OFFICIAL "SERVER PROBLEMS" THREAD - Page 200

Uncwilly · 2021-06-08, 12:41

Also, Ben and Curtis both run large groups of machines. Curtis's machines are distributed among many actual daily users and all fetch assignments in an automated manner. Ben's are cloud resources on stable hardware. Those two are trusted because the likelihood of one of their machines accidently getting another's cert assignment by random chance is high.

LaurV · 2021-06-17, 18:05

By playing with the link (i.e. manually editing the type to be 1011) we found out that actually there is a top producers list for P+1.

Any reason why it is not "publicly" accessible?
(i.e. can't be chosen from here, without editing the link, which edited link I didn't give, on purpose

in case it is deliberately kept hidden for whatever reason - now I suppose everybody will go there and edit the link, haha).

Uncwilly · 2021-06-17, 18:15

Quote:

Originally Posted by LaurV

By playing with the link (i.e. manually editing the type to be 1011) we found out that actually there is a top producers list for P+1.

Any reason why it is not "publicly" accessible?

This is an edge case. Stop being a whiner.

Viliam Furik · 2021-06-17, 18:51

Quote:

Originally Posted by LaurV

By playing with the link (i.e. manually editing the type to be 1011) we found out that actually there is a top producers list for P+1.

Any reason why it is not "publicly" accessible?
(i.e. can't be chosen from here, without editing the link, which edited link I didn't give, on purpose

in case it is deliberately kept hidden for whatever reason - now I suppose everybody will go there and edit the link, haha).

Here is the link for others; uncle LaurV doesn't want to give you sweets.

James Heinrich · 2021-06-17, 20:02

Quote:

Originally Posted by LaurV

Any reason why it is not "publicly" accessible?

No, just that none of the people who run the site remember all the little things that need updating when a new result type is added. I have updated that report page and included P+1 as a selectable report type.

LaurV · 2021-06-18, 10:33

Quote:

Originally Posted by James Heinrich

I have updated that report page and included P+1 as a selectable report type.

Ye'r ma' man!

Edit: @Unc:

Edit 2: WHAAATT?? James! PUT THOSE F'KING DATE FIELDS BACK AS THEY WERE ! I swear to all gods I will quit Gimps if you don't put them back. NO JOKE.

Happy5214 · 2021-06-22, 09:39

And now it limits the ranges to the top 500, cutting me out of most of the lists.

slandrum · 2021-07-04, 02:49

DC Cat 0 continues to shrink - it's now down to 71 exponents - 129 of the 200 that the server is counting to determine where the threshold is set are exponents that have had one LL test, and then a PRP test with certification.

In a few weeks/months there will no longer be any Cat 0 assignments made, cat 0 threshold will be below the line marking where all exponents have been tested and verified unless the server is corrected.

This will start to affect reaching milestones. If all the categories start to artificially shrink, the purpose they serve starts to go away, and DC assignments will be made with really long expirations to machines that shouldn't be getting them near the trailing edge.

If cat 1 eventually evaporates, then expirations will no longer be enforced except for the "haven't contacted the server in 60 days" expiration.

slandrum · 2021-07-04, 03:13

There's a problem with the 103M range: there are two assignments available that are never being assigned.

Additionally it shows 3699 exponents that have LTC/PRP with no double check, but the available assignments for DC and the ones actually assigned only add up to 3690. Outstanding CERT assignments may account for this, I don't think they show up as DC assignments on the work distribution map - but maybe they should?

slandrum · 2021-07-04, 15:34

Quote:

Originally Posted by slandrum

Additionally it shows 3699 exponents that have LTC/PRP with no double check, but the available assignments for DC and the ones actually assigned only add up to 3690. Outstanding CERT assignments may account for this, I don't think they show up as DC assignments on the work distribution map - but maybe they should?

There aren't any CERT assignments in the 103M range at this time, so that doesn't account for the discrepancy between the exponents needing DC vs. exponents available for DC + exponents assigned for DC.

slandrum · 2021-07-05, 00:54

Quote:

Originally Posted by slandrum

DC Cat 0 continues to shrink - it's now down to 71 exponents - 129 of the 200 that the server is counting to determine where the threshold is set are exponents that have had one LL test, and then a PRP test with certification.

It looks like this has been corrected. Thanks!

Just in time too, a large number of (now) cat 0 and cat 1 assignments expired today (an account with a large number of DC assignments stopped checking in 60 days ago).

2021-06-17, 18:05	#2191
LaurV Romulan Interpreter Jun 2011 Thailand 7·1,373 Posts	By playing with the link (i.e. manually editing the type to be 1011) we found out that actually there is a top producers list for P+1. Any reason why it is not "publicly" accessible? (i.e. can't be chosen from here, without editing the link, which edited link I didn't give, on purpose in case it is deliberately kept hidden for whatever reason - now I suppose everybody will go there and edit the link, haha). Last fiddled with by LaurV on 2021-06-17 at 18:07

2021-06-22, 09:39	#2196
Happy5214 "Alexander" Nov 2008 The Alamo City 688₁₀ Posts	And now it limits the ranges to the top 500, cutting me out of most of the lists. Last fiddled with by Happy5214 on 2021-06-22 at 09:40 Reason: Typo

2021-07-04, 02:49	#2197
slandrum Jan 2021 California 11·13 Posts	DC Cat 0 vanishing DC Cat 0 continues to shrink - it's now down to 71 exponents - 129 of the 200 that the server is counting to determine where the threshold is set are exponents that have had one LL test, and then a PRP test with certification. In a few weeks/months there will no longer be any Cat 0 assignments made, cat 0 threshold will be below the line marking where all exponents have been tested and verified unless the server is corrected. This will start to affect reaching milestones. If all the categories start to artificially shrink, the purpose they serve starts to go away, and DC assignments will be made with really long expirations to machines that shouldn't be getting them near the trailing edge. If cat 1 eventually evaporates, then expirations will no longer be enforced except for the "haven't contacted the server in 60 days" expiration. Last fiddled with by slandrum on 2021-07-04 at 02:57

2021-07-04, 03:13	#2198
slandrum Jan 2021 California 11·13 Posts	There's a problem with the 103M range: there are two assignments available that are never being assigned. Additionally it shows 3699 exponents that have LTC/PRP with no double check, but the available assignments for DC and the ones actually assigned only add up to 3690. Outstanding CERT assignments may account for this, I don't think they show up as DC assignments on the work distribution map - but maybe they should? Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Official "Faits erronés dans de belles-lettres" thread	ewmayer	Lounge	39	2015-05-19 01:08
Official "all-Greek-to-me Fiction Literature and Cinema" Thread	ewmayer	Science & Technology	41	2014-04-16 11:54
Official "Lasciate ogne speranza" whinge-thread	cheesehead	Soap Box	56	2013-06-29 01:42
Official "Ernst is a deceiving bully and George is a meanie" thread	cheesehead	Soap Box	61	2013-06-11 04:30
Official "String copy Statement Considered Harmful" thread	Dubslow	Programming	19	2012-05-31 17:49

2021-06-08, 12:41	#2190
Uncwilly 6809 > 6502 """"""""""""""""""" Aug 2003 101×103 Posts 263A₁₆ Posts	Also, Ben and Curtis both run large groups of machines. Curtis's machines are distributed among many actual daily users and all fetch assignments in an automated manner. Ben's are cloud resources on stable hardware. Those two are trusted because the likelihood of one of their machines accidently getting another's cert assignment by random chance is high.