Message boards :
Number crunching :
Sixtrack 4.67 available
Message board moderation
Author | Message |
---|---|
Send message Joined: 27 Sep 04 Posts: 282 Credit: 1,415,417 RAC: 0 |
With workunits! ;-) go grab them before they're gone |
Send message Joined: 2 Sep 04 Posts: 39 Credit: 4,910 RAC: 0 |
actually downloading, thx :) |
Send message Joined: 27 Sep 04 Posts: 282 Credit: 1,415,417 RAC: 0 |
> With workunits! ;-) go grab them before they're gone > they're gone now! ;-) |
Send message Joined: 1 Sep 04 Posts: 157 Credit: 82,604 RAC: 0 |
First time I notice this: the report deadline is 14 days minus 2 hours 40 minutes. Was this not 14 days round? I have 8 WU's that have been downloaded at 10:53:39 UTC, which is the correct time, but the deadline is stated on the website as 8:13:39 UTC for all of them. The reported deadline in Boinc Manager, v4.30, is stated as the same time as on the website. |
Send message Joined: 17 Sep 04 Posts: 1 Credit: 1,321,527 RAC: 0 |
> First time I notice this: the report deadline is 14 days minus 2 hours 40 > minutes. > Was this not 14 days round? > Little calculation: 1209600 seconds rounded to 1200000! |
Send message Joined: 1 Sep 04 Posts: 157 Credit: 82,604 RAC: 0 |
> Little calculation: > 1209600 seconds rounded to 1200000! Don' know if this could be the answer alariz. At S@H the deadline is exactly 14 days and at E@H it is exactly 7 days. |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
Why there is a difference in use of memory from the same application 4.67? See the picture bellow the sixtrack 4.67 once at 51'572 KB and once at 12'072 KB: I ask this because the second one I crunch now the 3th. time on the same host, once with 4.66 at alpha success, once here with 4.64 failed with 0 CPU-time, and now with 4.67: it hasn't finished yet. But I noticed this difference in using the memory. greetz littleBouncer |
Send message Joined: 2 Sep 04 Posts: 309 Credit: 715,258 RAC: 0 |
On my HT machine 4.67 is using 46,712kB and 46,732kB. |
Send message Joined: 1 Sep 04 Posts: 157 Credit: 82,604 RAC: 0 |
> Why there is a difference in use of memory from the same application 4.67? > > See the picture bellow the sixtrack 4.67 once at 51'572 KB and once at 12'072 > KB: > I ask this because the second one I crunch now the 3th. time on the same host, > once with 4.66 at alpha success, once here with 4.64 failed with 0 CPU-time, > and now with 4.67: it hasn't finished yet. But I noticed this difference in > using the memory. Hello littleBouncer, this had to do with the use of the graphics or screensaver AFAIK. Once one of them has been used, the used RAM drops down to some 10.9xxkB on my host but the peak consumption is shown as some 45.9xxkB in the taskmanager. sixtrack v4.67 |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
> Hello littleBouncer, > > this had to do with the use of the graphics or screensaver AFAIK. Once one of > them has been used, the used RAM drops down to some 10.9xxkB on my host but > the peak consumption is shown as some 45.9xxkB in the taskmanager. > > sixtrack v4.67 > > THX Thierry I didn't think at that(/this) . ( I'm not using the scrennsaver, but from time to time 'show graphic'. I saw it before and when 4.64 WU's began to exit with 0 CPU-time. That's why I was a little upset; first to see the same WU which previously failed, and second with this difference... Meanwhile this WU has finished with success (with 4.67).) greetz littleBouncer |
Send message Joined: 27 Sep 04 Posts: 13 Credit: 155,619 RAC: 0 |
>Meanwhile this WU has finished with success (with 4.67).) how can i re-run a 4.64 wu with 4.67 ? |
Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
I noticed that my name is moving about the top of the screensaver. Is this new? Or because i pressed a key, like F7? I'm not the LHC Alex. Just a number cruncher like everyone else here. |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
> >Meanwhile this WU has finished with success (with 4.67).) > > how can i re-run a 4.64 wu with 4.67 ? > You can do nothing . That is by send out the work from LHC; the last times they were always the same, first with 4.63, then 4.64 .... I noticed that first by the graphic. I was wondering me , that I see to often the same 'beginn' (on 'show graphic'). Then I discovered that every Simulation has at least 8 different (how Chrulle discribed) amplitude (s4_ - s18_). For example: v64boince6ib1-1s6_8660_1_sixvf_92_1 with 4.66 success on Alpha v64boince6ib1-19s6_8660_1_sixvf_812_0 with 4.64 failed v64boince6ib1-3s6_8660_1_sixvf_2542_2 with 4.67 success This is one, and the same WU was crunched with different executables on the same host the past week. greetz littleBouncer |
Send message Joined: 27 Sep 04 Posts: 13 Credit: 155,619 RAC: 0 |
> For example: > v64boince6ib1-1s6_8660_1_sixvf_92_1 with 4.66 success on Alpha > v64boince6ib1-19s6_8660_1_sixvf_812_0 with 4.64 failed > v64boince6ib1-3s6_8660_1_sixvf_2542_2 with 4.67 success > Jä so meinsch Du das... Oki, Now i understand, you don't meen the sam WU, you meen the same Units, but with different WorkParameters, and send from LHC self :-) I only ask, because i had cancel many 4.64 Units via textedito (Client 4.19). And i woundering there is a way to "save the unit" and edit the current_state.xml so that the unit is a 4.66 oder 4.67 Unit.. [workunit] [name]v64boince6ib1b-29s8_10615_1_sixvf_28345[/name] [app_name]sixtrack[/app_name] [version_num>466[/version_num] .... .... [/workunit] I ask because the version line ... :-) No, i never edit this, i don't cheat.. i only was woundering its possible and want ask, if so.. :-) |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
> > For example: > > v64boince6ib1-1s6_8660_1_sixvf_92_1 with 4.66 success on Alpha > > v64boince6ib1-19s6_8660_1_sixvf_812_0 with 4.64 failed > > v64boince6ib1-3s6_8660_1_sixvf_2542_2 with 4.67 success > you meen the same Units, but > with different WorkParameters, and send from LHC self :-) ----- @ bluumi Just to clearify: This example same Unit with the same parameters (s6_8660_1) but with different executables (sixtrack_4.6x_windows_intelx86). The other numbers are to catalog the unit (which run, which result, which place on sendout etc..) schöns tägly no greetz littleBouncer I will look into the results-table from 'boinclogx' to compare the times, and will make some notes here ( if I get in 60 minutes, then I update this post). [update] 4.67 : 11:18:49 4.64 : 00:00:00 (date: 4.10.05) 4.66 : 10:59:21 I noted that I crunched the '0-one' on 3.28.05 with 4.64 and success, maybe that's why it failed the second time?! 4.64 : 11:13:05 (date: 3.28.05) [/update] |
Send message Joined: 17 Sep 04 Posts: 69 Credit: 26,714 RAC: 0 |
Why me,,,,oh, why me. on the 9th, four days ago, I finally got to download some WUs (8). Like normal I didn't watch them, and just let them do there things. three days later (yesterday) I notice I still have 8 of them. 7 ready to run and one running. These were Boinc 4.64 units on boinc 4.27. The one that was running was only at 02:26:15 CPU time. I kept an eye on it and It'd reset to zero each time it restarted. I have my switch between projects set to 120 min. I ended up Aborting them all. Yesterday I downloaded eight more WUs, this time they are 4.67. I've been sort of watching them yesterday. This morning 0642 EST I noticed the first one had 02:27:03 CPU time and 12:01:33 to completion. I "Suspended" all other projects in Boinc CC 4.30. It's now 8:22 am EST (1 hr 40 min elapsed) and the CPU time says "02:28:42" and the "to comp" says 11:56:54. The other projects crunch just fine. What's going on?????? Is the clock "Days/hours/Min", instead of the regular? tony Formerly mmciastro. Name and avatar changed for a change The New Online Helpsytem help is just a call away. |
Send message Joined: 22 Oct 04 Posts: 6 Credit: 3,413,714 RAC: 0 |
From the looks of it, sixtrack does not do any periodic saves. Therefore whenever you unload the application from memory, all the work done is "forgotten". The other BOINC applications, at least the other 3 I subscribe to, are doing periodic saves. There is some loss of work, but at least itr is not 1 or 2 hours but just a matter of minutes. Hopefully, the save feature will be added to the application in the very near future. [img src="http://www.boincstats.com/signature/user_10138.gif"] |
Send message Joined: 17 Sep 04 Posts: 69 Credit: 26,714 RAC: 0 |
> From the looks of it, sixtrack does not do any periodic saves. Therefore > whenever you unload the application from memory, all the work done is > "forgotten". so basically, we can either run Lhc all by itself or NOT run Lhc at all. Why are they even a Boinc project if they don't want to share with others (cpu time). sorry, I know they're working on it, sometimes it gets so frustrating. Wanting to help a project, but not being able to. This still doesn't account for the 1 hour and 40 min of run time being displayed as 1 minute and change. Now CPU time is up to 02:28:58. it's advanced 16 seconds in the last 25 minutes. This specific WU is Vboince6ib1-15s4-6645-1 6vf-2803_2 thanks for your reply |
Send message Joined: 1 Sep 04 Posts: 139 Credit: 2,579 RAC: 0 |
Let's just clarify this point again: Sixtrack DOES do checkpointing ("periodic saves") and always has done, BUT there was a bug in version 4.64 which made this unreliable, leading to the bad effects you are describing. Later Sixtrack versions (4.65, 4.66, 4.67 ... ) have corrected this checkpointing bug. Users should stop using 4.64 as soon as possible. (You may have to reset your BOINC client to trigger the update of your Sixtrack version). Ben Segal / LHC@home |
Send message Joined: 18 Sep 04 Posts: 59 Credit: 317,857 RAC: 0 |
> Let's just clarify this point again: Sixtrack DOES do checkpointing ("periodic > saves") and always has done, BUT there was a bug in version 4.64 which made > this unreliable, leading to the bad effects you are describing. > > Later Sixtrack versions (4.65, 4.66, 4.67 ... ) have corrected this > checkpointing bug. > > Users should stop using 4.64 as soon as possible. (You may have to reset your > BOINC client to trigger the update of your Sixtrack version). > > Ben Segal / LHC@home > No need to reset the project. Install Boinc V4.2x and you can kill individual work unit. Questions? Answers are in the BOINC Wiki. Boinc V6.10.56 Recommended WinXP C2D 2.1G 3GB |
©2024 CERN