Newer milestone thread

[Prime95]

Originally Posted by cuBerBruce

Are exponents assigned before the rule changes expired according to the old rules?

Yes

[Uncwilly]

Originally Posted by rudy235

All exponents below 35,190,131 have been tested and double-checked.
All exponents below 63,350,927 have been tested at least once.

Countdown to first time checking all exponents below 64M: 2 (Estimated completion : 2016-05-21)
Countdown to first time checking all exponents below 65M: 3 (Estimated completion : 2016-05-21)
Countdown to first time checking all exponents below 66M: 4 (Estimated completion : 2016-05-21)
Countdown to first time checking all exponents below 67M: 32 (Estimated completion : 2016-05-21)

All exponents below 35,518,181 have been tested and double-checked.
All exponents below 63,350,927 have been tested at least once.

Countdown to first time checking all exponents below 64M: 1 (Estimated completion : 2016-05-21)
Countdown to first time checking all exponents below 65M: 1 (Estimated completion : 2016-05-21)
Countdown to first time checking all exponents below 66M: 1 (Estimated completion : 2016-05-21)
Countdown to first time checking all exponents below 67M: 17 (Estimated completion : 2016-06-07)


Estimated time until this gets poached????
I am against such activity. I suspect it will occur in less than 2 weeks.

[Mark Rose]

I think it will be poached before the 20th.

[rudy235]

Seems that exponent 63,350,927 is stuck at 54.1% for a long time.

[Madpoo]

Originally Posted by rudy235

Seems that exponent 63,350,927 is stuck at 54.1% for a long time.

I've been watching it to see if it will finish before expiring. Up until a week ago, I estimated it was going to be a real close shave and might actually finish a day or so before expiration.

If that were the case, I would have done something to extend it and let it finish before expiring, just so it wouldn't be reassigned only to have this original test check in a day later.

However, it was stuck, as you noticed, for about a week and then finally checked in again today: 8 days later and only 0.2 % progress.

I'm now calculating it to finish in 45 days, well past the 31 days it has left. So... that sucks. The user must have had some issue or their computer was shut off for a week. Unless it picked up the pace drastically or got moved to a faster machine, it will expire, based on it's typical daily progress of roughly 1% daily.

In hindsight, this system should have never been assigned this exponent, but that's neither here nor there and has been addressed already, but now what?

*Should* it be "poached", knowing it is unlikely to finish in time?

On the "NO" side, we'd have to speculate that it could be reassigned to another CPU just like this one, taking near (or over) the max amount of days to finish and maybe expire. Of course if that happened, this current assignment will probably limp past the finish line 2 weeks after it was reassigned.

On the "YES" side, the changes in priority assignment rules that George implemented stand a good chance of making sure the new assignment is done swiftly, preferably *before* the current assignment checks in.

Normally if I'm even considering poaching something at all, I'm going to look at whether the assignment was abandoned... no check-in/progress in several weeks AND it's close to expiring anyway. This is just one of those cases where the user actually is doing stuff, but way slower than anticipated, so I have mixed feelings.

What does the peanut gallery think? Let it be and let nature take it's course? Or wait for someone else to "do the dirty" and poach it so that we can publicly curse them but privately thank them? LOL

[retina]

Originally Posted by Madpoo

What does the peanut gallery think? Let it be and let nature take it's course? Or wait for someone else to "do the dirty" and poach it so that we can publicly curse them but privately thank them? LOL

First time checks are no problem with another user doing it also. But two things should be followed: 1) No more than two people at once, so coordinate. 2) Don't expire the exponent and have the system tell the user the work is now not needed. Still accept the work when it finally finishes.

And of course DC is a different matter. Don't waste cycles doing a second DC when someone is already actively working on it.

[chalsall]

Originally Posted by Madpoo

What does the peanut gallery think? Let it be and let nature take it's course? Or wait for someone else to "do the dirty" and poach it so that we can publicly curse them but privately thank them? LOL

Why don't YOU poach it yourself immediately?

My thinking is that you have the firepower to finish this in a couple of days or so, and it is currently holding up _three_ milestones (not that I personally care).

This has the advantage that no one else will be tempted (thus wasting resources), and you have the admin rights to grant Summy the "First LL Credit" in the exceptionally unlikely case where he actually finishes in time.

I'm really glad George implemented the most recently revised rules. Should eliminate almost all cases like this going forward.

(As an aside, it is amusing to note that Summy has seven other low assignments, none of which are projected to complete before expiring. He never should have been given Cat 1s in the first place.)

[Madpoo]

Originally Posted by chalsall

Why don't YOU poach it yourself immediately?

Oh, I'm tempted for sure, but I wanted to see what opinions others had.

Plus, Summy still has 31 days before it expires, and who knows, maybe he/she will pull a rabbit out of the hat and hit the hidden "turbo" button.

[chalsall]

Originally Posted by Madpoo

Plus, Summy still has 31 days before it expires, and who knows, maybe he/she will pull a rabbit out of the hat and hit the hidden "turbo" button.

Like what? Perhaps not running eight (8#) LL tests on a four core (4#) CPU?

Just poach it, before someone else does. Then it becomes a legitimate DC, and no throughput is lost.

[gjmccrac]

I would say NO to poaching.

George created the Cat 0 for a reason.

Let it actually work instead of thinking you know better.

How can you tell if the new Cat 0 is going to work if you circumvent it.

[ixfd64]

Are we really that impatient?

The assignment will be completed sooner or later. Is the difference of a few days that critical?