Newer milestone thread - Page 39

chalsall · 2012-02-22, 00:16

Quote:

Originally Posted by Dubslow

It's not unilaterally poaching if you have a have-finished expo that was unreserved due to a bug in either P95 or PM. It is poaching, in one sense, but in many others it is not. It is not knowingly and willingly starting an assignment from scratch while also knowing someone else is in progress and might not get credit if you beat them. Therefore I would not call it poaching in the usual sense.

Grasshopper talks about what he doesn't know...

Sigh...

Learn soon, he will...

Dubslow · 2012-02-22, 00:17

Quote:

Originally Posted by chalsall

Grasshopper talks about what he doesn't know...

Sigh...

Learn soon, he will...

???

...It has happened to me, where my AID went bad for no reason...

chalsall · 2012-02-22, 00:27

Quote:

Originally Posted by Dubslow

It has happened to me, where my AID went bad for no reason...

Candidates in question of interest are.

Dubslow · 2012-02-22, 03:08

Quote:

Originally Posted by chalsall

Candidates in question of interest are.

It's like your channeling davieddy. I have no idea what you're trying to say.

kladner · 2012-02-22, 03:11

Quote:

Originally Posted by Dubslow

Quote:
Originally Posted by chalsall http://www.mersenneforum.org/images/...s/viewpost.gif
Candidates in question of interest are.

It's like your channeling davieddy. I have no idea what you're trying to say.

I think he is asking for examples of candidates which were screwed up. He's just using Yoda syntax.

bcp19 · 2012-02-22, 03:31

Quote:

Originally Posted by cheesehead

So, communicate with GIMPS administrators. Wake up the old discussion threads about devising a more automated response to such situations.

Dubslow posted a link to the status numerous exponent assignments from 38103407 to 45000000 he put on his site from Jan 11 and Feb 21. In looking at all of the assignments over 365 days old that showed progress and were in both dates, there were 9 that showed no change, 20 that showed .1 to 1% progress, 22 from 1% to 5% and 9 with 5%+. Given that 41 days had passed and x% had completed, I used a spreadsheet to calculate remaining time on these 51 exponents based on those numbers, 32 of these 51 showed between 1 and 98 years remaining:

Code:

current
% done years  
57.5  1.1934931507
87.1  1.2075342466
71.4  1.7847792998
37.7  2.5918822933
90.6  2.6397260274
84.3  2.9392694064
62.3  3.024853229
53.3  3.4971689498
67.1  3.6956164384
11.5  4.7338551859
12.2  5.1907714492
12.2  5.1907714492
42.3  5.4011415525
42.7  5.8513075965
49.7  6.2779299848
43.9  6.3016438356
48.7  6.402739726
24.4  7.7200498132
23.8  7.7813200498
62.6  8.4021917808
53 8. 799086758
54.1  10.3117808219
44.7  15.5294520548
58.2  15.6511415525
20.9  22.2130136986
27.3  27.2210045662
25    28.0821917808
41.6  32.8
14.4  48.0767123288
1.2   55.4904109589
14    96.602739726
12.2  98.6246575342

cheesehead · 2012-02-22, 08:36

Quote:

Originally Posted by Dubslow

It's not unilaterally poaching if you have a have-finished expo that was unreserved due to a bug in either P95 or PM.

Poaching status has nothing to do with any reason or blame for unreservation. Poaching is knowingly or carelessly starting or resuming work on an exponent that's assigned to someone else but not to you.

Quote:

It is poaching, in one sense, but in many others it is not. It is not knowingly and willingly starting an assignment from scratch while also knowing someone else is in progress and might not get credit if you beat them. Therefore I would not call it poaching in the usual sense.

Yes, advocates and apologists for poaching can find and list numerous excuses.

But none of those excuses gives you a PrimeNet assignment. Without an assignment, it's poaching if someone else _does_ have an assignment for that exponent.

Quote:

Communicating with the current assignee is in general not possible; I am sure that many people join GIMPS because of the minimum of interaction required to participate in the project.

Yes, advocates and apologists for poaching like "minimum of interaction", because the less interaction they have, the less they have to think about the effects of their poaching on the real person to whom the exponent is assigned.

Quote:

Communicating with anyone, George/admin or not, is still a hassle, and falls under the "not-minimal-interaction".

Paying for an item at a store is a hassle. It's much more minimal to just walk out of a store after pocketing some item than it is to go through all the hassle of standing in line at a cashier and actually paying for the item.

That doesn't make shoplifting okay, and the aspects you mention don't justify poaching.

Quote:

It is far easier and simpler just to finish the work that had already been done so that I don't lose two months of work. No one will die.

It is far easier and simpler just to walk out of the store with the item and take it home than it is to stand in line at a cashier and pay for the item. No one will go broke. The insurance company will reimburse the store.

aketilander · 2012-02-22, 11:53

Quote:

Originally Posted by cheesehead

Let's be even more honest and empirical: Is there any ethical thing you could accomplish by unilateral poaching that you couldn't accomplish through communication and cooperation with one of the GIMPS administrators?

Well I will willingly be doing that. I told you my situation I have a few exponents where the original assignments were lost. It was completely due to circumstances which I did not have control over. When it has been possible I have registered similar cases as double-checks. The others have been reassinged a number of times since I had them but I can never catch them on the fly myself when they are available.

The exponents concerned are:

M51641341
M52502687
M53045947
M51394979
M52026859
M52834037

So please help me. Who should I communicate with and what should I ask for??

Maybe chalsall can help me with M51394979 since it is reserved for GPU factoring.

So far, due to this discussion, I have disconnected prime95 from PrimeNet so even if they finish the result will not be sent to the server.

... or should I just drop the exponents. They are on average around 70%.

KyleAskine · 2012-02-22, 11:58

Quote:

Originally Posted by chalsall

I don't disagree.

But, empricially, it seems there are only two trust levels with PrimeNet currently.

"Trusted" machines get "prefered" assignments.

Everyone else (including Anonymous machines) get anything above the prefered which are available; including those candidates at the "wave-back".

IMHO, PrimeNet needs some tender loving attention on its heuristics (and a bit of time on the table's indexes probably wouldn't hurt either)....

And some front end work...

LaurV · 2012-02-22, 12:07

I think both of you are right, and both of you are wrong (edit: chalsall-cheesehead dispute, sorry, did not see the last page of the thread). It is not possible to stop poaching. Poaching is happening since GIPMS exists, everyone did it occasionally, even now is going on, and it is simply to prove: if you add the DC assignments in the 30M range with the cleared expos, you get a surplus of about 200 expos. They are poacher's. If you look to active assignments under 24M (there should be NO LL and NO DC, all the range is cleared, according with the last registered milestone), there are still 20 or 30 active assignments. They were poached because the assignee was too slow, or are poached and the poacher is too slow.

Gimps survived for 10 years and poaching was never a big problem. Not all people understand that they are wasting resources if they poach. The most of us do understand, or do not care (they set and forget p95), but still that, sometime the curiosity beat the crap out of us. And we occasionally poach.

Of course this can give birth to unwanted situations, especially if some poacher finds a prime before the assignee does. These cases should be judged FROM CASE TO CASE. You can not make a general rule. Was the assignee sleeping? Had he billions of assignment or a very slow machine? Is the poacher... poacher? Does he do that everyday, or he did it just first/second/etc time? This is all recorded in the database.

All situations can be judged INDIVIDUALLY. You can not make a general rule, like for example "the assignee always take the money and the glory". This would be stupid and could result in abuses, for example I would reserve few thousands exponents and do just so much work on them to avoid Primenet kicking my ass off. Some "poacher" will get bored, and if he would find a prime I would take the money.

We could never interdict poaching by a rule. What you can do? Ban the poacher? You lose the computing power. Of course, we would never encourage poaching by a given rule. Fortunately most of the people understand that poaching is waste of time on their side. We already discussed this in many threads, many times.

And so on. Good/bad examples can be found on both cases.

ATH · 2012-02-22, 12:38

Here is George view on when assignments are too old, from the previous milestone thread: http://www.mersenneforum.org/showpos...&postcount=443

So if assignment is more than 365 days old and progress<50% or if assignment is more than 548 days old.

Apparently these recycling rules are still not implemented or not working properly, or your example M24077267 would have been reassigned 312 days ago.

2012-02-22, 12:07	#428
LaurV Romulan Interpreter Jun 2011 Thailand 3·3,221 Posts	I think both of you are right, and both of you are wrong (edit: chalsall-cheesehead dispute, sorry, did not see the last page of the thread). It is not possible to stop poaching. Poaching is happening since GIPMS exists, everyone did it occasionally, even now is going on, and it is simply to prove: if you add the DC assignments in the 30M range with the cleared expos, you get a surplus of about 200 expos. They are poacher's. If you look to active assignments under 24M (there should be NO LL and NO DC, all the range is cleared, according with the last registered milestone), there are still 20 or 30 active assignments. They were poached because the assignee was too slow, or are poached and the poacher is too slow. Gimps survived for 10 years and poaching was never a big problem. Not all people understand that they are wasting resources if they poach. The most of us do understand, or do not care (they set and forget p95), but still that, sometime the curiosity beat the crap out of us. And we occasionally poach. Of course this can give birth to unwanted situations, especially if some poacher finds a prime before the assignee does. These cases should be judged FROM CASE TO CASE. You can not make a general rule. Was the assignee sleeping? Had he billions of assignment or a very slow machine? Is the poacher... poacher? Does he do that everyday, or he did it just first/second/etc time? This is all recorded in the database. All situations can be judged INDIVIDUALLY. You can not make a general rule, like for example "the assignee always take the money and the glory". This would be stupid and could result in abuses, for example I would reserve few thousands exponents and do just so much work on them to avoid Primenet kicking my ass off. Some "poacher" will get bored, and if he would find a prime I would take the money. We could never interdict poaching by a rule. What you can do? Ban the poacher? You lose the computing power. Of course, we would never encourage poaching by a given rule. Fortunately most of the people understand that poaching is waste of time on their side. We already discussed this in many threads, many times. And so on. Good/bad examples can be found on both cases. Last fiddled with by LaurV on 2012-02-22 at 12:50

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Newer X64 build needed	Googulator	Msieve	73	2020-08-30 07:47
Performance of cuda-ecm on newer hardware?	fivemack	GMP-ECM	14	2015-02-12 20:10
Cause this don't belong in the milestone thread	bcp19	Data	30	2012-09-08 15:09
Newer msieves are slow on Core i7	mklasson	Msieve	9	2009-02-18 12:58
Use of large memory pages possible with newer linux kernels	Dresdenboy	Software	3	2003-12-08 14:47

2012-02-22, 12:38	#429
ATH Einyen Dec 2003 Denmark 6126₈ Posts	Here is George view on when assignments are too old, from the previous milestone thread: http://www.mersenneforum.org/showpos...&postcount=443 So if assignment is more than 365 days old and progress<50% or if assignment is more than 548 days old. Apparently these recycling rules are still not implemented or not working properly, or your example M24077267 would have been reassigned 312 days ago.