mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GPU to 72 (https://www.mersenneforum.org/forumdisplay.php?f=95)
-   -   results not needed (https://www.mersenneforum.org/showthread.php?t=19969)

Dubslow 2015-11-24 03:44

Such as these seven:

[code]Guilty-Spark 34568419 C - Verified 2015-09-09 11:16 79.3 6B4A1F93BE4EF0__ 40.9060
Guilty-Spark 34564379 C - Verified 2015-09-09 11:16 80.7 77A3E5018FF35C__ 40.9012
Guilty-Spark 34564399 C - Verified 2015-09-09 11:16 80.7 C518088875C47A__ 40.9012
Guilty-Spark 34543867 C - Verified 2015-09-09 11:16 89.0 06C1EED10BDC1C__ 40.8769
Guilty-Spark 34541833 C - Verified 2015-09-09 11:16 89.5 D7504BD8F7BA5E__ 40.8745
Guilty-Spark 34539227 C - Verified 2015-09-09 11:16 90.7 93B5328A78F157__ 40.8714
Guilty-Spark 34538591 C - Verified 2015-09-09 11:16 91.0 7E5331E8F3A4C9__ 40.8707[/code]

My Cat I DC computer was offline but crunching for several weeks. I "extended" the reservations via the manual reservation form, but apparently the 60 day deadline overrode those extensions.

Madpoo 2015-11-24 04:09

[QUOTE=chalsall;416994]Even with Primenet's LL / DC / P-1 efforts this happens. Someone makes a promise of completion but then doesn't deliver; the candidate is then reassigned and then the original assignee completes.

It is very difficult to manage independent actors....[/QUOTE]

Primenet *could* simply reject unassigned results coming in, but most of us probably agree with George that it would be weird to reject work just because it wasn't assigned (whether it was an expired one that just kept going, or was outright poached).

Madpoo 2015-11-24 04:13

[QUOTE=Dubslow;417076]Such as these seven:

[code]Guilty-Spark 34568419 C - Verified 2015-09-09 11:16 79.3 6B4A1F93BE4EF0__ 40.9060
Guilty-Spark 34564379 C - Verified 2015-09-09 11:16 80.7 77A3E5018FF35C__ 40.9012
Guilty-Spark 34564399 C - Verified 2015-09-09 11:16 80.7 C518088875C47A__ 40.9012
Guilty-Spark 34543867 C - Verified 2015-09-09 11:16 89.0 06C1EED10BDC1C__ 40.8769
Guilty-Spark 34541833 C - Verified 2015-09-09 11:16 89.5 D7504BD8F7BA5E__ 40.8745
Guilty-Spark 34539227 C - Verified 2015-09-09 11:16 90.7 93B5328A78F157__ 40.8714
Guilty-Spark 34538591 C - Verified 2015-09-09 11:16 91.0 7E5331E8F3A4C9__ 40.8707[/code]

My Cat I DC computer was offline but crunching for several weeks. I "extended" the reservations via the manual reservation form, but apparently the 60 day deadline overrode those extensions.[/QUOTE]

Yeah, the cat 1 stuff in the bleeding-edge have more aggressive expiration dates. But that's by design. My suggestion would be that if you have a machine that doesn't check in regularly and provide updates (not merely extending them from the website), uncheck the box saying you want those preferred assignments. Otherwise it's likely to happen again.

Might not have been the case for you... maybe you had those assignments for a really long time and the leading edge caught up with you?

Dubslow 2015-11-24 05:50

The offline but crunching was a one-off thing. You can see that the machine has turned in several results since then, and continues to crunch Cat I DCs.

chalsall 2015-11-24 19:49

[QUOTE=Dubslow;417087]The offline but crunching was a one-off thing. You can see that the machine has turned in several results since then, and continues to crunch Cat I DCs.[/QUOTE]

Let's be honest here... If one makes a promise, that's a contract.

If one doesn't do what they promise, then they've broken the contract.

I myself have done this; one of my stacks lost its' UPS and then came back online three weeks later. Unfortunately candidates were reassigned in the interim and my machines competed the work thus wasting work.

Dubslow 2015-11-25 03:46

[QUOTE=chalsall;417154]Let's be honest here... If one makes a promise, that's a contract.

If one doesn't do what they promise, then they've broken the contract.

I myself have done this; one of my stacks lost its' UPS and then came back online three weeks later. Unfortunately candidates were reassigned in the interim and my machines competed the work thus wasting work.[/QUOTE]

Yeah... at least now I know I must ask for a manual intervention, or share the work in progress.

kladner 2016-12-08 15:31

It seems that monst has taken 2 TF assignments which I had for about 3 weeks, but which had not expired. I discovered it this morning when one of them came up "Not Needed." I hope jumping the gun does not become a habit. That was about 7 hours work, for me.
[I]Manual[/I]
[B][URL="http://www.mersenne.org/report_exponent/default.php?exp_lo=81929833&full=1"]81929833[/URL][/B]LL TF73752016-11-15 05:11:572016-12-08 09:20:2070.048
[B][URL="http://www.mersenne.org/report_exponent/default.php?exp_lo=81929773&full=1"]81929773[/URL][/B]LL TF73752016-11-15 05:11:572016-12-08 05:20:2370.048

ixfd64 2016-12-08 22:38

Is there any good reason to not accept already-completed ranges?

Given than PrimeNet already accepts LL results beyond double checks, I don't see any reason to reject duplicate TF results. If anything, having someone independently verify a range isn't a bad idea anyway.

apocalypse 2017-02-22 05:41

FWIW, I just discovered that monst submitted TF results for these exponents which I had reserved through GPU72 and which were a few days from expiry (incl. 2 with factors).

[FONT="Courier New"]Computer Exponent WorkType From To Assigned * Completed GHz Days
Manual ‡ 83871163 LL TF 71 75 2017-01-25 04:50:40 2017-02-21 21:15:13 39.915
Manual . 83870947 LL TF 71 75 2017-01-25 04:50:40 2017-02-21 18:04:42 85.534
Manual . 83871079 LL TF 71 75 2017-01-25 04:50:40 2017-02-21 18:04:42 85.534
Manual ‡ 83870923 LL TF 71 75 2017-01-25 04:50:40 2017-02-21 18:04:22 85.534
[/FONT]

kladner 2017-02-22 05:43

I sympathize.. :beatdown:

0PolarBearsHere 2017-02-22 11:08

.


All times are UTC. The time now is 07:57.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.