Estimate to complete: feature request.

Message boards : Bug reports : Estimate to complete: feature request.

To post messages, you must log in.

AuthorMessage
Profile Death [Kiev]
Avatar

Send message
Joined: 21 Jul 08
Posts: 14
Credit: 124,289
RAC: 0
Message 2054 - Posted: 25 May 2011, 9:45:34 UTC

Good day.

http://www.enigmaathome.net/server_status.php

Available workunits:
m3_pldrv59_2
# of workunits: 540000
done 548518
remaining 0
in progress 325

can you please add one more line to every WU pool - time to complete?

we know: Workunits validated last hour 203

so TTC = in progess / WU's validated = time in hours.
======
wbr, Me. Dead J. Dona
ID: 2054 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile TJM
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 25 Aug 07
Posts: 843
Credit: 267,994,998
RAC: 0
Message 2077 - Posted: 24 Jun 2011, 22:51:58 UTC - in response to Message 2054.  

It won't work like that, mostly because the server doesn't know the actual processing rate. It jumps up and down when various daemons are started or stopped (the server automatically disables validators when the load gets too high); also it's totally impossible to predict *when* the workunit will come back from the client. Most of the workunits are back really fast (average turnaround time is less than 2 days), however the rest may come back close to deadline or never at all, so the processing rate slows down while the age of workunit batch increases. It's always slowest near the end, when the server has to wait for resends.



M4 Project homepage
M4 Project wiki
ID: 2077 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Death [Kiev]
Avatar

Send message
Joined: 21 Jul 08
Posts: 14
Credit: 124,289
RAC: 0
Message 2585 - Posted: 25 Jun 2013, 10:06:42 UTC
Last modified: 25 Jun 2013, 10:07:25 UTC

well that's why it's called ESTIMATE ))

i suppose it's better to have some even inaccurate rates, rather than nothing.


pleeease


======
wbr, Me. Dead J. Dona
ID: 2585 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile TJM
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 25 Aug 07
Posts: 843
Credit: 267,994,998
RAC: 0
Message 2586 - Posted: 25 Jun 2013, 15:38:35 UTC - in response to Message 2585.  

Then what kind of estimate would you like ?

The search for message key is based purely on luck multiplied by the strenght of the algorithm and dictionaries used.
Every single workunit has equal chance to break the message, so I can't predict when (or IF) it will be broken. Reaching end of the batch means that it failed and it's a thing I'd like to avoid.



M4 Project homepage
M4 Project wiki
ID: 2586 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
miki_e

Send message
Joined: 9 Jan 12
Posts: 5
Credit: 2,768,117
RAC: 0
Message 2589 - Posted: 26 Jun 2013, 15:48:16 UTC - in response to Message 2586.  

Hi,

Well, more statistics will not help us to solve the message, (faster or better algorithm will...) however, since we already have some stats on the server...

Lets suppose that algorithm is good and WILL find the solution. If we are really not lucky, the solution can happen to be THE LAST combination tried....

With this supposition in mind, "Death [Kiev]" would probably like to see what percentage of the whole job is making each work-unit batch, what percentage of the batch has been already returned (yes, everybody can calculate it from existing stats) and VERY ROUGH time estimate of the time needed to complete the batch based on AVERAGE return rate for the batch (for instance last 48 hours - or week average or something like this).... just to have a better idea where we are standing...
ID: 2589 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sword

Send message
Joined: 18 Nov 09
Posts: 11
Credit: 1,052,256
RAC: 0
Message 2590 - Posted: 26 Jun 2013, 17:43:22 UTC - in response to Message 2589.  

Given the above mentioned the daily rate of processing the new batch m4_vroln72_4 is now about 89000 work units in average.

With that speed the remaining batch of m4_vroln72_4 will terminate in about
500 days if no solution is found earlier.

As some have said before,nobody can estimate when the break eventually will come.

Anyway,I think this whole operation is very exciting indeed to participate in over more years to come.

And patience over time is needed trying to fulfill our goals here.
So did the real codebreakers also do in BP during WW2 !
ID: 2590 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Bug reports : Estimate to complete: feature request.




Copyright © 2024 TJM