some action over at LHC@Home Classic (SixTrack)

[Sunny129]

bumping this old thread to the top b/c work is available again for LHC@Home Classic (SixTrack).

i'm a little late on this notice, so i apologize in advance - i was well aware of the new influx of work to be crunched, but i had completely forgotten that i had started this "awareness/notifications" thread here for everyone until now.

so here's the scoop - only two days ago more LHC@Home Classic (SixTrack) work became available, and the number of tasks ready to send shot up to over 200,000. the infusion ceased, and the number of tasks ready to send steadily went back down to ~75,000 over the next ~36 hours, but then i discovered this morning that the number is climbing again. right now its up to 93,410 tasks ready to send.

for those who don't know, LHC@Home Classic is a project that is highly dependent on the goings on at CERN - sometimes they need massive amounts of compute power, and other times they don't need any. as such, it is a very spotty, intermittent, and irregular project for users and their hosts. so if you want to contibute to the project, now would be the time while there's still work available.

 

[Fardringle]

I have a couple of computers that frequently request work from LHC (while also working on other projects) but none of them have received any new work in the past few days, including a manual request I just sent through the BOINC Manager.

 

[Sunny129]

I have a couple of computers that frequently request work from LHC (while also working on other projects) but none of them have received any new work in the past few days, including a manual request I just sent through the BOINC Manager.

what happens if you temporarily set all of your other CPU projects to NNT (no new tasks) and increase your minimum work buffer by a day or two? i think you might be able to trick it into getting some LHC work that way...it works for me all the time. after you've received a day's worth of LHC work, you can decrease your work buffer again and re-enable work fetch on your other projects.

 

[Fardringle]

I tried that. The project says there is no work available even though the server status on the project site says that there is.

 

[Sunny129]

try resetting the project, and if that doesn't work, try detaching and reattaching to the project.

 

[zzuupp]

Okey Doke.

I've set one on LHC. It's got some asteroids to finish first.

 

[GLeeM]

The project says there is no work available even though the server status on the project site says that there is.

What version of BOINC are you running?
I get the same sometimes with 7.0.44.

 

[Bradtech519]

Looking at 7 hours 5 minutes per task. I'm in 100% though and will break from asteroids.

 

[Sunny129]

Looking at 7 hours 5 minutes per task. I'm in 100% though and will break from asteroids.

you'll soon find that there really is no way to accurately gauge what your LHC task run times should be. you may have to stay on the project for a day or two before you see it, but you'll notice that some tasks run for minutes, some for hours, and some for only seconds. from what i understand, the data we are crunching simulate the particles (streams of protons) traveling around the LHC at CERN. the slightest misalignment in the electromagnets that line the walls of the LHC tunnel and guide the beam of protons might only cause the slightest deviation in the course of the beam, but when they're taveling at 99.999999c (where c is the speed of light), they circle the LHC over 10,000 times every second. so if the electromagnets are not perfectly aligned, and it causes the stream of protons to deviate 1mm off course per revolution around the LHC, it only takes a single second for the beam to deviate 10m off center. i'm not sure how far it is from the center of the tunnel to its walls, but i do know that short-running tasks are simulations with highly unsteady beams (ones that veer off-center substantially and thus slam into the walls of the LHC tunnel shortly after the simulation begins). obviously very long running tasks simulate highly steady beams of protons.

 

[Bradtech519]

Yeah I just noticed I finished a task already. I've wanted to help with LHC for a while. Its the top of the list for me. Proving/disprovinf string theory would be a big break through on unification in physics. I'll keep it going as long as they feed me tasks. They have the other project also that runs in Oracle virtual box.

 

[Fardringle]

What version of BOINC are you running?
I get the same sometimes with 7.0.44.

7.0.28 on Windows XP Pro (32 bit)

I'll keep trying to snag some work for it but I haven't had any luck so far. Completely disconnecting and reconnecting to the project didn't change anything.

 

[Sunny129]

They have the other project also that runs in Oracle virtual box.

yeah, apparently i'm the only TeAm member crunching Test4Theory@Home currently...

7.0.28 on Windows XP Pro (32 bit)

I'll keep trying to snag some work for it but I haven't had any luck so far. Completely disconnecting and reconnecting to the project didn't change anything.

hmm...i'm stumped. i wonder if its a combination of the BOINC version and the OS you're running...and if that configuration works for several other projects on that host, then it might not be worth it to change anything just to see if you can get LHC@Home to run on it. on the plus side, you have time to keep trying, as there are now almost 200K tasks ready to send.

 

[Bradtech519]

I didn't realize they had any work now for it. I already have VMware workstation installed and running. I wonder if that will work in lieu of virtualbox.

yeah, apparently i'm the only TeAm member crunching Test4Theory@Home currently...


hmm...i'm stumped. i wonder if its a combination of the BOINC version and the OS you're running...and if that configuration works for several other projects on that host, then it might not be worth it to change anything just to see if you can get LHC@Home to run on it. on the plus side, you have time to keep trying, as there are now almost 200K tasks ready to send.

 

[Sunny129]

I didn't realize they had any work now for it. I already have VMware workstation installed and running. I wonder if that will work in lieu of virtualbox.

yeah, they've had work regularly for quite some time now. i have no idea if it'll work w/ VMware, as it was designed to be used with VirtualBox. i don't know what OS you're running or how much RAM you have available (and thus i don't know if you have room for another virtual machine on that host), but Test4Theory@Home was designed specifically to work with VirtualBox. installing VBox is pretty simple and straightforward too.

 

[Bradtech519]

I may test it out tomorrow. I've ran the sixtrack all day today.

 

[Bradtech519]

Just signed up and started task. Downloaded virtual box. Looks it downloaded a 250MB vmdk. Then started running.

 

[Bradtech519]

Impressed with how the 4theory works. See the vm self created itself in virtualbox. 22 hour task looks like. Got six other cores doing sixtrack and one core doing asteroids.

 

[Sunny129]

Impressed with how the 4theory works. See the vm self created itself in virtualbox. 22 hour task looks like. Got six other cores doing sixtrack and one core doing asteroids.

it'll take approx. 24 hours, give or take a few seconds, to complete 1 T4T@H task. your BOINC "estimated time remaining" is off by a few hours right now probably b/c this is your first T4T task. once you've run a handful of T4T tasks, and the BOINC DCF (duration correction factor) for this project settles down, your "estimated time remaining" for an unstarted T4T task should be approx. 24 hours. that's just the design of the project.

what's really happening is that the BOINC_VM (in VirtualBox) is sending and receiving data directly to and from CERN. once BOINC recognizes that this process has been going on for 24 hours, it terminates the task in BOINC, which in turn signals VirtualBox to terminate the BOINC_VM. its a bit of an unsophisticated technique in that the termination of the T4T task in BOINC (and therefore the termination of the BOINC_VM in VirtualBox) might occur while the BOINC_VM is in the process of crunching a CERN task. nevertheless, the vast majority of CERN tasks that your BOINC_VM processes over the duration of the 24-hour BOINC task to get completed and reported to CERN...so wasted CPU cycles is kept to a minimum in that respect.

 

[Fardringle]

yeah, apparently i'm the only TeAm member crunching Test4Theory@Home currently...

I ran Test4Theory for a while but the lack of checkpoints makes it pretty much impossible for me since I frequently use my computer for other things (and interrupt BOINC) so projects without checkpoints don't work very well.

hmm...i'm stumped. i wonder if its a combination of the BOINC version and the OS you're running...and if that configuration works for several other projects on that host, then it might not be worth it to change anything just to see if you can get LHC@Home to run on it. on the plus side, you have time to keep trying, as there are now almost 200K tasks ready to send.

It does work for pretty much every other project, so it's probably a compatibility problem either with my BOINC version or with Windows XP. Either way, while I like the idea of the LHC project, as you said, it's not worth messing everything else up trying to get it to work.

 

[Bradtech519]

I will probably just leave the one CPU task running and suspend other projects when I play a game.

 

[Sunny129]

a link to a brief but interesting 1-minute video was posted today in the LHC@Home Science sub-forum, so i thought i would share that link. it is a semi-timelapse video showing the relocation and positioning of one of the LHC's 1232 35-ton dipole magnets. i can only imagine the power of these magnets while the LHC is operating, as they are responsible for bending the path of a stream of protons (that carries the momentum of a freight train when traveling near the speed of light) around a circle only 17 miles in circumference! for reference, a single proton traveling at near c has approx. the same momentum as a mosquito in flight...that should give you an idea of just how many protons comprise this "small stream" of subatomic particles...

video: http://www.youtube.com/watch?v=BLXB3CGkOZ4&feature=youtu.be

 

[Sunny129]

hay Fardringle, i noticed that your host finally started crunching LHC@Home (i noticed via BOINCstats that you earned 784 points crunching LHC@Home yesterday)! what exactly did you have to do to get things working again?

 

[Bradtech519]

I wanna hit 500k in asteroids to get the badge. So I've paused the classic six track and dedicated seven cores to Asteroids and got one core going for test4theory. Once I hit 500k I'll jump back over to sixtrack.

 

[Fardringle]

hay Fardringle, i noticed that your host finally started crunching LHC@Home (i noticed via BOINCstats that you earned 784 points crunching LHC@Home yesterday)! what exactly did you have to do to get things working again?

Yeah, I finally got some work. I didn't change anything on my computer so I don't know what happened. Unfortunately, I don't think I'll keep running it much longer. The LHC app somehow manages to get my CPU up to 62 Celsius (which it really doesn't like) while other projects never push it any higher than 54 or 55. I'm not too worried about the temperature itself, but at 60+, the cooling fans run at max speed and they are really loud and annoying.

 

[Sunny129]

its been relatively quiet on the LHC@HOme front for the last month or so, with a few thousand WUs being distributed at a time every so often. well i never expected to see it on a Holiday like this, but i woke up to ~10,000 new WUs "ready to send" on the server status page this morning, which is one of the larger infusions i've seen this month. then again, maybe its not such a surprise considering they don't celebrate Thanksgiving over at CERN and that this is probably just another work day over there...i just always assumed that the project servers/admins/developers/etc. that linked us individual hosts w/ CERN was located stateside, but i suppose i may be wrong about that.

anyways, since i woke this morning, the number of "tasks in progress" has climbed substantially, and the number of "tasks ready to send" has gone up to ~12,000, which means that currently they must be producing work at a greater rate than it is being distributed and crunched, if only temporarily...so grab some LHC SixTrack WUs while they available!

...and may your hosts have a very productive Thanksgiving!