New milestone for the GIMPS leading edge
 

I saw where PrimeNet was handing out regular LL assignments over 50 million yesterday afternoon. I got assigned an exponent right below 50 million, but others were getting them as high as 50.02 million. :smile::showoff:

I think the P4 single-core system that got the just-below-50M assignment returned an estimated completion of about 70 days or so. IIRC, exponents in this range require the 3072K FFT.

[quote=NBtarheel_33;206371]I saw where PrimeNet was handing out regular LL assignments over 50 million yesterday afternoon. I got assigned an exponent right below 50 million, but others were getting them as high as 50.02 million. :smile::showoff:[/quote]My P-1 assignments are all in the mid-51 millions, which fits.

[quote]IIRC, exponents in this range require the 3072K FFT.[/quote]Yes.

TF assignments are getting bigger too
 

TF assignments have crossed the 70 million threshold. I got a TF assignment right around 70.5M this evening.

It's not going to be long before we're out of the "classical" GIMPS 79.3M assignment space.

Our only problem is if people start to lose patience with the assignments as they take longer and longer. For most seasoned GIMPSters, this won't be a problem, but it might be for the newer folks. At some point it probably becomes worthwhile to store intermediate save files on the server, and allow folks to complete (and receive credit for) *portions* of an LL test.

41 more to go. This prime carries a special significance for me; I joined GIMPS during its verification.

[QUOTE=jinydu;206444]41 more to go. This prime carries a special significance for me; I joined GIMPS during its verification.[/QUOTE]25 togo now.

[quote=Uncwilly;208101]25 togo now.[/quote]
Halved in 34 days.

[QUOTE=NBtarheel_33;206427]Our only problem is if people start to lose patience with the assignments as they take longer and longer. For most seasoned GIMPSters, this won't be a problem, but it might be for the newer folks.[/QUOTE]

Well in my 7'ish years experience hardware advancements have pretty much kept up with larger exponent ranges.

I was probably (unknowingly) one of the early "fools" to try a 10M digit test in 2003 on, what was at the time fairly leading edge (P400). It took 15 months.

From what has been posted xorbe is the only one so far to admit doing a full 100M digit LL test and also expects to take about 15 months....probably on all 4 cores or a fairly high-end Quad.

[quote=petrw1;208248]Well in my 7'ish years experience hardware advancements have pretty much kept up with larger exponent ranges.

I was probably (unknowingly) one of the early "fools" to try a 10M digit test in 2003 on, what was at the time fairly leading edge (P400). It took 15 months.

From what has been posted xorbe is the only one so far to admit doing a full 100M digit LL test and also expects to take about 15 months....probably on all 4 cores or a fairly high-end Quad.[/quote]

I've got one thread running for 100M. With a Core 2 Quad architecture you have too much loss if you run only one thread.

I'm planing to upgrade to a 750er sooner or later. After that my 100M process can get the full attention ^^ ...

[quote=NBtarheel_33;197055]Countdown to proving M(20996011) is the 40th Mersenne Prime: 100 as of 0538 GMT on 2009-11-26. It took 50 days to clear 100 exponents, or 1.5 times as long as it took to go from 300->200. Looks definitely like exponential decay on the rate of completion, and I would say that we'll be lucky to see this last 100 cleared by April 1, 2010.[/quote]
100 - 50 took 73 days
50 - 25 took 34 days
25 - 12 took 21 days

I'm abandoning the "exponential" model

David

PS The watched pot has not only boiled, but
is nearly boiling dry (like the paint;)

[quote=davieddy;210347]100 - 50 took 73 days
50 - 25 took 34 days
25 - 12 took 21 days

I'm abandoning the "exponential" model
[/quote]

12 - 6 took 9 days

5

20425091 to Jean-Michel Basson
20700073 to RKidson
20725153 to Tom Mueller
20742307 to Jamie Eubank
20756387 to Miles Huffman

So close it hurts!

I don't know how you got that info but I'm curious all the more now to know more about the specific progress on each. I guess you couldn't dig that up, could you, Primeinator?

[QUOTE=Primeinator;211823]20425091 to Jean-Michel Basson
20700073 to RKidson
20725153 to Tom Mueller
20742307 to Jamie Eubank
20756387 to Miles Huffman

So close it hurts![/QUOTE]

Actually:
20425091
Unverified LL 94F0D26AE5B293__ by "Jean-Michel Basson"
Unverified LL 7397B763B7615F__ by "Marc Steinbach" on 2009-10-21

20700073
Unverified LL 2F6F9580025A43__ by "RKidson"
Unverified LL B1747307EA3FA3__ by "Rainer Temme" on 2008-07-04
Assigned Double-checking to "hamlet3k" on 2009-08-17

20725153
Unverified LL 16366F6AF7CABD__ by "Tom Mueller"
Assigned Double-checking to "ANONYMOUS" on [COLOR="DarkOrange"]2009-02-10[/COLOR]

20742307
Unverified LL EBDB123B93B3AD__ by "Jamie Eubank"
Assigned Double-checking to "Thomas Sprenger" on [COLOR="#ff8c00"]2008-11-01[/COLOR]

20756387
Unverified LL 0E5459150324D3__ by "Miles Huffman"
Assigned Double-checking to "SAHS Math Dept" on [COLOR="#ff8c00"]2009-02-11[/COLOR]

[QUOTE=Uncwilly;211832]20742307Assigned Double-checking to "Thomas Sprenger" on [COLOR="#ff8c00"]2008-11-01[/COLOR]
[/QUOTE]

One must suspect this is not running 24/7 :ermm:

[quote=Uncwilly;211832]Actually:
20425091
Unverified LL 94F0D26AE5B293__ by "Jean-Michel Basson"
Unverified LL 7397B763B7615F__ by "Marc Steinbach" on 2009-10-21[/quote]Assigned Double-checking to "ANONYMOUS" on 2009-10-21

20425091 has: Assigned Double-checking to "ANONYMOUS" on 2009-10-21

Both 20425091 and 20700073 has 2 LLs but with different residues.


There should be a line on assigned numbers with last communication date and progress at that date, like you could see in status.txt back in the old server 4.0. I posted it in request thread long ago.

[QUOTE=cheesehead;211872]Assigned Double-checking to "ANONYMOUS" on 2009-10-21[/QUOTE]Yes:redface:

[QUOTE=imwithid;211830]I don't know how you got that info but I'm curious all the more now to know more about the specific progress on each. I guess you couldn't dig that up, could you, Primeinator?[/QUOTE]

No, that information is not available. Furthermore, it appears some of the information I provided was incorrect! Sorry!

"This information and this statistics -- they are slippery things, no?"

[QUOTE=imwithid;211830]I don't know how you got that info but I'm curious all the more now to know more about the specific progress on each. I guess you couldn't dig that up, could you, Primeinator?[/QUOTE]

It was possible to see the status of individual assignments in v4, but this feature was [url=http://www.mersenneforum.org/showpost.php?p=146514&postcount=2]removed in v5[/url] due to being "too slow."

[quote=davieddy;211492]5[/quote]
4

[quote=ixfd64;211996]It was possible to see the status of individual assignments in v4, but this feature was [URL="http://www.mersenneforum.org/showpost.php?p=146514&postcount=2"]removed in v5[/URL] due to being "too slow."[/quote]
Is there any chance of getting this back when we have the new server?

[quote=davieddy;212157]4[/quote]
3

(6 - 3 took 6 days)

[QUOTE=henryzz;212161]Is there any chance of getting this back when we have the new server?[/QUOTE]

It was interesting but there are sooo many more assisgments up to 1B under v5.

Mind you, it would be totally understandable if this report (were it to return) excluded all TF-LMH assignments, which make up the vast majority.

Poaching ?
 

[QUOTE=davieddy;212165]3

(6 - 3 took 6 days)[/QUOTE]When I look at who reported the latest three 20 M double checks, I must conclude that Carsten Kossendey got impatient and decided to poach some exponents :

20725153
Assigned Double-checking to "ANONYMOUS" on 2009-02-10
20725153 Carsten Kossendey mother 16366F6AF7CABD8E 2010-04-17 13:23

20742307
Assigned Double-checking to "Thomas Sprenger" on 2008-11-01
20742307 Carsten Kossendey mother EBDB123B93B3AD65 2010-04-17 15:12

20756387
Assigned Double-checking to "SAHS Math Dept" on 2009-02-11
20756387 Carsten Kossendey mother 0E5459150324D3E2 2010-04-17 18:04



Jacob

[QUOTE=S485122;212180]When I look at who reported the latest three 20 M double checks, I must conclude that Carsten Kossendey got impatient and decided to poach some exponents :

20725153
Assigned Double-checking to "ANONYMOUS" on 2009-02-10
20725153 Carsten Kossendey mother 16366F6AF7CABD8E 2010-04-17 13:23

20742307
Assigned Double-checking to "Thomas Sprenger" on 2008-11-01
20742307 Carsten Kossendey mother EBDB123B93B3AD65 2010-04-17 15:12

20756387
Assigned Double-checking to "SAHS Math Dept" on 2009-02-11
20756387 Carsten Kossendey mother 0E5459150324D3E2 2010-04-17 18:04



Jacob[/QUOTE]

It looks like it... Can you tell if he is about to do the same for the last 3, or is this not possible since they are not actually assigned to him?

[quote=Primeinator;212193]Can you tell if he is about to do the same for the last 3, or is this not possible since they are not actually assigned to him?[/quote]
Except through alternate routes, (e.g. anyone reading this is, or knows, the person doing this) I believe it's not possible.
I'd guess he is. Why poach 3 of the last 5 and leave the rest?

[QUOTE=Mini-Geek;212195]Except through alternate routes, (e.g. anyone reading this is, or knows, the person doing this) I believe it's not possible.
I'd guess he is. Why poach 3 of the last 5 and leave the rest?[/QUOTE]

True, it is unfortunate for the other testers; however, by the age of some of the assignments, it looks like they are either worked on very periodically or essentially not worked on at all.

2

[QUOTE=Mini-Geek;212195]Except through alternate routes, (e.g. anyone reading this is, or knows, the person doing this) I believe it's not possible.
I'd guess he is. Why poach 3 of the last 5 and leave the rest?[/QUOTE]
Carsten is [URL="http://www.mersenneforum.org/member.php?u=8395"]ckdo[/URL] He is a known heavy hitter and has been known to jump in on situations like this.
[code]
Overall stats
Total | Rank Change | GHz-days Ratios by Type, Percent (%)
Rank Member Name GHz-Days |90day 30day 7 day 1 day | TF P-1 LL LL-D ECM ECM-F
126 Carsten Kossendey 6463.486 | 49 11 4 2 | 39.2 27.1 8.5 19.6 1.4 4.1


TF
49 Carsten Kossendey 2532.192 203683 4128

LL
1138 Carsten Kossendey 551.636 12 0

LL-D
92 Carsten Kossendey 1268.015 80 68 | 96 21 5 2

P-1
27 Carsten Kossendey 1754.478 55396 1104

ECM
69 Carsten Kossendey 91.355 1538 0[/code]

I will just leave this here:
[attach]5044[/attach]
That is all.

[QUOTE=Mini-Geek;212195]I'd guess he is. Why poach 3 of the last 5 and leave the rest?[/QUOTE]I think it is because he does a few at a time :[code]20358929 Carsten Kossendey mother 30081562C71B71B2 2010-03-09 10:16
20580433 Carsten Kossendey mother 8165082A0514AD7E 2010-03-09 11:25
20806993 Carsten Kossendey mother 2E18A885965938CB 2010-03-09 12:22
20544763 Carsten Kossendey durandal 47ECD124836D5F56 2010-03-12 03:25
20940839 Carsten Kossendey durandal 0379857238187161 2010-03-12 03:25
20633659 Carsten Kossendey leela 8581430C000417DF 2010-03-14 11:52
20874143 Carsten Kossendey leela 1BE5E9E677AF9F04 2010-03-16 03:41
20490097 Carsten Kossendey leela 436300BC399FD30F 2010-03-16 06:58
20410259 Carsten Kossendey mother 47A26B4F0D0318ED 2010-03-19 17:44
20580953 Carsten Kossendey mother 787F2D99AE5082A8 2010-03-19 17:44
20832587 Carsten Kossendey mother E1F44072F22F7B65 2010-03-19 17:44
20890783 Carsten Kossendey leela E75CD9D6ACE898BE 2010-03-23 12:53
20854711 Carsten Kossendey mother 40D1A53CC17664E7 2010-04-01 00:26
20692627 Carsten Kossendey mother 8436C1FB2EC9F045 2010-04-01 06:44
20979949 Carsten Kossendey mother ED5066798CC3D193 2010-04-01 22:05
20774483 Carsten Kossendey leela 5DFEF974DC2BC37B 2010-04-04 23:45
20427137 Carsten Kossendey mother 96EB90EBD1A03770 2010-04-05 12:12
20584153 Carsten Kossendey mother D8A37BE60EBA12E6 2010-04-05 12:48
20590991 Carsten Kossendey leela 2A3E2E6195480DF0 2010-04-05 13:37
20557643 Carsten Kossendey mother C68546019B50E0B8 2010-04-06 01:21
20869103 Carsten Kossendey leela 29134C181873A7C0 2010-04-11 15:46
20944373 Carsten Kossendey leela D937CFAED7B36949 2010-04-12 13:39
20725153 Carsten Kossendey mother 16366F6AF7CABD8E 2010-04-17 13:23
20742307 Carsten Kossendey mother EBDB123B93B3AD65 2010-04-17 15:12
20756387 Carsten Kossendey mother 0E5459150324D3E2 2010-04-17 18:04[/code]I am afraid the last two will follow very soon.

Jacob

[quote=S485122;212252]I am afraid the last two will follow very soon.

Jacob[/quote]

Indeed.

I feel especially sorry for the 13 year old at SAHS who was
eagerly watching her progress.

FWIW I verified M(17Msomething) on a P2 350 running 24/7
for a few months. Shortly after that the power supply went bang,
taking the motherboard with it.

David

On another note:

Countdown to testing all exponents below M(30402457) once: 3

[quote=10metreh;212274]On another note:

Countdown to testing all exponents below M(30402457) once: 3[/quote]
Shush.
ckdo might hear.

Looks like I need to send ckdo my anti-poaching sermon. 101 mph and all that. (Where'd I put it?)

Do you think he will listen?

Looks like I need to send ckdo my anti-poaching [strike]sermon[/strike] pep talk. 101 mph and all that.

(Where'd I put it?

Seriously, folks: I've just done a forum search but can't find my "101 mph" explanation of why poaching's a bad idea. Anyone know where it was?)

[QUOTE=cheesehead;212338]Seriously, folks: I've just done a forum search but can't find my "101 mph" explanation of why poaching's a bad idea. Anyone know where it was?)[/QUOTE]

This one? [url]http://www.mersenneforum.org/showpost.php?p=149557&postcount=14[/url]

This one? [url]http://www.mersenneforum.org/showthread.php?t=10683&highlight=poaching[/url]

This one? [url]http://www.mersenneforum.org/showthread.php?t=3495[/url]

This one? [url]http://www.mersenneforum.org/showthread.php?t=305[/url]

Thank you for your effort. I should have written that I already searched for posts by myself with "poaching" or "poach", and looked through all those found, without success. Some of the ones you cite refer to my having earlier posted that "101 mph" explanation.

[QUOTE=davieddy;212266]
I feel especially sorry for the 13 year old at SAHS who was
eagerly watching her progress.
David[/QUOTE]

How do you know that?

I assumed that SAHS was a High School, then conjectured.
What age do folk start at high school?

David

[quote=davieddy;212404]What age do folk start at high school?[/quote]In the US, some high schools are grades 9-12, others are 10-12. Often the pairing is with middle schools (7-8) or junior high schools (7-9), respectively.

Most 12th-graders turn 18 during their senior year, so most 13-year-olds would be in 7th or 8th grade.

SAHS does end with 'HS'; but it seems a bit of a stretch to go from there to 'high school'. Still 2 more to go before M20996011.

[quote=jinydu;212518]SAHS does end with 'HS'; but it seems a bit of a stretch to go from there to 'high school'. Still 2 more to go before M20996011.[/quote]
[url]http://en.wikipedia.org/wiki/SAHS[/url]

[quote=henryzz;212549][URL]http://en.wikipedia.org/wiki/SAHS[/URL][/quote]
I also had "Math Dept" as a further clue.

I chose the age and sex for two reasons:

1) Someone you would least wish to discourage by poaching.

2) A jest I shared when meeting a forumite in person for the
first time a few weeks ago. Have to be careful about who you
meet on the internet :smile:

100...
 

There are now 100 remaining LL tests below 38M.

And only 23 less than 33219xxx (10M digits).

[quote=petrw1;212760]There are now 100 remaining LL tests below 38M.

And only 23 less than 33219xxx (10M digits).[/quote]
And boy are these dropping fast (not).

Each decrement is cause for a minor celebration:smile:

23 hasn't changed for ages.

[QUOTE=davieddy;212825]And boy are these dropping fast (not).

Each decrement is cause for a minor celebration:smile:

23 hasn't changed for ages.[/QUOTE]

I wish I could help get them down, but not possible without "poaching".

Again it would be great with a date on active exponents when they last sent progress update to the server. Then we could "poach" those that didn't connect for some number of months we agreed upon.

[QUOTE=ATH;212826]I wish I could help get them down, but not possible without "poaching".[/QUOTE]However since these are all first times, you could start a very early double check.
(ducks the flying cheese.)

[QUOTE=cheesehead;212338]Looks like I need to send ckdo my anti-poaching [strike]sermon[/strike] pep talk. 101 mph and all that.

(Where'd I put it?

Seriously, folks: I've just done a forum search but can't find my "101 mph" explanation of why poaching's a bad idea. Anyone know where it was?)[/QUOTE]Here?

[url]http://www.mersenneforum.org/showpost.php?p=40999&postcount=6[/url]

[quote=retina;212835]Here?

[URL]http://www.mersenneforum.org/showpost.php?p=40999&postcount=6[/URL][/quote]Hmm... That could be a reference to what I'm seeking, but in 2004 maybe it was a precursor instead.

Thanks for your effort, retina, but I'll just write another one.

Room 101 mph or "Ode to George"
 

Before I discovered Smirrnoff (sorry - Mersennefora)
I posted something to the effect that one would accept an error
rate of ~25% before opting for the next FFT size up.

I was flattered that George and Brian Beesleystreet replied
so patiently, often and politely.

The bottom line from George was that he didn't want to explain
to anyone that they had missed a Mersenne Prime through a software error.

David

[QUOTE=davieddy;212217]2[/QUOTE]

It's been 7 days, and neither has finished. Maybe we'll get a bit of a slower tail after all...

[QUOTE=jinydu;213082]It's been 7 days, and neither has finished. Maybe we'll get a bit of a slower tail after all...[/QUOTE]

It takes a while to poach, you know.

[QUOTE=jinydu;213082]It's been 7 days, and neither has finished. Maybe we'll get a bit of a slower tail after all...[/QUOTE]

... and then there was one.

And this too:

Countdown to testing all exponents below M(30402457) once: 2

So M20425091 is the last one standing... It is a triple-check; the double-check was completed on Oct. 21, 2009.

[QUOTE=imwithid;213128]... and then there was one.[/QUOTE]

Was it the same person again?

[quote=davieddy;212825]

Each decrement is cause for a minor celebration:smile:

23 hasn't changed for ages.[/quote]

Let´s celebrate...
It´s down to 22!

[QUOTE=Primeinator;213146]Was it the same person again?[/QUOTE]

[url]http://www.mersenne.org/report_exponent/?exp_lo=20700073&exp_hi=&B1=Get+status[/url]

What do you think?

[QUOTE]Verified LL B1747307EA3FA3AF by "ANONYMOUS" on 2010-04-25[/QUOTE]

Not unless our poacher is refusing to confess...

[quote=lycorn;213167]Let´s celebrate...
It´s down to 22![/quote]

I thought you had learnt your lesson about encouraging me:smile:

[URL]http://www.youtube.com/watch?v=3GwjfUFyY6M[/URL]

[QUOTE=lycorn;213167]Let´s celebrate...
It´s down to 22![/QUOTE]

These are very close too. I propose a celebration for when they conclude as well. Bring on the wine! :beer:

# Countdown to testing all exponents below M(32582657) once: 16
# Countdown to testing all exponents below M(37156667) once: 47

[QUOTE=davieddy;213172]I thought you had learnt your lesson about encouraging me:smile:

[URL]http://www.youtube.com/watch?v=3GwjfUFyY6M[/URL][/QUOTE]

Indeed...
This celebration is more appropriate for the next prime discovery!

It sure is taking a while for the last exponent- it looks like your anti-poaching lecture did not fall on deaf ears.

[quote=Primeinator;214344]It sure is taking a while for the last exponent- it looks like your anti-poaching lecture did not fall on deaf ears.[/quote]

It's called the Quantum Zeno Effect

David

[QUOTE=Primeinator;214344]It sure is taking a while for the last exponent- it looks like your anti-poaching lecture did not fall on deaf ears.[/QUOTE]

Also for the last few weeks NONE of the 22 LL tests below 33219xxx have completed either even though tests between that and 38M complete daily.

Are the majority of these tests just stragglers who have forgotten they are running Prime95 and the exponent just runs an hour a day and thus doesn't expire but takes forever to complete?

PrimeNet has mechanisms to take care of stragglers according to established standards. Let it work.

[quote=cheesehead;214659]PrimeNet has mechanisms to take care of stragglers according to established standards. Let it work.[/quote]
Wouldn't an exponent that's being worked on 1 hour a day (and therefore, expected completion dates updated daily) have its expiry date forever pushed off into la-la land and never be reassigned? (Or am I missing something here?)

[quote=mdettweiler;214701]Wouldn't an exponent that's being worked on 1 hour a day (and therefore, expected completion dates updated daily) have its expiry date forever pushed off into la-la land[/quote]No, not forever. Why exaggerate, if you have a sound argument?

[quote]and never be reassigned?[/quote]No, not never. Why exaggerate?

[quote](Or am I missing something here?)[/quote]a) Do we have actual evidence that an exponent is being worked on for only 1 hour/day?

b) Suppose we do. So, what? What harm, if any, does it do to GIMPS?

As I have explained multiple times, it is better to have a contribution from a "slow" computer while faster systems work on other exponents, than not to have that "slow" contribution. 1001 kph is faster than 1000 kph.

If we were to divert a "fast" system to perform the assignment being worked-on by the "slow" system, that would:

a) prevent the "fast" system from working on some other assignment, [I]thus guaranteeing a delay in a future milestone[/I], and

b) tend to discourage owners of "slow" systems from contributing. (But 1001 kph is faster than 1000 kph.)

So, if you're impatient, just turn your attention elsewhere. That's not only the best way to help GIMPS; it's also handy in other life situations.

Why would anyone want to delay a future milestone for the purpose of needlessly speeding up a current milestone, and perhaps (if the poaching discourages a "slow" system owner from future participation) slowing down overall GIMPS progress? Is it only because of shortsightedness?

GIMPS/PrimeNet has been designed to make efficient use of systems, whatever their speeds. Let it work.

[QUOTE=mdettweiler;214701]Wouldn't an exponent that's being worked on 1 hour a day (and therefore, expected completion dates updated daily) have its expiry date forever pushed off into la-la land and never be reassigned? (Or am I missing something here?)[/QUOTE]You are missing the following provision from the [url=http://www.mersenne.org/thresholds/]Assignment Thresholds[/url] : "The server may reassign exponents when the assignment is more than one year old." (With a little syntax correction on my part.)

What might be necessary is a review the assignment thresholds...

Jacob

[quote=cheesehead;214704]No, not forever. Why exaggerate, if you have a sound argument?

No, not never. Why exaggerate?

a) Do we have actual evidence that an exponent is being worked on for only 1 hour/day?

b) Suppose we do. So, what? What harm, if any, does it do to GIMPS?

As I have explained multiple times, it is better to have a contribution from a "slow" computer while faster systems work on other exponents, than not to have that "slow" contribution. 1001 kph is faster than 1000 kph.

If we were to divert a "fast" system to perform the assignment being worked-on by the "slow" system, that would:

a) prevent the "fast" system from working on some other assignment, [I]thus guaranteeing a delay in a future milestone[/I], and

b) tend to discourage owners of "slow" systems from contributing. (But 1001 kph is faster than 1000 kph.)

So, if you're impatient, just turn your attention elsewhere. That's not only the best way to help GIMPS; it's also handy in other life situations.

Why would anyone want to delay a future milestone for the purpose of needlessly speeding up a current milestone, and perhaps (if the poaching discourages a "slow" system owner from future participation) slowing down overall GIMPS progress? Is it only because of shortsightedness?

GIMPS/PrimeNet has been designed to make efficient use of systems, whatever their speeds. Let it work.[/quote]
I think you misunderstood what I was saying: I wasn't by any means advocating poaching to move ahead milestones, just asking because as I understood it at the time, assignments really would be kept reserved for all eternity as long as the assignee continued to report progress. Jacob's answer has since clarified that.

I thought my "Or am I missing something here" clarified my intended meaning adequately, but it seems not.

[QUOTE=S485122;214706]"The server may reassign exponents when the assignment is more than one year old." Jacob[/QUOTE]

This happened to me personally a few months ago. One of my team PC's which due to a long story, I no longer have access to was used VERY sporadically. Some assignments that due to this low activity did NOT finish in 365 days were dropped by the server.

[QUOTE=mdettweiler;214701]Wouldn't an exponent that's being worked on 1 hour a day (and therefore, expected completion dates updated daily) have its expiry date forever pushed off into la-la land and never be reassigned? (Or am I missing something here?)[/QUOTE]

[quote]No, not forever. Why exaggerate[/quote]
Theoretically, an exponent could be worked on for less than a second every 59 days on a slow computer. That comes out to about 1 iteration every 2 months, and the 60 day update limit will still be satisfied.

It'll take a year just to complete 6 iterations, and more than three million years to finally complete a 6 million digit candidate.

[quote]"The server may reassign exponents when the assignment is more than one year old." [/quote]
The key word is "may", not "will".
edit: This doesn't mean that I support poaching. However, if that one exponent holding up the double checking milestone is somehow still there after a few years, I bet the attitudes towards poaching would change.

I know this message wasn't directed at me, but I'll go ahead and respond.
[QUOTE=cheesehead;214704]
If we were to divert a "fast" system to perform the assignment being worked-on by the "slow" system, that would:

a) prevent the "fast" system from working on some other assignment, [I]thus guaranteeing a delay in a future milestone[/I],
[/quote]
Not necessarily. You're assuming that the fast system is always busy, while this isn't always the case. Let's say that I have a fast system that I usually use to play games. One day, I see the GIMPS milestone page and decide to poach an exponent. Instead of playing a computer game that day, I decide to leave it on until the exponent finishes and do something else less CPU-intensive, like surfing the web.

Now what happens if I didn't poach that exponent? No benefit to GIMPS or any other DC project; I just continue playing games in my spare time instead of doing some web surfing.

[quote]
b) tend to discourage owners of "slow" systems from contributing. (But 1001 kph is faster than 1000 kph.)
[/quote]
Not always. Many of those "slow" systems cannot participate on higher exponents due to factors such as a lack of memory or a lack of stability for multi-year periods. Even worse, they may be unattended and forgotten, and the original searcher of the exponent that got poached won't even know or care what happened.

[quote]
So, if you're impatient, just turn your attention elsewhere. That's not only the best way to help GIMPS; it's also handy in other life situations.
[/quote]
True, but patience has a limit. Waiting a few weeks, months, or even years is fine in certain cases, but you'll have a hard time finding anyone who's willing to wait a couple of centuries. In an extreme case, how would you feel about waiting for the sun to burn up?

[quote=S485122;214706]"The server may reassign exponents when the assignment is more than one year old."[/quote]
I wonder just what the criteria for this is, considering [URL="http://www.mersenne.org/report_exponent/?exp_lo=20742307"]M20742307[/URL]'s (which was poached) proper assignment is now over 18 months old and was never reassigned. In any case, [URL="http://www.mersenne.org/report_exponent/?exp_lo=20425091"]M20425091[/URL], the last one, has only been assigned for just under 7 months, so assuming ANONYMOUS has been checking in updates on the number, I don't see it as unusual that it's not been reassigned.
Have we heard from one of the people in charge (George or Scott(?)) whether the reassignment is working properly, and what more specifically is the requirements for the server to reassign? (since e.g. that 1.5-year-old assignment was not reassigned)

[QUOTE=Mini-Geek;214762]Have we heard from one of the people in charge (George or Scott(?)) whether the reassignment is working properly, and what more specifically is the requirements for the server to reassign? (since e.g. that 1.5-year-old assignment was not reassigned)[/QUOTE]I do not know if the reassignment of work-units older than a year work. I do not remember George or Scott commenting about it. I do remember something about (factoring ?) manual assignments in the higher ranges having a longer expiration time.

Perhaps some of the straggling exponents are at the bottom of long worktdodo.txt files ?

I must say that I find all those assignments of small exponents to some of the Anonymous accounts strange : how can they build up enough reliability and confidence ? Confidence is the number of LL (first time or double-check) tests completed without error. A computer needing more than a year to finish a test will have difficulty accumulating confidence. Once again the Assignment Thresholds must be periodically reviewed. For instance the threshold for unknown computers should be round 26 M, leaving the trailing edge to known computers.

For exponents above 24 930 000 the FFT size is 1536 KiB. The double checks will soon be out of reach for slower computers (200 ms iteration time still implies two month to complete a double check.) There has been so much LMH factoring done lately by quick computers (because of productivity expressed in GHz days) that even that could get out of reach for slower computers.

Jacob

The DC for M20742307 was poached, and then completed on April 17. Don't you mean something else?

[quote=Historian;214737]Theoretically, an exponent could be worked on for less than a second every 59 days on a slow computer. That comes out to about 1 iteration every 2 months, and the 60 day update limit will still be satisfied.

It'll take a year just to complete 6 iterations, and more than three million years to finally complete a 6 million digit candidate.[/quote]In other words,

[I]your argument is so weak that you have to make ridiculous exaggerations to try to justify it.[/I]

Couldn't prove your point with [U]actual[/U] data, could you?

[quote]However, if that one exponent holding up the double checking milestone is somehow still there after a few years,[/quote][U]That has never happened in the history of GIMPS.[/U]

Again: couldn't prove your point with [U]actual[/U] data, could you?

All you're trying to do is justify your barging in on someone else's assignment because you're so egotistical that you think you know better than the GIMPS/PrimeNet administrators, and you just can't stand to follow their rules, like others do.

[quote=Historian;214740]I know this message wasn't directed at me, but I'll go ahead and respond.[/quote]What a parallel (to your proposal that even though you know that a certain PrimeNet assignment isn't assigned to you, you'll go ahead and poach the assignment)!

Consistent, it is.

[quote][quote=cheesehead;214704]If we were to divert a "fast" system to perform the assignment being worked-on by the "slow" system, that would:

a) prevent the "fast" system from working on some other assignment, [I]thus guaranteeing a delay in a future milestone[/I],[/quote]Not necessarily. You're assuming that the fast system is always busy,[/quote]No, your assumption about my assumption is wrong. I was _not_ assuming that the fast system is always busy.

However, I was making a different assumption that I now see isn't necessarily correct, either:

I was assuming that the fast system would be used on a legitimate GIMPS assignment if it were not used to poach.

I neglected to consider that someone like you wouldn't care to contribute to GIMPS if he weren't poaching -- he wouldn't be interested in running legitimate GIMPS assignments.

I suppose I should thank you for cluing me in to your thinking.

[quote]Let's say that I have a fast system that I usually use to play games. One day, I see the GIMPS milestone page and decide to poach an exponent. Instead of playing a computer game that day, I decide to leave it on until the exponent finishes and do something else less CPU-intensive, like surfing the web.

Now what happens if I didn't poach that exponent?[/quote]You could've signed up for a legitimate GIMPS assignment!

No real difference in your system's use from what you propose (running a GIMPS assignment that you poached). Either way, you're accomplishing a GIMPS assignment ...

[I]except[/I]

... except that running a legitimate GIMPS assignment wouldn't have that thrill of poaching, the thrill of knowing that you were demonstrating your superior judgment -- superior to the dum-dum "holding up a milestone" with his oh-so-slow system, the thrill of punishing the guy who had the temerity to "hold up" a milestone you considered urgent.

[quote]No benefit to GIMPS or any other DC project;[/quote]Yes, that's very illuminating:

[U]You'd rather make the negative contribution of the things-wrong about poaching than make a positive contribution.[/U]

If you can't poach at GIMPS, you'd rather do nothing at all at GIMPS.

[B]You've certainly made your motivation clear, "Historian".[/B]

[quote=Historian;214740]
[quote=cheesehead;214704]b) tend to discourage owners of "slow" systems from contributing. (But 1001 kph is faster than 1000 kph.)[/quote]
Not always. Many of those "slow" systems cannot participate on higher exponents due to factors such as a lack of memory or a lack of stability for multi-year periods.[/quote]Yes, always. 1001 kph is faster than 1000 kph -- always.

You're trying to argue that just because a slow system might not be contributing in some ways, that means it can't contribute at all.

Either you're deceptive (deliberately omitting to mention certain ways a slow system can contribute) ... or you're so egotistical you think there can't be any way for a slow system to contribute other than what you've thought of.

[quote]Even worse,[/quote][I]Of course[/I] the "even worse" -- you can't justify your argument without conjuring up all kinds of extreme cases.

Again: couldn't prove your point with [U]actual[/U] data, could you?

[quote]they may be unattended and forgotten, and the original searcher of the exponent that got poached won't even know or care what happened.[/quote]... and then [U]the normal procedures we already have in place will rectify the situation[/U].

[quote][quote=cheesehead;214704]So, if you're impatient, just turn your attention elsewhere. That's not only the best way to help GIMPS; it's also handy in other life situations.[/quote]True, but patience has a limit.[/quote]... and you just can't [I]bear[/I] to refrain from interfering when your own particular standards of patience are exceeded, can you? Instead, you could ask the system administrators to make a judgment call about intervening to override the normal time limits -- but that wouldn't satisfy your ego either, would it?

You [U]must[/U] intervene directly because you [U]know[/U] that your judgment is superior to everyone else's.

[quote]Waiting a few weeks, months, or even years is fine in certain cases,[/quote]... which [U]you[/U] will select based on [U]your[/U] criteria, not the standard GIMPS/PrimeNet criteria

[quote]but you'll have a hard time finding anyone who's willing to wait a couple of centuries.[/quote]... because, once again, you can't make your case using real data or realistic typical situations -- you have to use exaggerations to justify your superior judgment.

(Say, are you the same egotistical guy who was making similar arguments to these, here at mersenneforum.org a few months ago, only under a different nym? I should go back and compare sentence structure, punctuation, word choice, typical mistakes, posting rhythm ... to see how parallel your posts are to his.)

[quote]In an extreme case, how would you feel about waiting for the sun to burn up?[/quote]Exaggeration all the way, isn't it?

You can't justify your proposals on the basis of real data or realistic hypothetical situations, can you?

[QUOTE=jinydu;214782]The DC for M20742307 was poached, and then completed on April 17. Don't you mean something else?[/QUOTE]

Silly me, Mini-Geek already mentioned that 20425091 is the last exponent standing in the same post.

[quote=mdettweiler;214713]I thought my "Or am I missing something here" clarified my intended meaning adequately, but it seems not.[/quote]It should've, but poaching is one of my hot-button topics.

I apologize for my hot-buttoned hot-headedness in your case.

[QUOTE=Historian;214740]True, but patience has a limit. Waiting a few weeks, months, or even years is fine in certain cases, but you'll have a hard time finding anyone who's willing to wait a couple of centuries. In an extreme case, how would you feel about waiting for the sun to burn up?[/QUOTE]

Well, since you're talking extreames. The sun WILL probably burn up before GIMPS is "done" since we think there is an infinite number of larger and larger mersenne primes to find. Yet it somehow doesn't really bother us!

[quote]even though you know that a certain PrimeNet assignment isn't assigned to you, you'll go ahead and poach the assignment)![/quote]
Relax, cheesehead.

If you want full disclosure, I've never poached an exponent before, nor have I attempted to do so. Even if I did have a lot of spare computing power and was annoyed at seeing one exponent hold up a milestone, I'd rather use it for something else instead of using it to compete with others who may be poaching that exact same exponent.

The "pro-poaching" argument was only me playing devil's advocate.

[quote]Say, are you the same egotistical guy who was making similar arguments to these, here at mersenneforum.org a few months ago, only under a different nym[/quote]
I have never said anything about poaching before yesterday. A few months ago, I had no clue this forum even existed, and if any mods want to, they can verify that my IP address and the IP address of that user do not match.

Pardon me if this sounds simplistic but...
 

With V5, every assignment has an ID.

Could not the Server notify Mr. P.O.Acher with: "Sorry this is NOT your assignment. Result rejected.". He could even be notified before he completes it if he is submitting intermediate status updates.

I realize V4 did not have IDs so it may not be as neat and obvious to notify "him" but I think the result should still be rejected.

Other scenarios include:
- Holding the result until the current assignee either completes it (Mr. Acher's result is discarded) or expires (Mr. Acher's result is accepted),
- Or in the case of an LL test Mr. Acher's result can be accepted as a DC if the current assignee completes his LL test.

[quote=Historian;214864]If you want full disclosure, I've never poached an exponent before, nor have I attempted to do so. Even if I did have a lot of spare computing power and was annoyed at seeing one exponent hold up a milestone, I'd rather use it for something else instead of using it to compete with others who may be poaching that exact same exponent.

The "pro-poaching" argument was only me playing devil's advocate.[/quote]Okay. (Damned devilish :)

[quote]I have never said anything about poaching before yesterday. A few months ago, I had no clue this forum even existed, and if any mods want to, they can verify that my IP address and the IP address of that user do not match.[/quote]Okay, I accept that you're not the same guy as before.

Animals
 

[quote=cheesehead;214900]Okay. (Damned devilish :)

Okay, I accept that you're not the same guy as before.[/quote]
I suspect you were thinking of Kevin (no shrinking violet himself).

You:
Please don't let me be misunderstood.
[URL]http://www.youtube.com/watch?v=d2FT4FprxDg[/URL]

Him:
We gotta get out of this place.
[URL]http://www.youtube.com/watch?v=jxNEiZhpinY&feature=related[/URL]

Me:
I'll spend my life in sin and misery in the house of the rising sun.
[URL]http://www.youtube.com/watch?v=mmdPQp6Jcdk&feature=related[/URL]

Lighten up Richard:smile:

David

Well, Kevin stopped running Prime95 due to that argy bargy with cheesehead. So we do not know if cheesehead's strident anti-poachery prevented any contributors from being discouraged but we do know that it drove Kevin away. and he was a very good contributor to GIMPS and TPR so I am pissed off on my team's behalf too.

I'm uncomfortable with the server sometimes deciding to completely reject a result. After all, poaching may be bad for GIMPS, but ignoring accurate data just because it wasn't assigned properly would be worse IMHO. Especially for things besides exponents of high interest. Imagine if you had a load of results (say, TF in a low-priority area) to submit, but hadn't reserved it all in advance. It wouldn't be good if the server just ignored the results.

It should definitely notify people however possible that they are working on (or just completed) something that was assigned to someone else, so they can avoid that in the future (not all poachers are doing it maliciously/intentionally, e.g. [URL="http://www.mersenneforum.org/showthread.php?p=214898#post214898"]this recent event with cheesehead and GrunWalderGIMP/Graff[/URL]).

Another scenario: poachers get no credit. If you have a result for something assigned to someone else, you can submit it and the server will accept it, but you will get no credit for your work, and the rightly-assigned person will still get full credit when they finally submit their work.

[QUOTE=Mini-Geek;214950]Another scenario: poachers get no credit. If you have a result for something assigned to someone else, you can submit it and the server will accept it, but you will get no credit for your work, and the rightly-assigned person will still get full credit when they finally submit their work.[/QUOTE]

I like that idea.

[quote=garo;214935]we do know that it drove Kevin away [/quote]Kevin's incessant poaching drove [I]me[/I] out of signing up for first-time LLs.

- -

[Edit: Also see comments in my next post about evidence inconsistent with Kevin's simply making random mistakes, and showing that he poached others as well as me.]

[quote=Mini-Geek;214950]I'm uncomfortable with the server sometimes deciding to completely reject a result. After all, poaching may be bad for GIMPS, but ignoring accurate data just because it wasn't assigned properly would be worse IMHO. Especially for things besides exponents of high interest. Imagine if you had a load of results (say, TF in a low-priority area) to submit, but hadn't reserved it all in advance. It wouldn't be good if the server just ignored the results.[/quote]I agree with all that.

[quote]It should definitely notify people however possible that they are working on (or just completed) something that was assigned to someone else, so they can avoid that in the future[/quote]I definitely agree with and encourage that.

[quote](not all poachers are doing it maliciously/intentionally, e.g. [URL="http://www.mersenneforum.org/showthread.php?p=214898#post214898"]this recent event with cheesehead and GrunWalderGIMP/Graff[/URL]).[/quote]... and an automatic notification by PrimeNet might help clear up or prevent future misunderstandings.

Right now, PrimeNet comes back to the second reporter (of TF or P-1) with only "result not needed" or something like that ... which doesn't necessarily equate to poaching. (It could be a setup mistake or forgetful self-duplication.)

I'd like to see PrimeNet keep some record of results reported while the exponent was assigned to someone else, and not only alert the person sending that report, but also alert the assignee at his next communication. The details would differ between TF, P-1, and L-L. E.g., poach of a first-time L-L still leaves the assignee's contribution as fully valuable for doublecheck, but poach of a TF level doesn't.

- - -

My incomplete understanding of why my first-time LLs were turned into DCs led me to abandon several of the latter without finishing them. My leading theory, before I understood poaching, was that it was a bug in PrimeNet. Once I understood that Kevin was poaching me, I wanted to communicate with him to straighten it out ... except that I didn't have his e-address. I avoided including his name in my initial complaint posts -- now I see that if I'd done so, someone else might've recognized Kevin's name and provided some way to send a message.

Had messages given me more of an explanation then, I might well have completed the DCs!

I recall that as the instances of my being poached kept accumulating (I'd start new first-time L-L assignments, only to have them poached, in turn, [I]repeatedly and apparently also consistently[/I]), the situation became more and more irritating. It seemed clear, from the timing of events, that Kevin was going to poach every new L-L assignment I took. (I tried waiting a while before requesting new assignments, but when I eventually did resume, I found them also poached within a short time -- faster than I could complete an assignment on my system.)

Note that I realized that the evidence I had did [I]not[/I] show that he was singling me out; he could've simply been poaching [I]anyone[/I]'s first-time LL assignments whose expected completion dates failed to meet his criteria.

However, the evidence I had did seem to indicate that his poaching was not accidental or random. After I learned how to extract sufficient evidence from the available reports, I tracked down some other exponents Kevin had L-Led ... and found that most of their assignees habitually reported more-distant-than-average completion dates.

I think there still could be an innocent explanation, but I can't come up with one that doesn't involve an unbelievably extreme string of coincidences.

[quote=garo;214935]I am pissed[/quote]

Another Americanism.

You may be pissed, but I suspect you meant "pissed off".

Home rool for England.

David