Panic Mode On (80) Server Problems?


log in

Advanced search

Message boards : Number crunching : Panic Mode On (80) Server Problems?

Previous · 1 . . . 3 · 4 · 5 · 6 · 7 · 8 · 9 . . . 25 · Next
Author Message
N9JFE David S
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 10760
Credit: 13,484,676
RAC: 14,618
United States
Message 1323659 - Posted: 2 Jan 2013, 17:24:45 UTC - in response to Message 1323623.

TBar wrote:
This part, Try NOT to use it, unless there is no other way.
I would consider being completely out of nVidia tasks for over half a day "no other way". This will be listed under "Anonymous Platform", which as far as I know, is wide open. If you can construct an App that preforms 10x better than the stock App, go for it. Just imagine I've spent scads of money on the latest Intel i9000 processor, I'm using a proto App, and it's cooking with gas. Seems to be a lot of concern over credits....

Just remember, Fred's comments were before the unrealistic Limits were imposed. Allotting just 100 units to the much faster GPUs is not realistic. I wouldn't be attempting this if my GPU hadn't have run out of tasks in less than half a day while the CPUs have days worth of tasks left. Maybe a more realistic allotment is in order? I would like to see at least 100 per GPU/Platform. This machine also has an AMD card which was fortunate enough to be working APs when the lights went out. It is just now approaching 50% cache. 100 for the nVidia card, 100 for the AMD, or 100 for each GPU. Then people such as myself wouldn't be involved in such endeavors. :-)


It's the point of view ..

For you - if you reschedule WUs around in BOINC - your PC can crunch a few minutes/hours longer/faster SAH WUs.
But then, all WUs calculated, no new WUs from SAH?
Maybe for you it's OK to get less Credits for the results.
This is your decision.

Other members/your wingmen look to their online overviews and are confused why a few results get less Credits - than normally.
It's not their decision.
They need to tolerate it because someone other rescheduled WUs in his BOINC?
They get less Credits. This is fair?

I used the BoincRescheduler tool before 'CreditNew'.
Since then never again, because I think it's not fair towards my wingmen.

For a long time I had EMail contact with the SETI@home Director Dr. David P. Anderson and he said:
'Since the new Credit System 'CreditNew' you shouldn't longer use a reschedule tool for to send CPU WUs to GPU or GPU WUs to CPU.'


* Best regards! :-) * Sutaru Tsureku, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. *

Take all your concern over some arbitrary numbers, plus a few 100 billion of those arbitrary numbers, and see what they will buy you. I'm afraid they won't buy much, and they certainly won't help you find ET any faster. I'm here for results, which means completing as many units as my means will allow. If I could afford it, I would buy the latest and greatest gear and not worry about any associated cost. I certainly wouldn't worry about some arbitrary number someone decided to 'grant' me. Those numbers will not help you in your final quest. I'm here to find results, not to amass some useless number. I'm sorry if your goal is not the same.
Have a good day.

As I understand this, you have a bunch of CPU tasks and you're out of GPU tasks, and you want to reassign the CPU tasks to the GPU so they'll get done faster. What's the point? Let's say you have 80 hours' worth of work and you transfer it and it gets done in 12 hours instead, but the servers are still down and you still can't get any more work? What have you gained? Nothing that I can see.

____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


TBar
Volunteer tester
Send message
Joined: 22 May 99
Posts: 1177
Credit: 41,717,751
RAC: 112,870
United States
Message 1323676 - Posted: 2 Jan 2013, 18:07:28 UTC - in response to Message 1323659.

TBar wrote:
This part, Try NOT to use it, unless there is no other way.
I would consider being completely out of nVidia tasks for over half a day "no other way". This will be listed under "Anonymous Platform", which as far as I know, is wide open. If you can construct an App that preforms 10x better than the stock App, go for it. Just imagine I've spent scads of money on the latest Intel i9000 processor, I'm using a proto App, and it's cooking with gas. Seems to be a lot of concern over credits....

Just remember, Fred's comments were before the unrealistic Limits were imposed. Allotting just 100 units to the much faster GPUs is not realistic. I wouldn't be attempting this if my GPU hadn't have run out of tasks in less than half a day while the CPUs have days worth of tasks left. Maybe a more realistic allotment is in order? I would like to see at least 100 per GPU/Platform. This machine also has an AMD card which was fortunate enough to be working APs when the lights went out. It is just now approaching 50% cache. 100 for the nVidia card, 100 for the AMD, or 100 for each GPU. Then people such as myself wouldn't be involved in such endeavors. :-)


It's the point of view ..

For you - if you reschedule WUs around in BOINC - your PC can crunch a few minutes/hours longer/faster SAH WUs.
But then, all WUs calculated, no new WUs from SAH?
Maybe for you it's OK to get less Credits for the results.
This is your decision.

Other members/your wingmen look to their online overviews and are confused why a few results get less Credits - than normally.
It's not their decision.
They need to tolerate it because someone other rescheduled WUs in his BOINC?
They get less Credits. This is fair?

I used the BoincRescheduler tool before 'CreditNew'.
Since then never again, because I think it's not fair towards my wingmen.

For a long time I had EMail contact with the SETI@home Director Dr. David P. Anderson and he said:
'Since the new Credit System 'CreditNew' you shouldn't longer use a reschedule tool for to send CPU WUs to GPU or GPU WUs to CPU.'


* Best regards! :-) * Sutaru Tsureku, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. *

Take all your concern over some arbitrary numbers, plus a few 100 billion of those arbitrary numbers, and see what they will buy you. I'm afraid they won't buy much, and they certainly won't help you find ET any faster. I'm here for results, which means completing as many units as my means will allow. If I could afford it, I would buy the latest and greatest gear and not worry about any associated cost. I certainly wouldn't worry about some arbitrary number someone decided to 'grant' me. Those numbers will not help you in your final quest. I'm here to find results, not to amass some useless number. I'm sorry if your goal is not the same.
Have a good day.

As I understand this, you have a bunch of CPU tasks and you're out of GPU tasks, and you want to reassign the CPU tasks to the GPU so they'll get done faster. What's the point? Let's say you have 80 hours' worth of work and you transfer it and it gets done in 12 hours instead, but the servers are still down and you still can't get any more work? What have you gained? Nothing that I can see.

Except that just as my nVidia card finished the last non-vlar, the servers came back up and now it is back to working 609s without losing a beat. I still had over half a days worth of vlars I was about to begin with the CPUs. As I see it, there was a point. I actually had done the math and considered the servers coming back up today a strong possibility. Years ago, when it wasn't really needed, Fred constructed an App to do basically what I had done. Why did Fred construct such an App back before the current Limits? What do you suppose was his motivation? I wasn't aware this project was filled with as many credit junkies as it appears to be. The last time I worked SETI all that mattered was Units completed and Hours spent. Maybe this project should revert to it's original intent.

N9JFE David S
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 10760
Credit: 13,484,676
RAC: 14,618
United States
Message 1323683 - Posted: 2 Jan 2013, 18:23:54 UTC
Last modified: 2 Jan 2013, 18:24:33 UTC

The crickets live!

And have you seen the numbers of results ready to send?
____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 26 May 99
Posts: 6821
Credit: 24,634,744
RAC: 26,829
United Kingdom
Message 1323687 - Posted: 2 Jan 2013, 18:30:09 UTC
Last modified: 2 Jan 2013, 18:38:01 UTC

My 303 reports spread across here an Beta disappeared in a flash!!

Unfortunately here has "No tasks available"
____________


Today is life, the only life we're sure of. Make the most of today.

Profile ivan
Volunteer tester
Avatar
Send message
Joined: 5 Mar 01
Posts: 579
Credit: 131,577,100
RAC: 114,011
United Kingdom
Message 1323691 - Posted: 2 Jan 2013, 18:35:42 UTC - in response to Message 1323687.

My 303 uploads spread across here an Beta disappeared in a flash!!

Unfortunately here has "No tasks available"

All my machines managed to report just after 1700 UTC but only two were able to get any new tasks before a tap was turned off somewhere.
____________

N9JFE David S
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 10760
Credit: 13,484,676
RAC: 14,618
United States
Message 1323697 - Posted: 2 Jan 2013, 18:47:53 UTC - in response to Message 1323691.

My 303 uploads spread across here an Beta disappeared in a flash!!

Unfortunately here has "No tasks available"

All my machines managed to report just after 1700 UTC but only two were able to get any new tasks before a tap was turned off somewhere.

None of mine asked for anything, but I realized one had a bunch of ghosts. I suspended Einstein on that one, then did another Seti update and they were resent.

____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


bill
Send message
Joined: 16 Jun 99
Posts: 859
Credit: 22,368,915
RAC: 17,056
United States
Message 1323708 - Posted: 2 Jan 2013, 19:10:03 UTC - in response to Message 1323659.


As I understand this, you have a bunch of CPU tasks and you're out of GPU tasks, and you want to reassign the CPU tasks to the GPU so they'll get done faster. What's the point? Let's say you have 80 hours' worth of work and you transfer it and it gets done in 12 hours instead, but the servers are still down and you still can't get any more work? What have you gained? Nothing that I can see.


Well, processing power is made available sooner to other projects. If you care about such things.

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3868
Credit: 107,175,532
RAC: 99,941
United States
Message 1323714 - Posted: 2 Jan 2013, 19:26:39 UTC - in response to Message 1323683.

The crickets live!

And have you seen the numbers of results ready to send?

"Results ready to send 4,738,529" [George Takei]Oh my.[/George Takei]
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

mikeej42
Send message
Joined: 26 Oct 00
Posts: 109
Credit: 787,755,191
RAC: 60,253
United States
Message 1323718 - Posted: 2 Jan 2013, 19:31:35 UTC

The WEB page is reporting "Results ready to send 4,743,749" but I have over a hundred systems that only get
"1/2/2013 1:24:47 PM | SETI@home | Project has no tasks available".
Even if I query from the same server every 15 seconds I can not get any tasks. I never get the message that the request was too recent.

Anyone have an explanation or what I need to change?
____________

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3868
Credit: 107,175,532
RAC: 99,941
United States
Message 1323723 - Posted: 2 Jan 2013, 19:42:23 UTC - in response to Message 1323718.
Last modified: 2 Jan 2013, 19:42:38 UTC

The WEB page is reporting "Results ready to send 4,743,749" but I have over a hundred systems that only get
"1/2/2013 1:24:47 PM | SETI@home | Project has no tasks available".
Even if I query from the same server every 15 seconds I can not get any tasks. I never get the message that the request was too recent.

Anyone have an explanation or what I need to change?

The work requests are fulfilled by the feeder. Which only holds a few hundred at a time. It also only refills every so few seconds. Originally I think it was 100 tasks every 60 seconds, but it has been estimated that it has to be much higher and faster these days.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

N9JFE David S
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 10760
Credit: 13,484,676
RAC: 14,618
United States
Message 1323726 - Posted: 2 Jan 2013, 19:49:14 UTC - in response to Message 1323723.

The WEB page is reporting "Results ready to send 4,743,749" but I have over a hundred systems that only get
"1/2/2013 1:24:47 PM | SETI@home | Project has no tasks available".
Even if I query from the same server every 15 seconds I can not get any tasks. I never get the message that the request was too recent.

Anyone have an explanation or what I need to change?

The work requests are fulfilled by the feeder. Which only holds a few hundred at a time. It also only refills every so few seconds. Originally I think it was 100 tasks every 60 seconds, but it has been estimated that it has to be much higher and faster these days.

The crickets seem to have reached a plateau, which is slightly but significantly lower than the pre-crash plateau.

____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


Keith White
Avatar
Send message
Joined: 29 May 99
Posts: 370
Credit: 2,704,577
RAC: 2,235
United States
Message 1323734 - Posted: 2 Jan 2013, 19:54:56 UTC - in response to Message 1323726.

The crickets seem to have reached a plateau, which is slightly but significantly lower than the pre-crash plateau.

"slightly but significantly" Is that like jumbo shrimp? :p

____________
"Life is just nature's way of keeping meat fresh." - The Doctor

Profile Dorsai
Avatar
Send message
Joined: 7 Sep 04
Posts: 464
Credit: 1,684,459
RAC: 1,428
United Kingdom
Message 1323736 - Posted: 2 Jan 2013, 19:59:32 UTC - in response to Message 1323726.

From a quick glance over the the above posts am I correct in assuming the situation can be summarized as follows:

There are plenty of work units to send, but they are not getting sent. (For whatever reason.)





____________

Foamy is "Lord and Master".
(Oh, + some Classic WUs too.)

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3868
Credit: 107,175,532
RAC: 99,941
United States
Message 1323743 - Posted: 2 Jan 2013, 20:18:07 UTC - in response to Message 1323736.

From a quick glance over the the above posts am I correct in assuming the situation can be summarized as follows:

There are plenty of work units to send, but they are not getting sent. (For whatever reason.)


The situation is normal after a period of machines running down their caches. Thousands of machines all requesting all the work they can at the same time.

Looking at my list of all tasks I have been allocated around 150 tasks across my machines in the past 15 minutes. Doing a quick check looks like they are all downloaded without any ghosts being made in the process.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

N9JFE David S
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 10760
Credit: 13,484,676
RAC: 14,618
United States
Message 1323746 - Posted: 2 Jan 2013, 20:24:39 UTC - in response to Message 1323736.

From a quick glance over the the above posts am I correct in assuming the situation can be summarized as follows:

There are plenty of work units to send, but they are not getting sent. (For whatever reason.)

The number of ready to send has dropped by over 40,000.

____________
David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.


mikeej42
Send message
Joined: 26 Oct 00
Posts: 109
Credit: 787,755,191
RAC: 60,253
United States
Message 1323749 - Posted: 2 Jan 2013, 20:29:38 UTC - in response to Message 1323723.


Even if I query from the same server every 15 seconds I can not get any tasks. I never get the message that the request was too recent.

Anyone have an explanation or what I need to change?

The work requests are fulfilled by the feeder. Which only holds a few hundred at a time. It also only refills every so few seconds. Originally I think it was 100 tasks every 60 seconds, but it has been estimated that it has to be much higher and faster these days.


I have over 150 machines that have not got any tasks today and they are getting tasks from Einstein which has a resource of 0.

Any idea why polling less than 303 seconds does not return the normal error about the request being too recent?
____________

Profile ivan
Volunteer tester
Avatar
Send message
Joined: 5 Mar 01
Posts: 579
Credit: 131,577,100
RAC: 114,011
United Kingdom
Message 1323750 - Posted: 2 Jan 2013, 20:34:48 UTC - in response to Message 1323746.

From a quick glance over the the above posts am I correct in assuming the situation can be summarized as follows:

There are plenty of work units to send, but they are not getting sent. (For whatever reason.)

The number of ready to send has dropped by over 40,000.

Yay! My out-of-work machine got one new GPU task!
____________

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3868
Credit: 107,175,532
RAC: 99,941
United States
Message 1323752 - Posted: 2 Jan 2013, 20:37:15 UTC - in response to Message 1323749.


Even if I query from the same server every 15 seconds I can not get any tasks. I never get the message that the request was too recent.

Anyone have an explanation or what I need to change?

The work requests are fulfilled by the feeder. Which only holds a few hundred at a time. It also only refills every so few seconds. Originally I think it was 100 tasks every 60 seconds, but it has been estimated that it has to be much higher and faster these days.


I have over 150 machines that have not got any tasks today and they are getting tasks from Einstein which has a resource of 0.

Any idea why polling less than 303 seconds does not return the normal error about the request being too recent?

They must have switched that off, perhaps temporally. I am not getting that from BETA either.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Profile ivan
Volunteer tester
Avatar
Send message
Joined: 5 Mar 01
Posts: 579
Credit: 131,577,100
RAC: 114,011
United Kingdom
Message 1323755 - Posted: 2 Jan 2013, 20:42:10 UTC - in response to Message 1323750.


Yay! My out-of-work machine got one new GPU task!

...which took 7'34" to complete...

____________

mikeej42
Send message
Joined: 26 Oct 00
Posts: 109
Credit: 787,755,191
RAC: 60,253
United States
Message 1323757 - Posted: 2 Jan 2013, 20:44:32 UTC - in response to Message 1323752.

Okay Thanks.

One of my machines finally got a task. Too many machines with empty caches fighting at the same time.

Guess I will just have to find some alcohol to get some patience, :^}
____________

Previous · 1 . . . 3 · 4 · 5 · 6 · 7 · 8 · 9 . . . 25 · Next

Message boards : Number crunching : Panic Mode On (80) Server Problems?

Copyright © 2014 University of California