LHC@home

I'm reactivating LHC@Home after being away for a while. I have just installed the latest BOINC and VirtualBox. and attached to the project.

However, according to the event log, there are no tasks available for any subproject. This contradicts what the web site Server Status says.

What can I do to fix this and get operating?

Thanks,
++PLS

What can I do to fix this and get operating?

(1) Ignore the message.
(2) Wait a while.
(3) It might work.
(4) No guarantees though.

Good luck.

Got Atlas, Theory and CMS tasks. One Atlas download failed, others seem to work.
Tullio

And I'm still not getting anything at all.

My computer ID is 10483643, is anyone knows what to look for.

The server says status page shows tasks ready to send. My event log just says
2020-07-11 02:01:57 | LHC@home | Sending scheduler request: Requested by project.
2020-07-11 02:01:57 | LHC@home | Requesting new tasks for CPU
2020-07-11 02:02:00 | LHC@home | Scheduler request completed: got 0 new tasks
2020-07-11 02:02:00 | LHC@home | No tasks sent
2020-07-11 02:02:00 | LHC@home | No tasks are available for SixTrack
2020-07-11 02:02:00 | LHC@home | No tasks are available for Theory Simulation
2020-07-11 02:02:00 | LHC@home | No tasks are available for ATLAS Simulation

Try enabling CMS too. Sometimes it helps to have them all enabled, at least to get started.
You might try disabling it later.

https://lhcathome.cern.ch/lhcathome/show_host_detail.php?hostid=10483643
This computer has no hardware virtualization support - or at least it is switched off.
See Yeti's checklist for more details:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4161&postid=29359

In addition you should ensure that "Run native if available?" is unchecked in your preferences:
https://lhcathome.cern.ch/lhcathome/prefs.php?subset=project

As a result the computer can't get any vbox tasks.


SixTrack runs independent from VirtualBox but really has no fresh work ATM, just very few resends from time to time.

Using this old thread as my problem fits the subject.

I have two hosts that have or had this problem. For the other machine (win 7 i7 3770) the server stopped sending me new tasks on Wednesday and only stating that no tasks are available. LHC was the only CPU project running in this host and all LHC tasks were finished. No matter what selections I made in the preferences the result was always the same 'No tasks available' even when requests were made every 5-10 minutes. I had to enable CPU usage in Einstein@home to get something running on CPU (Einstein is usually running on GPUs only). Finally for some reason yesterday LHC started sending me tasks, first some Theory and then some Atlas. And then a boat load of CMS tasks (total estimated run time 1830 hours = 13 hours/task, although cache is set for 0.65+0.1 days). Let's see how they actually run.

Last night my other host (win 10 i7-7820X) had also run out of LHC tasks with same message, This host was originally requesting only Theory and Atlas tasks and they had all run out. I have now enabled Einstein on the CPU like I did for the first host and changed LHC to accept all subprojects. Once all free CPUs are running Einstein tasks, I'll switch CPU back to requesting only LHC work. So far the response from LHC server is the same 'No tasks available' for every request (every 5-10 minutes). This host has the cache setting of 0.5+0.1 days, I'll increase that to 0.65+0.1 as well.

---

Hello Harri,
had the same constellation.
Cern-IT had upgrades Hardware on Thursday.
Have WCG for a second project.
Boinc need a flow of tasks from a project.
If this is not avalaible, no new tasks is the message.

My thinking is a fast Hardware, but problems on network in Cern-IT for Tier3.

Yep, something is wrong somewhere. The funny thing was that when win 7 machine had the problem the win 10 machine had no problem getting work at the same time. They are connected to same ethernet switch. Now that the win 7 machine is drenched with CMS tasks it is no longer requesting new tasks from LHC@home.

What puzzles me is also the occasional server response that 'There are AMD GPU tasks available but your preferences are set not to accept them' although there isn't actually any AMD tasks available as there is no application for it. Now I have enabled those tasks as well (although I don't have AMD GPU on these machines) just to see if that helps (it doesn't of course). So this makes me wonder what else is screwed in the server programs.

---

Yep, something is wrong somewhere. The funny thing was that when win 7 machine had the problem the win 10 machine had no problem getting work at the same time. They are connected to same ethernet switch. Now that the win 7 machine is drenched with CMS tasks it is no longer requesting new tasks from LHC@home.

I have had two identical Linux machines, one getting work and one not. Fortunately, both are getting work now. I will leave them alone.

What puzzles me is also the occasional server response that 'There are AMD GPU tasks available but your preferences are set not to accept them' although there isn't actually any AMD tasks available as there is no application for it. Now I have enabled those tasks as well (although I don't have AMD GPU on these machines) just to see if that helps (it doesn't of course). So this makes me wonder what else is screwed in the server programs.

That has been explained many times. There is a recent discussion for it somewhere.

That has been explained many times. There is a recent discussion for it somewhere.

Yes, I know that Boinc thinks that AMD GPU is supported because 'native_theory' and 'native_mt' contain the string 'ATI'. Anyway as there is actually no application for ATI GPU available, no work can be dedicated to it. So what happens to those tasks if they are meant for ATI but nobody can download them? Or do they actually get downloaded but are done with CPU by a host that accepts them or are they removed somehow from the queue?

---

So what happens to those tasks if they are meant for ATI but nobody can download them?

There aren't any tasks at all for ATI. It is all a figment of the BOINC server imagination.

My problem seems to be over. The win 10 machine has downloaded some new work. I don't know if the solution was changing the value of max# of jobs but that was the last change I did before new tasks were downloaded. The win 7 machine was set to max 8 tasks and I changed that to No limit. The win 10 machine was at No limit and I changed that to 8. A couple of hours later tasks were downloaded, so effect was not immediate. I then changed the win 10 machine back to No limit because a buffer of 8 tasks is not enough for a 16 thread CPU.

After about 90 minutes the downloaded Einstein tasks should be finished, LHC tasks should start and I'll see if it will download more tasks. So fingers crossed here.

---