issue, let alone any other kind of communication. Had I not come to the Mersenne Forum on my own initiative, to judge from the subsequent lack of contact I'd have concluded long ago that GIMPS is at best a stagnant or perhaps even moribund project.

PR to one's own membership is one way to promote a voluntary project's health and growth. Keeping your members informed decreases (not increases) churn and dropouts, as more casual members are reminded of the project and remain aware that things are happening.

Those of us who participate in the forum might easily lose sight of this, but we are but a fraction (anybody have stats?) of the totality of GIMPS participants, representing a "hard core" of those involved. The way to reach the most members the fastest is to go beyond the forum and utilize that apparently untapped resource in PrimeNet.

Independently of the newsletter, the PrimeNet signup page includes the statement that,

A contact email address is recommended for account change confirmations, password recovery if you forget your login or password, and to reach you if your computer makes a discovery or for other project support needs [emphasis added].

So there is a basis for using this resource to recruit members to do P-1 work specifically, or to propose increasing their memory allocation. To deal with the "surprise" element, the author (presumably George or a ghost writer working with him) could acknowledge that we've been bad about communicating with the members and that we'll do better in future, and assure the reader that this is a "real" circular (perhaps by providing some "inside" details); and then go on to explain the reason for the circular by describing the need re: P-1. Doubters could be reassured by providing the URL (not a link!) to a relevant explanatory page on mersenne.org that would offer further details.

How's that sound?

Rodrigo

[chalsall]

Originally Posted by Rodrigo

Glad you mentioned that. In general, GIMPS could stand to use more frequent communication "from on high" with its members (as opposed to conversations on this forum). I did check the box for the newsletter when I signed up, and have yet to receive a single issue, let alone any other kind of communication. Had I not come to the Mersenne Forum on my own initiative, to judge from the subsequent lack of contact I'd have concluded long ago that GIMPS is at best a stagnant or perhaps even moribund project.

You make many very valuable points. You empirically have experience with volunteer efforts.

I've been with GIMPS for over ten years now, and to the best of my knowledge I have *never* received a newsletter.

It's the whole issue about "Push" vs. "Pull" communications. And I while I agree that one should minimize the "push" (read: emails), if someone opts in and the emails are relevant, it would probably be a positive thing.

I also agree that this Forum is probably not even known about by many GIMPS participants (even though there is a link from the main site -- it's surprising how few drill-down on all the available links of a site); this is unfortunate.

[Xyzzy]

There used to be newsletters:

http://www.garlic.com/~wedgingt/newsletters.html

[Xyzzy]

In our opinion, this statement is not true:

You can increase your chances of finding a Mersenne prime very slightly if you let the program occasionally use more memory.

You don't find Mersenne primes doing P-1 work, and the time spent doing P-1 work could be spent doing more LL work.

[Rodrigo]

Thanks for the kind words, chalsall!

Rodrigo

[ckdo]

Originally Posted by Xyzzy

You don't find Mersenne primes doing P-1 work, and the time spent doing P-1 work could be spent doing more LL work.

Right. You do find non-primes using P-1, however, and faster than by doing an LL test. Hence, given a fixed time frame, you will be able to test more candidates, thus (marginally) increasing your chance to find a Mersenne prime.

[KyleAskine]

Originally Posted by ckdo

Right. You do find non-primes using P-1, however, and faster than by doing an LL test. Hence, given a fixed time frame, you will be able to test more candidates, thus (marginally) increasing your chance to find a Mersenne prime.

It's true. Imagine if GIMPS was a cluster owned by one large person. We do P-1 because it does increase the odds we find a Prime by culling out compound numbers. While you would spend more time doing LL if you never did P-1, you would also spend more time doing worthless LL that could have been avoided.

The statement is true.