Newer milestone thread

[mdettweiler]

Originally Posted by davieddy

Well ***k it. I never wanted to be a mod in the first place,
and I don't think my old keyboard could stand it anyway.

FYI, I'm not a mod in this particular subforum...just in the "Conjectures 'R Us" and "No Prime Left Behind" subforums under Prime Search Projects. Outside of those I'm just another member of the forum. (The green usernames can be a little confusing that way...)

[davieddy]

Originally Posted by mdettweiler

FYI, I'm not a mod in this particular subforum...just in the "Conjectures 'R Us" and "No Prime Left Behind" subforums under Prime Search Projects. Outside of those I'm just another member of the forum. (The green usernames can be a little confusing that way...)

No offence intended, either to you or the mods in general.
If I intended anything meaningful (which some
suspect I usually don't) I would guess it involved the number
of pixels your post (inter alia) used up.

I am aware that this site has many facets, some of which have
never heard of me or my foibles.

David

[Primeinator]

Originally Posted by davieddy

foibles.

David

Nice word.

[Uncwilly]

Originally Posted by davar55

When SUPERGIMPS solves M100 I'm going to celebrate.

How about MM127?

[NBtarheel_33]

What about sending an e-mail like the following? Surely this isn't rude, or pushy, or disrespectful.

Let's have a discussion that acknowledges both sides of the issue, not just a one-poster, one-sided argument featuring page after page of turgid, wordy posts and angry font effects, OK?

My proposal would be that George could send out an e-mail like the following, once the milestone clock ticked down to 1. Let me know what you think, minus the flames, please.

>>>>>>>>>>>>>>>>>>>>>>

Dear (User),

Allow us to take this opportunity to thank you for contributing your spare computer time to the GIMPS project since (date joined). It is because of loyal, enthusiastic contributors such as yourself that the GIMPS project not only continues to create and record mathematical history for future generations, but also continues to attract new users and spark new interest worldwide in number theory and computing.

We are writing today to inform you that one of the Mersenne numbers you are testing as part of the GIMPS project, namely M31494937, has become the last open first-time testing assignment below the Mersenne primes M32582657 and M37156667. This means that once M31494937 has had a completed first LL test, there will be no remaining candidates needing a first-time test below what are possibly the 44th and 45th Mersenne primes, respectively. Since the GIMPS standard requires that all numbers below a putative n-th Mersenne prime receive two LL tests that produce matching LL residues, before formally naming the candidate as "the n-th Mersenne prime", you can see that completing all of the first-time tests below a Mersenne prime candidate is a key milestone on the way to GIMPS being able to prove a given candidate's order on the list of Mersenne primes.

As you might guess, there are many GIMPS participants (perhaps including yourself) who take great satisfaction in seeing the project reach such milestones. To many participants, these milestones are a measure of GIMPS' strength and success as a project, and it is for this reason that some users enjoy seeing these milestones reached as soon as possible. We therefore would like to respectfully ask your assistance, if you would be willing to oblige us, in expediting your LL test of M31494937 by perhaps allowing the machine running the test to remain on for more time each day, or by allowing the test to run on one of your faster machines. If it is possible for you to expedite this special LL test in this fashion, we thank you; if this is not possible right now, we certainly understand and invite you to continue to contribute at whatever pace you are comfortable with. The bottom line is that the GIMPS project will always value every single user and every single contribution.

If you have any questions regarding this request, we invite you to visit www.mersenneforum.org, where you may post your GIMPS-related questions and receive answers from other seasoned GIMPSters. If you are interested in tracking the status of major GIMPS milestones, we invite you to visit http://www.mersenne.org/report_milestones.

Thank you for your time, and as always, thank you for your contributions to GIMPS.

Best Regards,

The Participants of GIMPS

[NBtarheel_33]

Originally Posted by Primeinator

How long before you expect GPUs to "take off" as you put it?

Given that there is now a CUDA program that makes use of the IBDWT, and has reduced the time for a verification of M42643801 to 2-1/2 days or so, I'd say that a large contingent of serious GIMPSters (and maybe even the more casual GIMPSters who simply have nice gaming machines) will be running GIMPS on their GPUs in the very near future, e.g. within the next year or so.

Barring serious errors in large tests undertaken by GPUs, I would not be surprised to see 100M digit tests become commonplace by the end of 2012.

[NBtarheel_33]

Originally Posted by davar55

When SUPERGIMPS solves M100 I'm going to celebrate.

Need a thread for predicting that arrival.

(Must take into account SON-OF-GIMPS ^ n, for n = 1 , k)

Eight-core systems, RAM in the tens of gigabytes available right now (for the right price)...and imagine what we'll have in five, ten, twenty years. Think of where we were re: Mersenne primes 25 years ago. Now think of where we might be in another 25.

Not only that, but think of how many businesses, schools, universities, etc. we could work on recruiting for GIMPS!

[davieddy]

Originally Posted by NBtarheel_33

Eight-core systems, RAM in the tens of gigabytes available right now (for the right price)...and imagine what we'll have in five, ten, twenty years. Think of where we were re: Mersenne primes 25 years ago. Now think of where we might be in another 25.

Not only that, but think of how many businesses, schools, universities, etc. we could work on recruiting for GIMPS!

Even with Moore's Law continuing to hold, we may well not find any more
Mersenne primes before 2025, let alone a 100M digit one.

There is obviously scope for a vast expansion in GIMPS participation,
but I don't detect (on this forum at least) much enthusiasm for the
sexing-up/dumbing-down which would make this likely.

David

I'll expand on this a bit.

In 15 years, GIMPS has raised height of the Mersenne bar
from exponents of 1M to 50M.
The computation expected to be needed per prime found
has gone up by more than 50³ ~ 2¹⁷
The expression "victim of its own success" occurs to me.
.

[Brian-E]

Originally Posted by NBtarheel_33

What about sending an e-mail like the following? Surely this isn't rude, or pushy, or disrespectful.

Let's have a discussion that acknowledges both sides of the issue, not just a one-poster, one-sided argument featuring page after page of turgid, wordy posts and angry font effects, OK?

My proposal would be that George could send out an e-mail like the following, once the milestone clock ticked down to 1. Let me know what you think, minus the flames, please.

>>>>>>>>>>>>>>>>>>>>>>
<snip>

I think however politely and professionally you word an Email - and you have done that well in my opinion - the result is still likely to be one of two possibilities:

(1) The mail is ignored as spam or as uninteresting.

(2) The underlying message - beneath all the politeness and compliments - is deciphered and its meaning is digested by the receiver of the mail.

In this case, the underlying message which I would read is: "we are all waiting for you, we think you are not using all the resources you could be using to get this test finished, could you get a move on please?". My reaction would be one of shame and embarrassment. The mail would not have the desired effect on me. I would respond by giving up the test, apologising, leaving GIMPS and feeling that much less worthy as a human being when I had thought I was contributing to a wonderful team effort.

[davieddy]

Originally Posted by Brian-E

I think however politely and professionally you word an Email - and you have done that well in my opinion - the result is still likely to be one of two possibilities:

(1) The mail is ignored as spam or as uninteresting.

(2) The underlying message - beneath all the politeness and compliments - is deciphered and its meaning is digested by the receiver of the mail.

In this case, the underlying message which I would read is: "we are all waiting for you, we think you are not using all the resources you could be using to get this test finished, could you get a move on please?". My reaction would be one of shame and embarrassment. The mail would not have the desired effect on me. I would respond by giving up the test, apologising, leaving GIMPS and feeling that much less worthy as a human being when I had thought I was contributing to a wonderful team effort.

We've tried the subtle warm and friendly approach to this
nonproblem.

How about a good kick up the arse?

Margaret Thatcher

[chalsall]

Originally Posted by Brian-E

I think however politely and professionally you word an Email - and you have done that well in my opinion - the result is still likely to be one of two possibilities:

(1) The mail is ignored as spam or as uninteresting.

(2) The underlying message - beneath all the politeness and compliments - is deciphered and its meaning is digested by the receiver of the mail.

Why did you not articulate the third possibility: that the email might be interpreted as it is intended?

Quite possibly this contributor doesn't realize he's inadvertently holding up what *some* other contributors consider to be an important milestone. This contributor likely doesn't even know about this forum, and might be interested in learning about it.

And, as has been documented by someone else above, this contributor does have another, much faster, machine which could complete the work if they *choose* to expedite it.

P.S. NBtarheel_33 -- a very well written and diplomatic email!