the messages from last night are gone, cause I started and restarted BOINC.
I will see if this happens again and post the messages.
The following is a little different, both WUs are willing to run, but I still show as being overcommitted, although not in EDF???......with only 4 hours total runtime for both units together.
7/4/2005 11:23:13 AM||General prefs: no separate prefs for home; using your defaults
7/4/2005 11:23:13 AM||Remote control not allowed; using loopback address
7/4/2005 11:23:13 AM|Einstein@Home|Deferring computation for result l1_0755.5__0755.9_0.1_T03_S4lA_1
7/4/2005 11:23:13 AM|LHC@home|Resuming computation for result wjun2A_v6s4hvpac_mqx__13__64.259_59.269__4_6__6__50_1_sixvf_boinc17759_3 using sixtrack version 4.67
7/4/2005 11:23:13 AM||Suspending work fetch because computer is overcommitted.
I'm suspecting that this weirdness might be due to something involved with the LHC WU and running a version beyond BOINC 4.19. Perhaps the scheduler or something is having problems figuring out whats going on with the LHC units.
Pretty sure I read over there, that they hadn't updating their end beyond 4.19 yet, since they were offline for awhile.
When I have time, I'll hop over to the LHC forums, and post some info. I want to see what happens when I download a new unit first though...........
7/4/2005 11:33:44 AM||request_reschedule_cpus: project op
7/4/2005 11:33:44 AM||Allowing work fetch again.
7/4/2005 11:33:45 AM|LHC@home|Sending scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi
7/4/2005 11:33:45 AM|LHC@home|Requesting 0 seconds of work, returning 0 results
7/4/2005 11:33:46 AM|LHC@home|Scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi succeeded
7/4/2005 12:33:44 PM|Einstein@Home|Restarting result l1_0755.5__0755.9_0.1_T03_S4lA_1 using einstein version 4.79
7/4/2005 12:33:44 PM|LHC@home|Pausing result wjun2A_v6s4hvpac_mqx__13__64.259_59.269__4_6__6__50_1_sixvf_boinc17759_3 (removed from memory)
7/4/2005 12:33:44 PM||request_reschedule_cpus: process exited
7/4/2005 12:33:44 PM||Suspending work fetch because computer is overcommitted.
7/4/2005 12:33:44 PM||Using earliest-deadline-first scheduling because computer is overcommitted.
I have one workunit with 2hr to completion(E@H) and one other with 20 minutes(LHC)......and am put into EDF mode.
seems trippy.
PC----uhh, Pentium III 1.0Ghz, 383 Ram, XP.......curiously though, my computer type listed here under the account is weird......let me see......
GenuineIntel
x86 Family 6 Model 8 Stepping 10 1096MHz
what's that? used to say Pentium III....have recently upgraded from Me to XP...
7/4/2005 11:33:44 AM||request_reschedule_cpus: project op
7/4/2005 11:33:44 AM||Allowing work fetch again.
7/4/2005 11:33:45 AM|LHC@home|Sending scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi
7/4/2005 11:33:45 AM|LHC@home|Requesting 0 seconds of work, returning 0 results
7/4/2005 11:33:46 AM|LHC@home|Scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi succeeded
7/4/2005 12:33:44 PM|Einstein@Home|Restarting result l1_0755.5__0755.9_0.1_T03_S4lA_1 using einstein version 4.79
7/4/2005 12:33:44 PM|LHC@home|Pausing result wjun2A_v6s4hvpac_mqx__13__64.259_59.269__4_6__6__50_1_sixvf_boinc17759_3 (removed from memory)
7/4/2005 12:33:44 PM||request_reschedule_cpus: process exited
7/4/2005 12:33:44 PM||Suspending work fetch because computer is overcommitted.
7/4/2005 12:33:44 PM||Using earliest-deadline-first scheduling because computer is overcommitted.
I have one workunit with 2hr to completion(E@H) and one other with 20 minutes(LHC)......and am put into EDF mode.
seems trippy.
PC----uhh, Pentium III 1.0Ghz, 383 Ram, XP.......curiously though, my computer type listed here under the account is weird......let me see......
GenuineIntel
x86 Family 6 Model 8 Stepping 10 1096MHz
what's that? used to say Pentium III....have recently upgraded from Me to XP...
who knows.
still crunching.
How many E@H do you have now? and LHC@H??
I run a p4 3.0 and it take about 7.78 hrs to complete E@H and about 2.86 for LHC@H and you say your running a P3 1.0. It may have something to do with speed and debt....
Message Type: Status Message
This Message means that the BOINC Client Software has determined that one of the following triggers is true:
Computer is over committed,
Less than 1 day until deadline, or
Deadline is before reconnect time.
Until all of these conditions are false Work Units will be processed in Earliest Deadline First order and will not switch.
The CPU Scheduler is running the Work Unit with the earliest Deadline instead of the Work Unit with the highest Short Term Debt.
Note:
This is the so-called "panic-mode".
[edit]Version Information
I have 1 E@H with about 2 hours left to complete, and one LHC with about 20 minutes to complete.
JKeck Wrote:
"What are your active fractions for that computer?"
What's an "active" fractioin? Each of the two projects is set at 50%.
have new message. Can't don't have the exact to copy again,
but the gist was this: When communicated with the Einstein server/BOINC scheduler or whatever, got a red message that said.............."computer is only running 13% of the time, and BOINC ~30% of that time......not enough to complete unit, entering EDF mode."
THis causing the problem would make sense, but the figure stating the PC is only running 13% of the time is incorrect. My PC is on all the time.......the BOINC figure is probably correct in that I have BOINC running about 30% of the time....but thats 30% of 24hrs.....or about 8 hours a day. Which is enough time to complete one EAH unit, and one LHC unit within the deadlines, by a wide margin.
Times........E@H units take me about 12hrs I think. and LHC...6 or 7 hours. per unit......rough estimate. Still should come out to enough time in my book.
I think BOINC thinking my PC is only on for 13% of the day..around 3 hours, and BOINC only getting 30% of that, is what's causing the problem.
Only thing is, is that my PC is on 100% of the day, and BOINC gets 30% of that time.......which is enough to complete work.
What's an "active" fractioin? Each of the two projects is set at 50%.
have new message. Can't don't have the exact to copy again,
but the gist was this: When communicated with the Einstein server/BOINC scheduler or whatever, got a red message that said.............."computer is only running 13% of the time, and BOINC ~30% of that time......not enough to complete unit, entering EDF mode."
That is the active fractions and it is what is causing EDF mode as well. It will correct itself over time.
OK, that's got to be what it is, as you say......as long as all is normal, and will correct with time, I'm happy, and now know a little more about what's going on.
Thanks for the input JKeck and Fuzzy!
(typing in these forums is kinda weird, characters get all messed up.. used to the other....)
Multiply on_frac * active_frac and you have the % of time that BOINC believes that your computer is running BOINC programs. These numbers multiplied together giva an answer of about 0.04 giving BOINC the impression that your computer is spending almost no time running BOINC apps. This will slowly self correct over the next couple of months.
You may find a BOINC manager for OS X published lately on the download pages. However it has some ... well ... unexpected flaws regarding the interface design
yeah, i can confirm that.
- it consumes FAR TOO MUCH CPU time. looks like it updates every second which is absolutely unnecessary. every minute or so would be sufficient.
- it is possible to make the columns wider, but you can't swap columns or sort the data in some custom order
- when i select open boinc manager from the dock icon menu, the window opens behind all other windows. i have to click the icon to bring it forward.
- the window should appear when the dock iocon is clicked, whether the window is open behind other windows or not.
- the graphs under the statistics tab should all have the same horizontal scale, and that scale should be linear.
i'm sure there's more, but i should have been in bed an hour ago.
does some admin or developer ever read this thread?
the messages from last night
)
the messages from last night are gone, cause I started and restarted BOINC.
I will see if this happens again and post the messages.
The following is a little different, both WUs are willing to run, but I still show as being overcommitted, although not in EDF???......with only 4 hours total runtime for both units together.
7/4/2005 11:23:13 AM||General prefs: no separate prefs for home; using your defaults
7/4/2005 11:23:13 AM||Remote control not allowed; using loopback address
7/4/2005 11:23:13 AM|Einstein@Home|Deferring computation for result l1_0755.5__0755.9_0.1_T03_S4lA_1
7/4/2005 11:23:13 AM|LHC@home|Resuming computation for result wjun2A_v6s4hvpac_mqx__13__64.259_59.269__4_6__6__50_1_sixvf_boinc17759_3 using sixtrack version 4.67
7/4/2005 11:23:13 AM||Suspending work fetch because computer is overcommitted.
I'm suspecting that this weirdness might be due to something involved with the LHC WU and running a version beyond BOINC 4.19. Perhaps the scheduler or something is having problems figuring out whats going on with the LHC units.
Pretty sure I read over there, that they hadn't updating their end beyond 4.19 yet, since they were offline for awhile.
When I have time, I'll hop over to the LHC forums, and post some info. I want to see what happens when I download a new unit first though...........
Update: here's the log
)
Update:
here's the log after the above:
7/4/2005 11:33:44 AM||request_reschedule_cpus: project op
7/4/2005 11:33:44 AM||Allowing work fetch again.
7/4/2005 11:33:45 AM|LHC@home|Sending scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi
7/4/2005 11:33:45 AM|LHC@home|Requesting 0 seconds of work, returning 0 results
7/4/2005 11:33:46 AM|LHC@home|Scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi succeeded
7/4/2005 12:33:44 PM|Einstein@Home|Restarting result l1_0755.5__0755.9_0.1_T03_S4lA_1 using einstein version 4.79
7/4/2005 12:33:44 PM|LHC@home|Pausing result wjun2A_v6s4hvpac_mqx__13__64.259_59.269__4_6__6__50_1_sixvf_boinc17759_3 (removed from memory)
7/4/2005 12:33:44 PM||request_reschedule_cpus: process exited
7/4/2005 12:33:44 PM||Suspending work fetch because computer is overcommitted.
7/4/2005 12:33:44 PM||Using earliest-deadline-first scheduling because computer is overcommitted.
I have one workunit with 2hr to completion(E@H) and one other with 20 minutes(LHC)......and am put into EDF mode.
seems trippy.
PC----uhh, Pentium III 1.0Ghz, 383 Ram, XP.......curiously though, my computer type listed here under the account is weird......let me see......
GenuineIntel
x86 Family 6 Model 8 Stepping 10 1096MHz
what's that? used to say Pentium III....have recently upgraded from Me to XP...
who knows.
still crunching.
RE: Update: here's the log
)
How many E@H do you have now? and LHC@H??
I run a p4 3.0 and it take about 7.78 hrs to complete E@H and about 2.86 for LHC@H and you say your running a P3 1.0. It may have something to do with speed and debt....
Message Type: Status Message
This Message means that the BOINC Client Software has determined that one of the following triggers is true:
Computer is over committed,
Less than 1 day until deadline, or
Deadline is before reconnect time.
Until all of these conditions are false Work Units will be processed in Earliest Deadline First order and will not switch.
The CPU Scheduler is running the Work Unit with the earliest Deadline instead of the Work Unit with the highest Short Term Debt.
Note:
This is the so-called "panic-mode".
[edit]Version Information
Link to Unofficial Wiki for BOINC, by Paul and Friends
What are your active
)
What are your active fractions for that computer? I'm thinking if it's anything below 50% (0.50) that is what is causing it to go into EDF so often.
BOINC WIKI
BOINCing since 2002/12/8
Fuzzy wrote: "How many E@H
)
Fuzzy wrote:
"How many E@H do you have now? and LHC@H??"
I have 1 E@H with about 2 hours left to complete, and one LHC with about 20 minutes to complete.
JKeck Wrote:
"What are your active fractions for that computer?"
What's an "active" fractioin? Each of the two projects is set at 50%.
have new message. Can't don't have the exact to copy again,
but the gist was this: When communicated with the Einstein server/BOINC scheduler or whatever, got a red message that said.............."computer is only running 13% of the time, and BOINC ~30% of that time......not enough to complete unit, entering EDF mode."
THis causing the problem would make sense, but the figure stating the PC is only running 13% of the time is incorrect. My PC is on all the time.......the BOINC figure is probably correct in that I have BOINC running about 30% of the time....but thats 30% of 24hrs.....or about 8 hours a day. Which is enough time to complete one EAH unit, and one LHC unit within the deadlines, by a wide margin.
Times........E@H units take me about 12hrs I think. and LHC...6 or 7 hours. per unit......rough estimate. Still should come out to enough time in my book.
I think BOINC thinking my PC is only on for 13% of the day..around 3 hours, and BOINC only getting 30% of that, is what's causing the problem.
Only thing is, is that my PC is on 100% of the day, and BOINC gets 30% of that time.......which is enough to complete work.
so, what's up with the wrong stats?
What's an "active" fractioin?
)
What's an "active" fractioin? Each of the two projects is set at 50%.
have new message. Can't don't have the exact to copy again,
but the gist was this: When communicated with the Einstein server/BOINC scheduler or whatever, got a red message that said.............."computer is only running 13% of the time, and BOINC ~30% of that time......not enough to complete unit, entering EDF mode."
That is the active fractions and it is what is causing EDF mode as well. It will correct itself over time.
BOINC WIKI
BOINCing since 2002/12/8
Ahh haa! C:-Program
)
Ahh haa!
C:-Program Files-BOINC-client_state
time_stats
on_frac --0.130946--on_frac
connected_frac--0.388443--connected_frac
active_frac--0.388424--active_frac
last_update--1120503867.936565--last_update
This what your looking for??
OK, that's got to be what it is, as you say......as long as all is normal, and will correct with time, I'm happy, and now know a little more about what's going on.
Thanks for the input JKeck and Fuzzy!
(typing in these forums is kinda weird, characters get all messed up.. used to the other....)
Multiply on_frac *
)
Multiply on_frac * active_frac and you have the % of time that BOINC believes that your computer is running BOINC programs. These numbers multiplied together giva an answer of about 0.04 giving BOINC the impression that your computer is spending almost no time running BOINC apps. This will slowly self correct over the next couple of months.
BOINC WIKI
RE: You may find a BOINC
)
yeah, i can confirm that.
- it consumes FAR TOO MUCH CPU time. looks like it updates every second which is absolutely unnecessary. every minute or so would be sufficient.
- it is possible to make the columns wider, but you can't swap columns or sort the data in some custom order
- when i select open boinc manager from the dock icon menu, the window opens behind all other windows. i have to click the icon to bring it forward.
- the window should appear when the dock iocon is clicked, whether the window is open behind other windows or not.
- the graphs under the statistics tab should all have the same horizontal scale, and that scale should be linear.
i'm sure there's more, but i should have been in bed an hour ago.
does some admin or developer ever read this thread?