Gah... I don't know what happened, but lately Curtis' account has been churning out a lot of self-verified work. It looks like some old assignments expired and got reassigned to another of his computers, and then the original machines got turned back on and are continuing the work, but as a double-check now.

Or... They're all in the 67M range which makes me think they're being expired quicker under the new rules and probably shouldn't have been assigned to those slower machines in the first place.

The assignment rules won't give out an assignment to the same person who did the first check, but there's nothing preventing a user from getting a new assignment on something they already had, but expired. That's probably fine most of the time, but not right now.

Maybe once the new expiration rules have settled into place and all these expirations of stuff assigned under the old rules has worked through, it won't be an issue...

There are about 5 or 6 right now, plus I count another 32 or so that are probably going to finish up soon as more self-verified work.

I've been keeping up with the periodic entries that show up here and there, but is there anyone interested in doing independent checks on these?

Here's what needs independent checks currently:
[CODE]DoubleCheck=67588013,75,1
DoubleCheck=67657747,75,1
DoubleCheck=67665629,75,1
DoubleCheck=73309741,75,1
DoubleCheck=73312229,75,1
DoubleCheck=92408501,75,0[/CODE]

Would it be a good idea to tweak the assignment rules to never give the same exponent to the same person (even if expired)?

Mismatched DC against a 2007 result:
36556253
Factored to 71.

These Cat0 exponents are already taken automatically before anyone can read your message here :)

[QUOTE=ATH;438131]These Cat0 exponents are already taken automatically before anyone can read your message here :)[/QUOTE]
Oh. OK. I guess I have not paid enough attention, but will remember hence.

On another note, CUDALucas on a GTX 460 agrees with your result for 41518229. :smile:

[QUOTE=kladner;438134]On another note, CUDALucas on a GTX 460 agrees with your result for 41518229. :smile:[/QUOTE]

Thank you :smile:

Here are the thresholds: [url]http://www.mersenne.org/thresholds/[/url]
Anything in the 36M and maybe 37M are probably reassigned very quickly.

Thanks! I had not appreciated that aspect of running Smallest Exponent DCs.

[QUOTE=kladner;438139]Thanks! I had not appreciated that aspect of running Smallest Exponent DCs.[/QUOTE]

I took a big chunk (~200) of 36M exponents a couple weeks back to supplement some 60-70M testing I'm doing (they work well in pairs like that).

I cherry picked my assignments to include any where the previous tester had at least one bad result, and I've been surprised so far at how many of those I ended up mismatching. I guess it might be close to the 5% average, but it's still kind of fun and surprising since I wasn't really expecting much. Definitely not part of my strategic double-checking, I just needed smaller exponents to work alongside the larger ones and avoid memory contention. :smile:

[QUOTE=Madpoo;438077]Gah... I don't know what happened, but lately Curtis' account has been churning out a lot of self-verified work. It looks like some old assignments expired and got reassigned to another of his computers, and then the original machines got turned back on and are continuing the work, but as a double-check now.

Or... They're all in the 67M range which makes me think they're being expired quicker under the new rules and probably shouldn't have been assigned to those slower machines in the first place.

The assignment rules won't give out an assignment to the same person who did the first check, but there's nothing preventing a user from getting a new assignment on something they already had, but expired. That's probably fine most of the time, but not right now.

Maybe once the new expiration rules have settled into place and all these expirations of stuff assigned under the old rules has worked through, it won't be an issue...

There are about 5 or 6 right now, plus I count another 32 or so that are probably going to finish up soon as more self-verified work.

I've been keeping up with the periodic entries that show up here and there, but is there anyone interested in doing independent checks on these?

Here's what needs independent checks currently:
[CODE]DoubleCheck=67588013,75,1
DoubleCheck=67657747,75,1
DoubleCheck=67665629,75,1
DoubleCheck=73309741,75,1
DoubleCheck=73312229,75,1
DoubleCheck=92408501,75,0[/CODE][/QUOTE]

Looks like no one else has spoken up to claim these, so I've added them to my worktodo lists.

[QUOTE=endless mike;438252]Looks like no one else has spoken up to claim these, so I've added them to my worktodo lists.[/QUOTE]

Thanks!

I would have tackled these, but I recently took on a batch of about 400 exponents in the 60-70M range that needed triple-checking and realized these would take my systems several months to churn through.

I'd only been grabbing a couple weeks of work at a time but I thought I'd save myself some effort by doing a big batch at once and filling up the worktodo, but then I realized how fun it was to be able to squeeze in some other fun stuff along the way. Oh well... I'll go back to that arrangement once this mega load finishes.

could use a TC
I have already done a TC on a different machine on the side.
Mine is correct.
Thanks

41937001 LL Unverified;2009-06-27;-Anonymous-;8DB80F8E430E75__
41937001 LL Unverified;2016-06-03;srow7;D88C505E833AE8__