Message boards : Number crunching : dear Neasan & Alex
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
bass4lhc

Send message
Joined: 28 Sep 04
Posts: 43
Credit: 249,962
RAC: 0
Message 17813 - Posted: 11 Sep 2007, 4:05:37 UTC

again there is work but no cpu-time donator can get it.

i have been watching this for a few hours. (i am not a fulltime boinc watcher) :)

but if this happens 3 times in a row i don't think this is a coincidence anymore.

and yes, in a few hours you will tell: there was a problem but now work is flowing and it all went while we were watching. we hope you got some.

why not just tell what is going on?

i truly like to donate my cpu time to this project.
i don't think i get informed on what is going on. boinc-wise that is.
ID: 17813 · Report as offensive     Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 24 Nov 06
Posts: 76
Credit: 7,914,481
RAC: 27,114
Message 17814 - Posted: 11 Sep 2007, 4:29:18 UTC - in response to Message 17813.  

again there is work but no cpu-time donator can get it.

upload/download server is disabled. See here:

http://lhcathome.cern.ch/lhcathome/server_status.php
Dublin, California
Team: SETI.USA

ID: 17814 · Report as offensive     Reply Quote
KAMasud

Send message
Joined: 7 Oct 06
Posts: 114
Credit: 23,192
RAC: 0
Message 17815 - Posted: 11 Sep 2007, 5:05:48 UTC


:-) Zombie67 LoL the point is, like bass4lhc is pointing out, the problem is now getting chronic in nature :)it also now seems that while we are swarming around like flies to grab a couple Alex and Neasan are rolling on the floor in their mirth watching our antics :( just a feeling that i am getting :) may not be true ;)
Regards
Masud.
ID: 17815 · Report as offensive     Reply Quote
Christoph

Send message
Joined: 25 Aug 05
Posts: 69
Credit: 306,627
RAC: 0
Message 17816 - Posted: 11 Sep 2007, 9:29:43 UTC

Hallo, I only just now also recognized that all the WUs are still sitting there. Obviously the upload/download & scheduler services are not running. But in my eyes the server status page is not very accurate on these two programs. even when work was flowing previously, they always had the state "disabled". Also since a few weeks there are 3 work units "in progress", which should have either been returned by now for sure, or be re-distributed due to meeting deadline. Hope Alex and Neasan will get this fixed soon, at least that the work is flowing out. The correct indication on the server status page is nice to have, but of course just now not priority.

Good luck, Christoph
Christoph
ID: 17816 · Report as offensive     Reply Quote
Profile Ocean Archer
Avatar

Send message
Joined: 13 Jul 05
Posts: 143
Credit: 263,300
RAC: 0
Message 17818 - Posted: 11 Sep 2007, 13:18:34 UTC

Hey - I can remember in the not-too-far-distant past where we had similar problems with SETI. Always seemed to happen over a weekend. Finally traced down (I think) to some knuckle-dragging sanitation engineer who would liberate power plugs from service receptacles just so he could run his floor buffer.

Not saying that it's the problem here, but one never knows.


If I've lived this long, I've gotta be that old
ID: 17818 · Report as offensive     Reply Quote
Profile Neasan
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 30 Nov 06
Posts: 234
Credit: 11,078
RAC: 0
Message 17842 - Posted: 13 Sep 2007, 9:37:45 UTC

It is not a problem any more. We are holding the jobs because we hope to get more very soon so you'll have even more to crunch.
ID: 17842 · Report as offensive     Reply Quote
Ariel
Avatar

Send message
Joined: 7 Mar 07
Posts: 59
Credit: 7,906
RAC: 0
Message 17861 - Posted: 13 Sep 2007, 19:12:16 UTC - in response to Message 17842.  

It is not a problem any more. We are holding the jobs because we hope to get more very soon so you'll have even more to crunch.


Thanks Neasan. I got a nice fistful of WU's this morning from 8:46am to 8:50am PST... Four minutes of happiness in my otherwise dreary wasteland of a life so bereft of excitement, wonder, and inspiration.

Actually, I'm a little exited about the LHC switch-on this November ;)




Ariel: Certified "Too Cute for LHC" Cruncher!


. . . . . . . . . . . . -- Consider the lilies.
ID: 17861 · Report as offensive     Reply Quote
bass4lhc

Send message
Joined: 28 Sep 04
Posts: 43
Credit: 249,962
RAC: 0
Message 17877 - Posted: 14 Sep 2007, 1:26:37 UTC - in response to Message 17842.  

It is not a problem any more. We are holding the jobs because we hope to get more very soon so you'll have even more to crunch.


i am sorry, i am not a very smart human.
but this is like: i will not eat this week because there might be more food next week.

even stupid me doesn't understand this.


ID: 17877 · Report as offensive     Reply Quote
PovAddict
Avatar

Send message
Joined: 14 Jul 05
Posts: 275
Credit: 49,291
RAC: 0
Message 17878 - Posted: 14 Sep 2007, 1:34:53 UTC - in response to Message 17877.  
Last modified: 14 Sep 2007, 1:48:08 UTC

It is not a problem any more. We are holding the jobs because we hope to get more very soon so you'll have even more to crunch.


i am sorry, i am not a very smart human.
but this is like: i will not eat this week because there might be more food next week.

even stupid me doesn't understand this.

Don't blame yourself; what he said indeed doesn't make much sense...
ID: 17878 · Report as offensive     Reply Quote
Profile rbpeake

Send message
Joined: 17 Sep 04
Posts: 99
Credit: 30,618,118
RAC: 3,938
Message 17879 - Posted: 14 Sep 2007, 1:35:46 UTC - in response to Message 17877.  

It is not a problem any more. We are holding the jobs because we hope to get more very soon so you'll have even more to crunch.


i am sorry, i am not a very smart human.
but this is like: i will not eat this week because there might be more food next week.

even stupid me doesn't understand this.


We did just get fed, they did not hold back for the hungry hordes! ;)

Regards,
Bob P.
ID: 17879 · Report as offensive     Reply Quote
picantecomputing

Send message
Joined: 1 Jan 07
Posts: 19
Credit: 44,943
RAC: 0
Message 17880 - Posted: 14 Sep 2007, 3:03:35 UTC - in response to Message 17878.  

Don't blame yourself; what he said indeed doesn't make much sense...

Actually, if I'm interpreting their intentions correctly, holding back WUs until there are lots to go around actually makes quite a bit of sense. If they only release work in small batches, the same farms that are always banging on the door and caching large numbers of WUs for themselves are going to get the lion's share very quickly at the beginning each (small) release, since they're spaced fairly far apart. Having larger batches allows more hosts (and more users) to get at least some work over a larger window of opportunity - and definitely more than most have been able to get up to now. So, yes, we have to wait - but I think it's ultimately better for everyone this way. Unless my logic is flawed.
ID: 17880 · Report as offensive     Reply Quote
daemion

Send message
Joined: 4 Dec 06
Posts: 33
Credit: 75,491
RAC: 0
Message 17881 - Posted: 14 Sep 2007, 3:24:21 UTC - in response to Message 17880.  

Don't blame yourself; what he said indeed doesn't make much sense...

Actually, if I'm interpreting their intentions correctly, holding back WUs until there are lots to go around actually makes quite a bit of sense. If they only release work in small batches, the same farms that are always banging on the door and caching large numbers of WUs for themselves are going to get the lion's share very quickly at the beginning each (small) release, since they're spaced fairly far apart. Having larger batches allows more hosts (and more users) to get at least some work over a larger window of opportunity - and definitely more than most have been able to get up to now. So, yes, we have to wait - but I think it's ultimately better for everyone this way. Unless my logic is flawed.



FWIW this is the way i understood it as well
ID: 17881 · Report as offensive     Reply Quote
Daxa

Send message
Joined: 29 Dec 06
Posts: 100
Credit: 184,937
RAC: 0
Message 17887 - Posted: 14 Sep 2007, 6:43:45 UTC
Last modified: 14 Sep 2007, 6:44:07 UTC

"We are holding the jobs because we hope to get more very soon so you'll have even more to crunch."
[sarcasm]
I think it's such a good idea, I've decided not to eat any of the food in the kitchen until my wife does the shopping next week. That way, I will lose weight and when I finally do eat there will be lots of food to throw away and that makes me feel like a prosperous American. I love this country so much.
[/sarcasm]



_______

"Three quarks for Muster Mark!"
. . . . . . . - James Joyce, Finnegans Wake . . . .

ID: 17887 · Report as offensive     Reply Quote
Profile Trog Dog

Send message
Joined: 25 Nov 05
Posts: 39
Credit: 41,119
RAC: 0
Message 17889 - Posted: 14 Sep 2007, 11:32:17 UTC - in response to Message 17887.  

"We are holding the jobs because we hope to get more very soon so you'll have even more to crunch."
[sarcasm]
I think it's such a good idea, I've decided not to eat any of the food in the kitchen until my wife does the shopping next week. That way, I will lose weight and when I finally do eat there will be lots of food to throw away and that makes me feel like a prosperous American. I love this country so much.
[/sarcasm]




Pity the minus button doesn't work.
ID: 17889 · Report as offensive     Reply Quote
Daxa

Send message
Joined: 29 Dec 06
Posts: 100
Credit: 184,937
RAC: 0
Message 17900 - Posted: 15 Sep 2007, 0:37:55 UTC - in response to Message 17889.  
Last modified: 15 Sep 2007, 0:38:08 UTC

Pity the minus button doesn't work.

Lighten up, Dawg.

(maybe you missed the irony within the sarcasm, or maybe silver is the new gold, or maybe I'm just pretending to be an annoying hipster douchebag.)


_______

"Three quarks for Muster Mark!"
. . . . . . . - James Joyce, Finnegans Wake . . . .

ID: 17900 · Report as offensive     Reply Quote
bass4lhc

Send message
Joined: 28 Sep 04
Posts: 43
Credit: 249,962
RAC: 0
Message 17911 - Posted: 18 Sep 2007, 3:10:26 UTC - in response to Message 17880.  

Don't blame yourself; what he said indeed doesn't make much sense...

Actually, if I'm interpreting their intentions correctly, holding back WUs until there are lots to go around actually makes quite a bit of sense. [......]So, yes, we have to wait - but I think it's ultimately better for everyone this way. Unless my logic is flawed.


i'm sorry, but yes i think your logic is flawed.
it's not that difficult, if you have more cpu's you get more wu's.
if there are 10 wu's 9 go to boinc farms.
if there are 1000 wu's 900 or more go to boinc farms.
holding back wu's makes no difference.

who cares, we all want to donate to science.
this is not a problem. more cpu's is good boinc-wise.

there are so many boinc projects to keep your cpu busy.

information from the projects we donate to would be nice however. (hint to lhc)



ID: 17911 · Report as offensive     Reply Quote
picantecomputing

Send message
Joined: 1 Jan 07
Posts: 19
Credit: 44,943
RAC: 0
Message 17912 - Posted: 18 Sep 2007, 3:49:33 UTC - in response to Message 17911.  

i'm sorry, but yes i think your logic is flawed.
it's not that difficult, if you have more cpu's you get more wu's.
if there are 10 wu's 9 go to boinc farms.
if there are 1000 wu's 900 or more go to boinc farms.
holding back wu's makes no difference.

Well, I'm not sure I totally agree. If there's actually a quota per CPU (and I'm not sure offhand if one is in force or not), then yes, the farms would still get the lion's share of the work. BUT, if the quota were low enough, then that would prevent the farms from running down ALL the WUs in a larger batch before x number of other users could get a share of the work. If there's no quota, then yes, the size of the batch is pretty much irrelevant, to some extent.

And if there ISN'T a quota, I guess my question would be, why not?
ID: 17912 · Report as offensive     Reply Quote
Adam23

Send message
Joined: 30 May 06
Posts: 40
Credit: 219,288
RAC: 10
Message 17914 - Posted: 18 Sep 2007, 10:18:49 UTC - in response to Message 17912.  
Last modified: 18 Sep 2007, 10:24:07 UTC


Well, I'm not sure I totally agree. If there's actually a quota per CPU (and I'm not sure offhand if one is in force or not), then yes, the farms would still get the lion's share of the work. BUT, if the quota were low enough, then that would prevent the farms from running down ALL the WUs in a larger batch before x number of other users could get a share of the work. If there's no quota, then yes, the size of the batch is pretty much irrelevant, to some extent.

And if there ISN'T a quota, I guess my question would be, why not?


Yes. And the quota for the first batch could be set to say for example 5 WUs, yes five WUs per CPU for a start. Then everyone could grasp some WUs to have pleasure of :) Then of course upon completion of them, another batch would be sent. So that even speed-of-light machines probably would not know any difference.
ID: 17914 · Report as offensive     Reply Quote
Profile Neasan
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 30 Nov 06
Posts: 234
Credit: 11,078
RAC: 0
Message 17915 - Posted: 18 Sep 2007, 11:00:50 UTC

There are two quotas:
1) The amount of WUs you can download at a time
2) The number of WUs you can download for each CPU

We have tried to set these very low and keep the number of jobs high so (as some have guessed, picantecomputing and daemion) more people get the work when we release it, this worked the last few times (i.e. it took a while for the 20,000 or so jobs to go and a good few different people seemed to get them) but due to a hiccup with the server status script people knew jobs were coming so prepared to download as many as they could anyway they could.

I'm sorry if you don't like how we were trying to do it but as can plainly be seen we have to do it some how. Some people are complaining after getting 80 WUs when others got none. We thought holding back till we had around 20,000 and the releasing them with low quotas would evenly distribute the work as it has done before but we were wrong.

p.s. LHC turn on has now been delayed it will be May 2008 at the earliest.
ID: 17915 · Report as offensive     Reply Quote
Profile [B^S] Dora

Send message
Joined: 27 Apr 06
Posts: 26
Credit: 13,559
RAC: 0
Message 17916 - Posted: 18 Sep 2007, 13:26:22 UTC

I am pretty much hopelessly ignorant about how severs work, but from my limited perspective, it seems that it is not just the quota, or number, of wu's, but where, actually, your particular cpu is in the polling process cycle when the wu's hit?

If my machine happens to be somewhere around the 1 min to 3 min range and you release a batch, chances are I will get some.

But, I just checked my messages and LHC just deferred communication for 1 hr 27 min and 41 sec. If you released them right now I would not get any, most likely, unless it was a HUGE batch. Or I played some games with the BOINC manager and forced it out of cycle.

So isn't this a BOINC manager thing more than a LHC thing? Isn't there a way to alter how often my machine can poll LHC? Say 10 minute intervals?

That will not help with small batches, of course, because we seem to have some very determined wu suckers.

Personally, I think you have done just about everything you can from your end with that size batch.

And thank you. We ARE interested in the science, too, I hope. It's kind of amusing that we are fighting over how many cookies we get instead of the best recipe. I am sure you are actually more interested in what the wu's that we return tell you than in how many you give us ;-)






ID: 17916 · Report as offensive     Reply Quote
1 · 2 · Next

Message boards : Number crunching : dear Neasan & Alex


©2024 CERN