|
1)
Message boards :
Theory Application :
MadGraph5
(Message 50269)
Posted 28 May 2024 by Henry Nebrensky Post: Thanks... I wasn't sure if there was some clean-up check that would catch it, apart from the 10-day mop-up. I should have said earlier: upshot was that the task was taking up a slot but using zero CPU - hence my dredging up an old thread about madgraph doing exactly that. |
|
2)
Message boards :
Theory Application :
MadGraph5
(Message 50267)
Posted 28 May 2024 by Henry Nebrensky Post: Just aborted task 411417290 as runRivet.log showsIt also has significant stretches of not actually using CPU at all.e.g I recently killed task 281349801 precisely because it was holding two cores but idle - it's reported as using just 50 mins in 20 hours :( ===> [runRivet] Tue May 28 13:26:22 UTC 2024 [boinc pp zinclusive 13000 - - madgraph5amc 2.6.0.atlas nlo 100000 160]and then INFO: Result for check_poles: I'd have left it running to see what happens, but the machine will get powered off soon. Looks like internal fault detection isn't being passed up the chain and stopping the task, or did I just get too impatient? Edit: upshot was that the task was taking up a slot but using zero CPU - hence my revisiting an old thread about madgraph doing exactly that. |
|
3)
Message boards :
LHC@home Science :
Collision energies
(Message 49951)
Posted 16 Apr 2024 by Henry Nebrensky Post: Thanks for that - I'll have to read it through properly. It's nice to see LHC experiments other than the big four get a mention. For reference., p21 answers a similar question I think: Lead-Lead has a total of 1150 TeV available but there is actually only 5.02 TeV available to the colliding pair of nucleons within; possibly I'm misunderstanding the physics but shouldn't there be a similar downshift between the overall proton(nucleon) energy and that of the actual colliding constituents? |
|
4)
Message boards :
LHC@home Science :
Collision energies
(Message 49898)
Posted 6 Apr 2024 by Henry Nebrensky Post: Hi all, Does anybody here know roughly what the maximum "physics energy" is for an LHC proton-proton collision? Since protons are composite particles the collisions are actually between individual quarks/gluons/whatever, and I'd like to get an idea of the "inefficiency" compared to the beam proton energy. I couldn't see this in a quick look at Wikipedia, probably because I'm not sure what the technical terms would be! Thanks |
|
5)
Message boards :
Cafe LHC :
Eviction of Russian scientists
(Message 48897)
Posted 4 Nov 2023 by Henry Nebrensky Post: (Who's "we"?) That's odd considering we welcome doctors from India/Pakistan.Only because they accept NHS pay rates... So scientists aren't considered important?No - (foreign) scientists are just as much seen as cheap exploitable disposable labour. That's why the UK Government's R&D people and culture strategy avoids any commitment to retaining experienced staff or supporting returns after a career break, but when it comes to getting in naive junior staff there's bullet points a-plenty! |
|
6)
Message boards :
Cafe LHC :
CMS: I'm (almost) back
(Message 48896)
Posted 4 Nov 2023 by Henry Nebrensky Post: Hi Ivan, glad to hear you've made it home at last... I've been trying to arrange another trip into London to visit en route, but not been able to stitch (geddit!) all the various stop-offs needed together :( Focus on your recovery! |
|
7)
Message boards :
Theory Application :
How long may Native-Theory-Tasks run
(Message 48154)
Posted 30 May 2023 by Henry Nebrensky Post: Theory native logs can be checked, e.g for a task running in slot 0 If you "head" the runRivet.log, it will tell you the code in use and how many events that specific task is to generate: [boinc pp jets 8000 170,-,2960 - pythia8 8.301 dire-default 57000 482] 57k in this case. If you then "tail" the log you can see how far it's got and if it's making progress... |
|
8)
Message boards :
LHC@home Science :
Ready to restart!
(Message 47245)
Posted 12 Sep 2022 by Henry Nebrensky Post: It looks like they are indeed training the magnets these days, there are magnet quenches with no beam.Possibly, but given the potential side-effects of a quench - remember those magnets scattered around the tunnel after the first startup - I know I would be wary of trying to train to 14TeV current settings before getting 13TeV data in the can. I'm intrigued by the cryptic references to UFOs in that blog post - sounds suspiciously like someone's dropping their spanners on the machine. (Unless that's a dated baguette reference?) |
|
9)
Message boards :
LHC@home Science :
Ready to restart!
(Message 47195)
Posted 27 Aug 2022 by Henry Nebrensky Post: Meanwhile the LHC magnets remain at 2K, allowing more time to train the magnets.What does that involve? |
|
10)
Message boards :
LHC@home Science :
Proton Physics : Stable Beams
(Message 47054)
Posted 30 Jul 2022 by Henry Nebrensky Post: Why does the luminosity at an experiment's IP vary relative to the circulating beam? (e.g. what happened at Alice before/after 20:00-ish?) Does this mean that the LHC dynamically alters the lattice even within a fill? |
|
11)
Message boards :
Theory Application :
Sherpa tasks run okay for long time, then they fail
(Message 47024)
Posted 15 Jul 2022 by Henry Nebrensky Post: 7.100 events (8d 17h) atm.But isn't that going to take more like 14 days to complete, rather than 10? |
|
12)
Message boards :
Number crunching :
CentOS9 squid Win11pro
(Message 46857)
Posted 8 Jun 2022 by Henry Nebrensky Post: CentOS8 have no Mirror, out of support since 2022/01.As announced a couple of years ago, CentOS became "CentOS Stream". Conversion instructions for CentOS8 to CentOS Stream8 are at https://www.centos.org/centos-stream/. (I've not tried them; I've been reverting to C7 instead). |
|
13)
Questions and Answers :
Getting started :
How To Setup VirtualBox With LHC@HOME On Command Line Only Server?
(Message 46645)
Posted 15 Apr 2022 by Henry Nebrensky Post: I got some cloud VMs and want to contribute to the project, but the native applications always fail, and I couldn't fix it. I am wondering how can I setup VirtualBox with LHC.As these are already VMs, I doubt Virtualbox will run correctly inside them. Never tried it though. I did simply `sudo apt install virtualbox -y`, is that all? Answering the subject line, my old notes for CentOS7/8 (Minimal install) say:
yum install gcc make elfutils-libelf-devel kernel-devel kernel-headers setools-libs libsemanage-python libseccomp Install the downloaded VirtualBox RPM: yum localinstall VBox-version.RPM user running BOINCmust be in VBox group: usermod --append --groups vboxusers boinc Rebuild the kernel modules: vboxconfig Start VBox services: service vboxautostart-service start service vboxdrv restart Optional: VBoxManage extpack install /tmp/Oracle_VM_VirtualBox_Extension_Pack-version.vbox-extpack
VBoxManage list extpacks VBoxManage list vms VBoxManage showvminfo 1e6cf289-3176-476f-b80c-6fa3fbe42922 VBoxManage unregistervm 1e6cf289-3176-476f-b80c-6fa3fbe42922 --delete Distro-specific and SysV-to-systemd translation left as an exercise for the reader. Note that I've not tested this in a couple of years as I don't have any suitable machine at the moment. Any problems see Yeti's Checklist |
|
14)
Message boards :
Number crunching :
CentOS9
(Message 46488)
Posted 20 Mar 2022 by Henry Nebrensky Post: Mirror for CentOS8 is out of Service since 22/01/31.Isn't that part of the change from CentOS to CentOS Stream? This has caused some "discussion" in the community, including proposals for LHC to move away from CentOS to some other distro. I don't know if there's any clear route on this yet. CVMFS is not supported for CentOS9 (only CentOS7 and CentOS8).May well be low-priority due to above. |
|
15)
Message boards :
ATLAS application :
Atlas task slowing right down near the end but still using all cores - continue?
(Message 46414)
Posted 3 Mar 2022 by Henry Nebrensky Post: ... and angles, (which is where clock faces came from?).Are they the ones responsible for clocks?I would vote for 240.Actually, I'd vote for 360 - them Babylonians knew what they were doing ... But, actually computezrmle was right: 240 also gets you a division by 16, for future expansion. |
|
16)
Message boards :
ATLAS application :
Atlas task slowing right down near the end but still using all cores - continue?
(Message 46403)
Posted 2 Mar 2022 by Henry Nebrensky Post: I would vote for 240.Actually, I'd vote for 360 - them Babylonians knew what they were doing - but if people are already struggling with compute times then it would be better to stick to low hanging fruit. :( |
|
17)
Message boards :
ATLAS application :
Atlas task slowing right down near the end but still using all cores - continue?
(Message 46402)
Posted 2 Mar 2022 by Henry Nebrensky Post: Doesn't that depend on the variance, which I've never studied? In any case, the aim is that the averaging overcomes the variance, which is why divisibility is important for avoiding a small number of events left over.Even within a task, as the number of threads is reduced then each thread must run more events, and it's more likely - but not guaranteed - that they will average out across the threads.Even with the full 8 cores, that's 25 events per thread, which is more than enough to average things out. Thank you - I did put some effort into setting those machines up...A few minutes in 4 hours is nothing.Also, do you have a figure for how much time is wasted? Since it's 200 in the pool, the wasted cores at the end are probably a fraction of a percent of inefficiency.IIRC, back when I was running 8-core native Atlas I would generally see the active threads reduce over usually 1-2 minutes, 5 if slow, for tasks of about 4 hrs total wall-clock. (There might be numbers in some ancient post here, but the laptop's tired tonight :( ) |
|
18)
Message boards :
ATLAS application :
Atlas task slowing right down near the end but still using all cores - continue?
(Message 46399)
Posted 2 Mar 2022 by Henry Nebrensky Post: Even within a task, as the number of threads is reduced then each thread must run more events, and it's more likely - but not guaranteed - that they will average out across the threads.... I'm surprised there's a difference though, considering the wide variance in event times.... they are random sizes.You are looking on just 1 task but you would have to look at the long term averages. Was the wide variance I saw unusual?I've no idea. Also, do you have a figure for how much time is wasted? Since it's 200 in the pool, the wasted cores at the end are probably a fraction of a percent of inefficiency.IIRC, back when I was running 8-core native Atlas I would generally see the active threads reduce over usually 1-2 minutes, 5 if slow, for tasks of about 4 hrs total wall-clock. (There might be numbers in some ancient post here, but the laptop's tired tonight :( ) |
|
19)
Message boards :
ATLAS application :
Atlas task slowing right down near the end but still using all cores - continue?
(Message 46390)
Posted 1 Mar 2022 by Henry Nebrensky Post: As CP mentioned each ATLAS task processes 200 events from a pool.It has struck me before, that changing the task's pool size to 180 or 240 events would give better divisibility. |
|
20)
Message boards :
Sixtrack Application :
Sixtrack BOINC application on track to be ported to GPUs?
(Message 46205)
Posted 9 Feb 2022 by Henry Nebrensky Post: More to the point, you don't even need to start worrying about Squid until after you've been through the VirtualBox/CVMFS palaver at least once (to create useful traffic for a Squid to cache), and then only for the larger scale crunchers.... and now SQUID. That is a lot to ask of volunteer crunchers.Set up once, runs for months... If Sixtrack becomes as painful to get running as, say, Atlas, then surely this will be a big turn-off when trying to encourage new, casual volunteers to join the project. |
Next 20
©2024 CERN