[QUOTE=chalsall;405663]I'll take it.[/QUOTE]

Just realized; this has now finished.

[QUOTE=chalsall;405803]Just realized; this has now finished.[/QUOTE]

Yeah, I saw that a little bit ago. Thanks for that, much appreciated.

Sometimes...
 

Every now and then I run across an exponent like this:
[URL="http://www.mersenne.org/M46013437"]M46013437[/URL]

Poor other user did it 3 times and still got it wrong. :smile:

That poor guy! Probably he was using faulty hardware i guess..
I am TF'ing it to usual 72 bits, lets see if a pentacheck is needed.

[QUOTE=manfred4;406135]That poor guy! Probably he was using faulty hardware i guess..
I am TF'ing it to usual 72 bits, lets see if a pentacheck is needed.[/QUOTE]

Yup... and it's the exponents like those that make me think there really could be a hidden prime that double-checking will find.

Here's another example I just checked in...
[URL="http://www.mersenne.org/M47223833"]M47223833[/URL]

The fact that I've been finding some where the first 2 (or 3) checks were all wrong is interesting.

I've noticed that the Primenet server will automatically mark a result as "suspect" if it has certain error codes, and those "suspect" results are kind of ignored for the time being, in the sense that the server will hand out that same exponent again as if it was a first-time check, rather than wait for double-checking to get to it.

That's a smart way to do it, because about 50% of the runs with certain error codes (higher than 50% in some cases) will end up with a bad residue.

What I've been interested in are the thousands of known bad results where the error code was zero. Those don't get handed out again right away, of course. Looking back at machines that have historically had bad results with zero error codes, and very few verified good results, I can find their unverified results and run a double-check.

I've done that in a few cases and as expected I showed their result to be bad. Well, I assume they're bad... won't know until someone else does a triple-check, but I trust my machines. :smile: Sometimes I match the original residue and that's cool too... counts as a double-check.

At the moment I have my machines doing a bunch of needed triple-checks for all exponents where the same user did the first and second checks and they didn't match. Curtis has a bunch, but there's a couple hundred from other users too.

Once those are done I may continue my analysis of these "suspect" machines and try to target the most obvious ones... systems that have 10-to-1 bad to good residues, historically, and not too many in the unknown category where there was a zero error code. I cherry picked a handful and I'm testing those now just to see how it plays out with more than one or two. :smile:

I think it would be nice, if you tell us before, what exponents you do, so we (I) can TF them further. As you may have noticed, I don't really like low exponents being LL'ed when they really should be TF'ed 3 or even more levels deeper before! And I think that was one of your goals of this thread, right?

Agreed.

I'm going to TF M47223833.

PrimeNet is not letting me manually reserve it for whatever reason though.

[QUOTE=manfred4;406164]I think it would be nice, if you tell us before, what exponents you do, so we (I) can TF them further. As you may have noticed, I don't really like low exponents being LL'ed when they really should be TF'ed 3 or even more levels deeper before! And I think that was one of your goals of this thread, right?[/QUOTE]

Oh, sure that's cool. Here's a list of all the exponents I currently have assigned to myself with TF less than 71 bits (41 of them). If you feel like taking any of them any farther that's fine with me. I think there's a way to lookup a certain user's assignments and get a full list. I have 264 right now, so the rest of them not listed below are all at 71 bits or above, but just in case you felt like looking at any of those as well...
[CODE]Exponent CurrentTF
43125469 70
43196987 70
43221091 70
43468573 70
43478711 70
43542743 70
43649303 70
43949921 70
44304157 70
44359429 70
44438113 70
44631007 69
44718437 70
44900447 70
44966351 70
45023621 69
45109459 69
45594581 69
45608891 69
46009603 69
46032733 69
46194703 69
46276553 69
46303441 69
46336001 70
46535087 69
46679401 70
46834589 70
46908703 70
46993819 70
47041747 69
47054339 69
47164823 70
47268643 70
47450303 70
47526881 69
48737089 69
49232801 69
49491391 69
50128459 70
50328797 70[/CODE]

[QUOTE=Madpoo;406176]I think there's a way to lookup a certain user's assignments and get a full list. I have 264 right now, so the rest of them not listed below are all at 71 bits or above, but just in case you felt like looking at any of those as well...[/QUOTE]

Oh, maybe not. I must be thinking of the report to show *results* by user, but can't do it for assignments. Oh well.

Here's the list of all the 71+ assignments I have:
[CODE]Exponent CurrentTF
34618891 71
34644007 71
34684453 71
34693397 71
34709803 71
34713001 71
34717981 71
34756739 71
34761523 71
34767791 71
34769473 71
34775801 71
34794007 71
34807043 71
34835161 71
34844519 71
34856743 71
34877671 71
34900589 71
34903541 71
34921087 71
34935709 71
34942091 71
34945909 71
34949689 71
34986779 71
34990667 71
34996849 71
34997201 71
34997213 71
34997741 71
35012623 71
35097871 71
35171561 71
35220019 71
35485759 71
35517907 71
35539891 71
35633677 71
35638639 71
35674481 71
35759939 71
35795779 71
35859541 71
35868887 72
35883167 71
35888659 71
35942717 71
36189233 71
36199781 71
36237881 71
36263053 71
36269881 71
36270203 71
36319301 71
36383569 71
36392387 71
36396727 71
36415319 71
36440051 71
36498421 71
36498731 71
36600259 71
36618787 71
36626587 71
36747407 71
36759847 71
36807349 71
36855521 71
36856531 71
37077791 71
37102957 71
37186733 71
37228129 71
37244719 71
37245499 71
37272493 71
37272629 71
37272931 71
37284367 71
37296587 71
37357711 71
37407511 71
37446287 71
37448933 71
37481677 71
37487887 71
38527997 71
38546999 71
38891939 71
38897039 71
38949707 71
38994863 71
39051871 71
39081641 71
39116053 71
39116257 71
39146453 71
39162359 71
39514081 71
39633457 71
39961193 71
40433671 72
40524257 72
40533539 72
40556653 72
40556983 72
40835909 72
40848961 72
40860517 72
40898527 72
40918783 72
40926211 72
40951727 72
40981891 72
40985897 72
41222689 72
41280011 71
41281609 71
41355557 71
41368543 71
41413781 71
41415029 71
41491979 71
41492051 71
41505901 71
41536529 71
41543413 71
41547251 72
41554949 71
41573341 71
41617861 71
41617867 71
41666749 71
41690947 71
41709793 71
41741503 71
41759611 71
41787743 71
41845519 71
41849641 71
41888131 71
41930293 71
41931817 71
41948747 71
41963771 71
41993869 71
42002707 71
42010261 71
42079043 71
42103637 71
42122611 71
42177661 71
42200309 71
42228569 71
42229129 71
42245627 71
42335581 71
42421823 71
42427211 71
42430627 71
42459293 71
42489983 71
42502111 71
42549821 71
42626471 71
42649811 71
42725759 71
42739063 71
42748711 71
42754223 71
42755303 71
42772889 71
42781799 71
42816911 71
42829121 71
42884651 71
42900733 71
42955547 71
42961517 71
42977551 71
43099729 71
43237781 71
43553401 71
43560481 71
43638977 71
43823191 71
43844147 71
43894699 71
43938941 71
44717027 73
45046171 72
45726851 72
45849691 72
45956227 71
46299763 72
46567541 72
46664521 72
46669759 72
47316187 71
47953303 71
48653383 72
49654117 72
49675459 72
50923991 71
51781111 72
52237331 72
52331561 72
52436017 72
53964947 72
54325283 72
54427367 72
54575839 72
54585137 73
54644257 72
55208323 73
55412029 72
55570813 72
55864219 72
55975649 72
56302951 72
57207373 72
57927197 73[/CODE]

[QUOTE=Madpoo;406176]Oh, sure that's cool. Here's a list of all the exponents I currently have assigned to myself with TF less than 71 bits (41 of them). If you feel like taking any of them any farther that's fine with me. I think there's a way to lookup a certain user's assignments and get a full list. I have 264 right now, so the rest of them not listed below are all at 71 bits or above, but just in case you felt like looking at any of those as well...
[/QUOTE]

I'll take all these to an appropriate level. Should take about a day.

[QUOTE=Mark Rose;406178]I'll take all these to an appropriate level. Should take about a day.[/QUOTE]

If you find any factors, can you also report them in this thread so I don't miss it? Then I can remove any of those from my worktodo files.

Thanks for that. It takes ~ 13 hours to do a 34M exponent on a 10-core chip, so even finding factors for two of the smallest in the list would be a net gain. And not all my systems are dual 10-core beasties either, so that's just a conversation starter there... if you found a factor for a 50M+ exponent that I had running on a 6-core CPU it'd save me about a week, for instance.

(in truth I tend to put the larger exponents on the higher-cored chips)