Newer milestone thread

[rudy235]

Originally Posted by Madpoo

You may be right. I'd be in heap big doo doo if it turned out to be prime and I had encouraged someone to poach it.

In this case I was basing my comment on the fact that I highly doubted the current assignee would complete before it expired, so it was a preemptive strike. Might not hold much weight in a court of law, but I stand by that assessment anyway. LOL

For what it's worth, that assignment for M74095187 still hasn't updated since Oct 1 ... yeah, it's expired now but since it's in progress it would normally finish... the client won't unassign it and abandon it or anything.

I may be surprised and see him turn in a result in the next couple weeks, proving me wrong, but I doubt it.

But the exponent 74095187 had been already tested by mikeblas. Suspect result or not, this was a second testing. The result of ATH verified the previous result of AE5C73EB430E2D72 The other user "icedragon" verification is moot because if it coincides with the other 2 it is a third and thusly unneeded and if it differs it will not be taken into account.

[Madpoo]

Originally Posted by rudy235

But the exponent 74095187 had been already tested by mikeblas. Suspect result or not, this was a second testing. The result of ATH verified the previous result of AE5C73EB430E2D72 The other user "icedragon" verification is moot because if it coincides with the other 2 it is a third and thusly unneeded and if it differs it will not be taken into account.

True, it's no longer needed, but the assignment will keep going.

In fact it checked in again today... still at the same 80.3%, no change since last time (6 days ago).

[rudy235]

Originally Posted by Madpoo

True, it's no longer needed, but the assignment will keep going.

In fact it checked in again today... still at the same 80.3%, no change since last time (6 days ago).

3 weeks later the user is still at 80.3%


2017-07-30 Icedragon LL double-check LL 80.3 % 2017-10-24 2017-10-05

[rudy235]

Less than 20,000 primes remain to be tested for the first time to complete the classic view of 79.3 million.



This was the situation only 22 months ago (Jan 11, 2016)


Just about 9 times more!

[Madpoo]

With PRP assignments and results now in the mix, it was messing up the milestone page. PRP assignments were showing up even with "exclude doublechecks".

Well, technically they were NOT doublechecks of another PRP test, but there was already an LL test done for the ones showing up so... not really a first time check either.

It was also failing to take into account when a PRP test was done but not an LL test.

Anyway, I got that all sorted out and now the milestone page should be accurate, and the assignments page with "exclude doublechecks" will NOT include PRP assignments when an LL test has already been done. If you want to see those overlapping PRPs, don't exclude doublechecks.

Basically, for now I'm going to pretend that a PRP test for an exponent that's already had an LL test is technically a double-check.

[retina]

Originally Posted by Madpoo

With PRP assignments and results now in the mix, it was messing up the milestone page. PRP assignments were showing up even with "exclude doublechecks".

Well, technically they were NOT doublechecks of another PRP test, but there was already an LL test done for the ones showing up so... not really a first time check either.

It was also failing to take into account when a PRP test was done but not an LL test.

Anyway, I got that all sorted out and now the milestone page should be accurate, and the assignments page with "exclude doublechecks" will NOT include PRP assignments when an LL test has already been done. If you want to see those overlapping PRPs, don't exclude doublechecks.

Basically, for now I'm going to pretend that a PRP test for an exponent that's already had an LL test is technically a double-check.

So now those PRP "double checks" only seem to show up if I untick both the LL and DC boxes together.

https://www.mersenne.org/assignments...&exp1=1&extf=1

That doesn't feel right to me.

[Madpoo]

Originally Posted by retina

So now those PRP "double checks" only seem to show up if I untick both the LL and DC boxes together.

https://www.mersenne.org/assignments...&exp1=1&extf=1

That doesn't feel right to me.

I'll have to think about this more when I'm not as tired...

Basically though, the idea is that exponents can be checked with either PRP or LL tests, and one or the other counts as a first-time check. If it's been LL tested and someone is doing a PRP test (which is the problem I saw and worked around), that's not a double-check.

Only a double-check of the same type (LL or PRP) is really a double-check, otherwise it's a different first-time check.

Anyway, if I somehow missed some use case that didn't seem obvious, I'll give it a re-think.

[ATH]

Originally Posted by Madpoo

Only a double-check of the same type (LL or PRP) is really a double-check, otherwise it's a different first-time check.

If an exponent has an LL test without any errors and a user tries to start a PRP test, maybe the server should force the assignment into an LL double check to avoid wasted work? And by the same logic if there is a successful PRP test without errors and the user tries to start an LL test, force it into a PRP double check?

Originally Posted by Madpoo

Basically though, the idea is that exponents can be checked with either PRP or LL tests, and one or the other counts as a first-time check. If it's been LL tested and someone is doing a PRP test (which is the problem I saw and worked around), that's not a double-check.

So PRP double checks (when there already is an existing PRP first time test) should only show up if the "Exclude double-check assignments" is unchecked, and PRP first time test (regardless of existing LL test or not) should only show up when "Exclude first-time LL/PRP assignments" is unchecked.

[rudy235]

75838877 LL LL, 93.80% 7 -22 2017-09-17 2017-11-05 2017-11-06 2017-11-06 Edward Miller
75909439 LL LL, 59.30% 8 -19 2017-09-22 2017-11-06 2017-11-07 2017-11-09 Edward Miller


These two exponents have not been updated since the first week of November. They will eventually expire in 7 0r 8 days but in the meantime...

[Uncwilly]

Originally Posted by rudy235

These two exponents have not been updated since the first week of November. They will eventually expire in 7 0r 8 days but in the meantime...

I doubt there will be a need for a "Christmas Miracle" this year. We have been mowing through milestones so well, we have no need to be hasty. Also, since they are first time checks.....

Seriously, 5 of the million level milestones cleared this year after last year's 10. We have passed the largest known prime, so any new prime found by a first time check will be a WR prime. The past 2 years have seen 7 of the million level milestones cleared too. This is really doing well.

[rudy235]

One of those two is gone.. And now there are 8 in total to go.

75909439 LL LL, 59.30% 8 -19 2017-09-22 2017-11-06 2017-11-07 2017-11-09 Edward Miller