LHC@home

As you can see from the following log for this afternoon, the same LHC WU was processed over and over again. When I watched it restart, I noticed that it processed for an hour (as per preference setting) then Boinc would switch to another project. When it came time to finish the approx. 1/2hr of the WU it would restart at 0% instead of the point at which it stopped. I had to suspend the other projects (I'm using Boinc v4.23) to allow the LHC WU to go to completion.
I've never seen this behavior with any of the other projects. I hope someone has an explanation and solution for this oddity.
For the moment, I've adjusted my preference to 1.5 hr between project changes, which should give the LHC WU enough time to be processed to completion in one turn.
-----------------------------------------
2/23/05 12:27:34 PM|LHC@home|Starting result v64lhc87-35s8_10515_1_sixvf_1021_0 using sixtrack version 4.64
2/23/05 1:14:11 PM|SETI@home|Restarting result 04ja05aa.19146.20801.611064.37_1 using setiathome version 4.09
2/23/05 1:14:11 PM|LHC@home|Pausing result v64lhc87-35s8_10515_1_sixvf_1021_0 (removed from memory)
2/23/05 1:14:41 PM||Suspending network activity - user request
2/23/05 2:14:18 PM|SETI@home|Pausing result 04ja05aa.19146.20801.611064.37_1 (removed from memory)
2/23/05 2:14:19 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 2:14:20 PM|climateprediction.net|Restarting result 1rfx_200103062_1 using hadsm3 version 4.04
2/23/05 2:14:20 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 2:14:22 PM|climateprediction.net|Pausing result 1rfx_200103062_1 (removed from memory)
2/23/05 2:14:22 PM|Einstein@Home|Restarting result H1_0209.4__0209.8_0.1_T16_Test02_1 using einstein version 4.79
2/23/05 2:14:33 PM|Einstein@Home|Pausing result H1_0209.4__0209.8_0.1_T16_Test02_1 (removed from memory)
2/23/05 2:14:34 PM|LHC@home|Restarting result v64lhc87-35s8_10515_1_sixvf_1021_0 using sixtrack version 4.64
2/23/05 2:14:35 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 2:14:35 PM|LHC@home|Pausing result v64lhc87-35s8_10515_1_sixvf_1021_0 (removed from memory)
2/23/05 2:14:35 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 2:14:36 PM|LHC@home|Restarting result v64lhc87-35s8_10515_1_sixvf_1021_0 using sixtrack version 4.64
2/23/05 2:14:37 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 2:14:37 PM|LHC@home|Pausing result v64lhc87-35s8_10515_1_sixvf_1021_0 (removed from memory)
2/23/05 2:14:39 PM|climateprediction.net|Restarting result 1rfx_200103062_1 using hadsm3 version 4.04
2/23/05 2:14:39 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 2:14:41 PM|climateprediction.net|Pausing result 1rfx_200103062_1 (removed from memory)
2/23/05 2:14:41 PM|LHC@home|Restarting result v64lhc87-35s8_10515_1_sixvf_1021_0 using sixtrack version 4.64
2/23/05 3:14:42 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 3:14:42 PM|LHC@home|Pausing result v64lhc87-35s8_10515_1_sixvf_1021_0 (removed from memory)
2/23/05 3:14:43 PM|climateprediction.net|Restarting result 1rfx_200103062_1 using hadsm3 version 4.04
2/23/05 3:14:43 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 3:14:44 PM|climateprediction.net|Pausing result 1rfx_200103062_1 (removed from memory)
2/23/05 3:14:44 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 3:14:46 PM|climateprediction.net|Restarting result 1rfx_200103062_1 using hadsm3 version 4.04
2/23/05 3:14:46 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 4:14:47 PM|climateprediction.net|Pausing result 1rfx_200103062_1 (removed from memory)
2/23/05 4:14:48 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 5:14:50 PM|Einstein@Home|Restarting result H1_0209.4__0209.8_0.1_T16_Test02_1 using einstein version 4.79
2/23/05 5:14:50 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 6:14:51 PM|SETI@home|Restarting result 04ja05aa.19146.20801.611064.37_1 using setiathome version 4.09
2/23/05 6:14:51 PM|Einstein@Home|Pausing result H1_0209.4__0209.8_0.1_T16_Test02_1 (removed from memory)
2/23/05 7:14:52 PM|SETI@home|Pausing result 04ja05aa.19146.20801.611064.37_1 (removed from memory)
2/23/05 7:14:53 PM|LHC@home|Restarting result v64lhc87-35s8_10515_1_sixvf_1021_0 using sixtrack version 4.64
2/23/05 8:14:54 PM|climateprediction.net|Restarting result 1rfx_200103062_1 using hadsm3 version 4.04
2/23/05 8:14:54 PM|LHC@home|Pausing result v64lhc87-35s8_10515_1_sixvf_1021_0 (removed from memory)
2/23/05 9:14:55 PM|climateprediction.net|Pausing result 1rfx_200103062_1 (removed from memory)
2/23/05 9:14:56 PM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/23/05 10:14:57 PM|Einstein@Home|Restarting result H1_0209.4__0209.8_0.1_T16_Test02_1 using einstein version 4.79
2/23/05 10:14:57 PM|ProteinPredictorAtHome|Pausing result t0212E_1_134741_1 (removed from memory)
2/23/05 11:14:57 PM|SETI@home|Restarting result 04ja05aa.19146.20801.611064.37_1 using setiathome version 4.09
2/23/05 11:14:57 PM|Einstein@Home|Pausing result H1_0209.4__0209.8_0.1_T16_Test02_1 (removed from memory)
2/23/05 11:55:48 PM|SETI@home|Pausing result 04ja05aa.19146.20801.611064.37_1 (removed from memory)
2/23/05 11:55:49 PM|LHC@home|Restarting result v64lhc87-35s8_10515_1_sixvf_1021_0 using sixtrack version 4.64
2/24/05 12:27:55 AM|ProteinPredictorAtHome|Restarting result t0212E_1_134741_1 using mfoldB125 version 4.22
2/24/05 12:27:55 AM|LHC@home|Pausing result v64lhc87-35s8_10515_1_sixvf_1021_0 (removed from memory)
-----------------------------------------

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

I noticed (CC 4.19, WinXP Pro, HT):

LHC@home - 2005-02-24 09:21:36 - Restarting result v64boince6ib1-13s4_6660_1_sixvf_604_7 using sixtrack version 4.63
Einstein@Home - 2005-02-24 09:21:36 - Restarting result H1_0231.9__0232.3_0.1_T15_Test02_7 using einstein version 4.79
ProteinPredictorAtHome - 2005-02-24 09:21:36 - Pausing result t0212E_1_128508_4 (removed from memory)
ProteinPredictorAtHome - 2005-02-24 09:21:36 - Pausing result t0212E_1_128681_2 (removed from memory)
LHC@home - 2005-02-24 09:21:37 - Pausing result v64boince6ib1-13s4_6660_1_sixvf_604_7 (removed from memory)
Einstein@Home - 2005-02-24 09:21:37 - Restarting result H1_0688.4__0688.6_0.1_T02_Test02_4 using einstein version 4.79

and in the work tab this WU has then wrong #,
CPUtime:'08:36:24' ; adv.time:00:00:00 ; to completion: 11:33:00 ; paused
(v64boince6ib1..-WU will take normally the predicted time: 11:33:00)
next time this appl. restarts, then 'adv.time' and 'to completion' will be shown right after a few sec.!

Everytime when projects apps. change, it starts one from another project and few millisec. this will pause. If there is a leak , then it will this do again and again...(like an 'infinity-loop')

It is not the LHC-WU, what make this problem. Next time, when apps. change, it will be another WU from another project with discribed behaviour.

Sorry it is verry hard to explain in english. I hope somebody can interprete my thinkings.

greetz littleBouncer

---

I've had more of the same overnight with a LHC WU that took a little longer than my switching time. The processing apparently restarted at 0% instead of its stopping point. This is not happening with any other projects.

Boinc CC 4.23
LHC sixtrack 4.64

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

Maybe you guys can try to change your preferences to keep the WU in memory instead of removing it from memory.
In your preferences put
" Leave applications in memory while preempted? " to Yes

I don't know if this helps since my LHC units is finished before it switches to another project but I know it speeds up the crunching.

Hope it works,

kind regards,
Krikke

---

> Maybe you guys can try to change your preferences to keep the WU in memory
> instead of removing it from memory.
> In your preferences put
> " Leave applications in memory while preempted? " to Yes
>
> I don't know if this helps since my LHC units is finished before it switches
> to another project but I know it speeds up the crunching.
>
> Hope it works,
>
> kind regards,
> Krikke
>
>
As I recall, this option has it's own problems. Sometimes a project doesn't release fully and 2 WU seem to run at the same time, making for extra long processing times. I have had this problem in the past.

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

> > Maybe you guys can try to change your preferences to keep the WU in
> memory
> > instead of removing it from memory.
> > In your preferences put
> > " Leave applications in memory while preempted? " to Yes
> >
> > I don't know if this helps since my LHC units is finished before it
> switches
> > to another project but I know it speeds up the crunching.
> >
> > Hope it works,
> >
> > kind regards,
> > Krikke
> >
> >
> As I recall, this option has it's own problems. Sometimes a project doesn't
> release fully and 2 WU seem to run at the same time, making for extra long
> processing times. I have had this problem in the past.
>
Result data is only saved at checkpoints. If your computer does not reach a checkpoint before the application is removed from memory because of a context switch, the WU will start crunching at the previous checkpoint, in this case, the beginning of the WU.

---

BOINC WIKI

Sometimes when switching between WU's it will restart at 0%, but after 15 secs it will jump back up to or close to where it was when paused.

---

> Sometimes when switching between WU's it will restart at 0%, but after 15 secs
> it will jump back up to or close to where it was when paused.
>
If you check the time log in the original post you will see that the WU ran for a full hour only to be restarted latter at 0%. This occurred several time.

@John McLeod
The WU were more than 2/3 complete when they were paused after one hour. Unless there are no checkpoint at all in the WU, it should not be reseting at 0%.

I have changed the option to keep the work in memory. Unfortunately my system did a reboot while I was at work today. I don't do an autorun on Boinc, so I don't have any new information on this problem.

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

Result data is only saved at checkpoints. If your computer does not reach a checkpoint before the application is removed from memory because of a context switch, the WU will start crunching at the previous checkpoint, in this case, the beginning of the WU.
==========

The problem I'm having is if I Exit the v4.19 GUI when I start it back up again the Time restarts at 0:00 again. It doesn't happen on all my Computers just some of them. It can happen after 2 hours of running also so I know it's reached a check point by then, at least I would think it had.

This problem doesn't seem to happen with the newer v4.20 & above just with the v4.19 Client. I just reset my Preferences to Keep in Memory to see if that helps, I thought I had it to do that but when I checked I didn't here at the LHC Site, so maybe that was the problem I don't know ...

---

Using CC v4.24

I set my Preferences to Keep in Memory last night and I woke up this morning with 3 projects ticking away in the CPU time column. This is an old known side effect of the Keep in Memory option. Apparently it still has not been resolved. I'm reverting to the 'NO' option.

One good thing ... Setting my switching time to 2 hr has allowed 1 LHC WU to run to completion. It actually only took 45 min for it process.

Luckily all this foolishness with LHC has had no ill effect on my other projects, which seem to be happily doing their thing.

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

> Using CC v4.24
>
> I set my Preferences to Keep in Memory last night and I woke up this morning
> with 3 projects ticking away in the CPU time column. This is an old known side
> effect of the Keep in Memory option. Apparently it still has not been
> resolved. I'm reverting to the 'NO' option.
>
> One good thing ... Setting my switching time to 2 hr has allowed 1 LHC WU to
> run to completion. It actually only took 45 min for it process.
>
> Luckily all this foolishness with LHC has had no ill effect on my other
> projects, which seem to be happily doing their thing.
>
Yup the time clock issue is a problem of the interaction of the Application (not Boinc) and Win98. If you continued to watch, and are crunching Einstein, then you'd notice that the clock worked well with Einstien, but not Seti. Removing from memory does fix the clock issue also. Seti(Rom) is aware that it's an Application issue.

tony

---

Formerly
mmciastro. Name and avatar changed for a change

The New Online Helpsytem help is just a call away.

I somewhat agree with what John Mcleod is saying. The 0 processor time showing after a pause resets to proper time once resume happens. Noooo this IS an LHC prob in some workunits I believe. Just found one resetting over and over deleted zip file and let LHC resend me the file 1st off as it seems some downloads may have been corruppted due to server overload. Instead of resetting anything I fooled my puters clock into running this unit w/o pause. Will see if it continues after new unit starts and I get rid of this 1. Only happenning on 1 of 4 machines. I do have save in memory enabled but since I downloaded BOINC 4.19 it always gives message removed from memory but workunits always ALWAYS restart properly after pause til this LHC work unit. Am running up to 4 projects each on 4 hosts I also think Aurora that this thread should be renamed to something like ahh workunits resetting to 0 after pause to get admins attention :) ahhh thus far after deleting zip and LHC resending the time to competion has come way down to puters "norm" was adding approx 1hr or 25% to last zipfile

---

@jrenkar
The problem definitely has to be within the LHC WU or application since it doesn't exist in any other projects.
I've processed several WU today by setting the project change time at 2hr. This guarantees that the WU are processed to completion every time.

I would change the thread name but I don't know how. Edits have to be done within 60 min. of posting.

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

> @jrenkar
> The problem definitely has to be within the LHC WU or application since it
> doesn't exist in any other projects.
> I've processed several WU today by setting the project change time at 2hr.
> This guarantees that the WU are processed to completion every time.
>
> I would change the thread name but I don't know how. Edits have to be done
> within 60 min. of posting.
Yes I agree Aurora have increased time also seems best fix for problem at the moment.Got all mine through too ... strange how this occured with many at same time when 1st batch of ver4.64 didn't do this... Happy crunching... :)

---

See also my experience described here.

---

Best greetings from Belgium
Thierry

I'm returning to a 60 min. change time.
My latest LHC file, v64boince6ib1-51s8_10615_1_sixvf_12204_2, is a long process (est. at 12 hr) and has restarted properly once. I can only hope that it contains plenty of checkpoints.

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

Yes Aurora Borealis, I think you are on to something.
I have also noticed that check pointing in LHC wu is not working.
As a workaround, I have set to leave apps in memory and this works ok ‘til exiting BOINC or reboot.
Today as a test I exited BOINC while running LHC wu at 76% progress.
Restart BOINC and the wu keeps CPU time but progress begins at 0%.

Win98. BOINC 4.24

---

> Yes Aurora Borealis, I think you are on to something.
> I have also noticed that check pointing in LHC wu is not working.
> As a workaround, I have set to leave apps in memory and this works ok ‘til
> exiting BOINC or reboot.
> Today as a test I exited BOINC while running LHC wu at 76% progress.
> Restart BOINC and the wu keeps CPU time but progress begins at 0%.
>
> Win98. BOINC 4.24
>

Unfortunately Boinc doesn't always play nice with Win98. When apps are left in memory the paused apps CPU time will continue ticking away and the active app runs more slowly. As mentioned earlier in this thread, this is a know bug. It is not a good idea to use this option with Win98.

It appears that the problem is with WU labeled v64lhc. They are short run WU. On my system slightly over or under 1 hr. They don't appear to have any checkpoint, so they restart at 0% after being paused. Yesterday, I had a WU labled v64boince, it ran for approx. 6hr, and restarted properly after beeing paused.

The following is an extract of the LHC WU for this afternoon. The first WU was very short time and completed properly. The second was just over 1 hr. and kept restarting at 0%. I'm going to compromise and increase my switch time to 90 min. to process the 6 WU currently on my system.

2/28/05 1:41:50 PM|LHC@home|Starting result v64lhc88-55s16_18515_1_sixvf_3582_2 using sixtrack version 4.64
2/28/05 1:42:49 PM|LHC@home|Computation for result v64lhc88-55s16_18515_1_sixvf_3582 finished
2/28/05 1:42:49 PM|LHC@home|Starting result v64lhc92-54s10_12575_1_sixvf_3797_0 using sixtrack version 4.64
2/28/05 2:42:50 PM|LHC@home|Pausing result v64lhc92-54s10_12575_1_sixvf_3797_0 (removed from memory)
2/28/05 6:42:55 PM|LHC@home|Restarting result v64lhc92-54s10_12575_1_sixvf_3797_0 using sixtrack version 4.64
2/28/05 7:42:58 PM|LHC@home|Pausing result v64lhc92-54s10_12575_1_sixvf_3797_0 (removed from memory)

Boinc 4.24

---

Questions? Answers are in the BOINC Wiki.

Boinc V6.10.56 Recommended
WinXP C2D 2.1G 3GB

> Yup the time clock issue is a problem of the interaction of the Application
> (not Boinc) and Win98. If you continued to watch, and are crunching Einstein,
> then you'd notice that the clock worked well with Einstien, but not Seti.
> Removing from memory does fix the clock issue also. Seti(Rom) is aware that
> it's an Application issue.

Tony,
Thanx for the info... I didn't know if was the OS or old/slow machine as I often see this on my old Win98se/266Mhz/128MB machine but I don't recall seeing it on the only slightly faster win2K/400Mhz/256MB or any of the newer/faster WinXP machines. I assume/hope that if I set up some separate preferences for (home/school or work) and assign that computer to those preferences that I set it to not leave in memory and have all the other machines leave apps in memory. Has anyone done this successfully?

---

- da shu @ HeliOS,
"Free software” is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech”, not as in “free beer".

> > Yup the time clock issue is a problem of the interaction of the
> Application
> > (not Boinc) and Win98. If you continued to watch, and are crunching
> Einstein,
> > then you'd notice that the clock worked well with Einstien, but not Seti.
>
> > Removing from memory does fix the clock issue also. Seti(Rom) is aware
> that
> > it's an Application issue.
>
> Tony,
> Thanx for the info... I didn't know if was the OS or old/slow machine as I
> often see this on my old Win98se/266Mhz/128MB machine but I don't recall
> seeing it on the only slightly faster win2K/400Mhz/256MB or any of the
> newer/faster WinXP machines. I assume/hope that if I set up some separate
> preferences for (home/school or work) and assign that computer to those
> preferences that I set it to not leave in memory and have all the other
> machines leave apps in memory. Has anyone done this successfully?

Update: I set up "work" venue and have this working now. No biggie.

---

- da shu @ HeliOS,
"Free software” is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech”, not as in “free beer".