Is it OK to pick up expired exponents?

[patrik]

I just want to ask if it is okay to pick up expired exponents (by the "usual" method of putting them into the worktodo.txt and communicating).

The reason I ask is that they no longer just disappear from the re-designed exponent status pages, but show as assigned but expired a few days ago.

For over three years (IIRC) I have been working in the 10M digit DC range when the DC wavefront hit this range in late January. The reservations made in January are now beginning to expire. (If I reserve them, I'll be putting them only on fast hardware.)

[chalsall]

Originally Posted by patrik

I just want to ask if it is okay to pick up expired exponents (by the "usual" method of putting them into the worktodo.txt and communicating).

I would argue yes, so long as Primenet gives you an AID using the "usual method" of putting the candidates into the worktodo.txt without an AID, and having Prime95/mprime communicate with Primenet. Any which don't get an AID should be removed from the worktodo.txt file.

George (of course) can speak in more depth with regards to how Primenet handles this situation.

[patrik]

As a test, I tried to reserve two exponents.

33340063 Assigned 2014-01-30 ANONYMOUS D expired on 2014-03-31
33343897 Assigned 2014-01-28 ANONYMOUS D expired on 2014-03-29

The first gave me:

PrimeNet error 40: No assignment
ra: already assigned, exponent 33340063, A: 1, b: 2, c: -1

while the second registered as it should.

However, the exponent I picked up now shows as

33343897 Assigned 2014-01-28 Patrik Johansson D expired on 2014-03-29

Maybe someone else already picked up the first one?

I discovered, just before posting, that none of the exponents show up on my assignments page (this was expected for the first one, but not for the second). So I think I'd better stick to exponents without the "assigned but expired" attribute.

[chalsall]

Originally Posted by patrik

I discovered, just before posting, that none of the exponents show up on my assignments page (this was expected for the first one, but not for the second). So I think I'd better stick to exponents without the "assigned but expired" attribute.

Four axioms, if I may...

1. Code is hard.

2. Humans are error prone.

3. Deterministic systems do *exactly* what you tell them to.

4. Testing is good.

[snme2pm1]

Originally Posted by patrik

is okay to pick up expired exponents

I recognise that your thrust is for Lucas–Lehmer or related double tests.
More generally, in accord with the thread title, I might venture that I have no hesitation to recognise an expired Trial Factor assignment within my interest range and simply provide a quick resolution.
By quick, I mean that upon recognising a particularly obstinate delayed job, and prior to PrimeNet making the exercise publicly available, I might prosecute the work but immediately prior to submitting a result, check that nobody has acquired such assignment.
Only once in the past year did I have such reason to abstain from submitting a result due to somebody acquiring the assignment.
(For that exception, it took the acquirer a couple of weeks to resolve even though I knew what the factor gained result would be, but that's ok).

[TheMawn]

Originally Posted by snme2pm1

I recognise that your thrust is for Lucas–Lehmer or related double tests.
More generally, in accord with the thread title, I might venture that I have no hesitation to recognise an expired Trial Factor assignment within my interest range and simply provide a quick resolution.
By quick, I mean that upon recognising a particularly obstinate delayed job, and prior to PrimeNet making the exercise publicly available, I might prosecute the work but immediately prior to submitting a result, check that nobody has acquired such assignment.
Only once in the past year did I have such reason to abstain from submitting a result due to somebody acquiring the assignment.
(For that exception, it took the acquirer a couple of weeks to resolve even though I knew what the factor gained result would be, but that's ok).

This seems fair, though I might suggest taking work in smaller batches if your assignments expire frequently.

[snme2pm1]

Originally Posted by TheMawn

I might suggest taking work in smaller batches if your assignments expire frequently.

To clarify, I was talking about languishing TF jobs that other members have abandoned until expiry.
I believe that jobs assigned to me have never been left long enough to expire.

[chalsall]

Originally Posted by snme2pm1

To clarify, I was talking about languishing TF jobs that other members have abandoned until expiry.

To clarify, could you give a single example of what you're talking about? Leave off the last three or so digits and replace them with "x"'s if you so wish.

[snme2pm1]

Originally Posted by chalsall

single example.

The most recent example of an expired exponent that I wanted to clear out was 78086249.
Assigned 2013-01-07;<SomeoneElse>;TF;expired on 2014-04-05
I nudged it from 70 bits up to 72 bits soon after spotting it, to save PrimeNet from reassigning it, then further to 74 bits.

[chalsall]

Originally Posted by snme2pm1

The most recent example of an expired exponent that I wanted to clear out was 78086249.

OK. Not terribly valuable work in the short term, but worthwhile in about a year or so. And honorable that you didn't do this work while it was assigned.

http://www.mersenne.org/report_expon...exp_hi=&full=1