Questions and Answers :
Unix/Linux :
Runs seti and CP at the same time on a mono cpu!
Message board moderation
Author | Message |
---|---|
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
Ok, this has now happened twics, on two differnt linux boxes - redhat 9. I'm running CP and Seti. When a CP WU completes (every month or so), boinc downloads a new Wu, as it should. The new CP wu starts, but the SETI wu stays running! (both are grabbing whatever cycles they can). Restarting Boinc fixes the problem, and since CP wu's take so long, it doesn't happen that often. The progress on the SP WU stays at 0%, and after 6 hours, the seti wu is only .7% (seti normally takes 4h to complete). It appears that that is a bug in task startup/switching. (boinc 4.13) |
KWSN - MajorKong Send message Joined: 5 Jan 00 Posts: 2892 Credit: 1,499,890 RAC: 0 |
I have the same issue on my 1-cpu PII-400 box (Gentoo Linux). PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 8259 majorkon 39 19 58784 46m 1144 R 46.7 14.8 15:27.25 mfoldB120_4.21_ 8183 majorkon 39 19 18716 15m 2504 R 44.7 5.0 78:03.58 setiathome_4.02 Boinc v4.13 Linux pii400gen 2.6.9-gentoo-r4 #1 Sun Nov 21 17:17:00 CST 2004 i686 Pentium II (Deschutes) GenuineIntel GNU/Linux glibc version 2.3.4.20040808-r1 The other day, i noticed that while predictor was running, seti would grab about 0.01 sec of cpu once every minute or two, but predictor would not grab time from seti. I happened to notice the load average today, and investigated. Both were running at the same time, sharing the cpu time. [edit]Investigating the messages tab that boincview (0.9.2b, running on one of my windows boxes) keeps, it looks like setiathome_4.02 did not pause when asked to so that mfoldB120_4.21 could run. I wish i could copy/paste the messages here, but I can not figure out how to get boincview to let me copy them (it gets them via RPC), and I have stdout and stderr going to /dev/null on that linux box. Oh, and I have my prefs set to: Leave applications in memory while preempted? (suspended applications will consume swap space if 'yes') yes Switch between applications every (recommended: 60 minutes) 60 minutes Also, the one hour time period that predictor was set to run has elapsed, and predictor has paused. seti just kept right on going, and now is getting 99.7% of the cpu time (as normally it does). It looks like setiathome_4.02 does not properly pause when boinc asks it to do so. https://youtu.be/iY57ErBkFFE #Texit Don't blame me, I voted for Johnson(L) in 2016. Truth is dangerous... especially when it challenges those in power. |
Charles Dennett Send message Joined: 29 Apr 00 Posts: 27 Credit: 18,785 RAC: 0 |
I've noticed the same thing with both the Seti and Mfold applications running at the same time. I was running RedHat 9 until a couple of weeks ago when I upgraded to Fedora Core 3. Unfortunatley, I do not have a record of which application refused to pause when the core client wanted to start the other. If I noticed it in time, a stop and restart of the core client fixed it. Otherwise, the it would switch back to the first application after an hour. System is using an Asus A7V8X motherboard with an AMD XP2600+ CPU. Charlie |
KWSN - MajorKong Send message Joined: 5 Jan 00 Posts: 2892 Credit: 1,499,890 RAC: 0 |
I just noticed it happening again: retyping the messages as best I can: SETI@home Pausing result 27mr04aa.17618.3906.417318.60_2 (left in memory) ProteinPredictorAtHome Starting result t0239D_1_4283_0 using mfoldB120 version 4.21 ProteinPredictorAtHome Result t0239D_1_4293_0 exited with zero status but no 'finished' file ProteinPredictorAtHome If this happens repeadedly, you may need to reset the project. ProteinPredictorAtHome Restarting result t0239D_1_4293_9 using mfoldB120 version 4.21 from the stderr.txt file in the slot (1) predictor is running in: majorkong@pii400gen 1 $ cat stderr.txt No heartbeat from core client for 30.026051 sec - exiting seti@home's stderr.txt is empty. this results in both seti and predictor running at the same time (each getting about 50% of the cpu) for an hour, until it is time for it to switch back to seti, which it does. majorkong@pii400gen boinc $ uname -a Linux pii400gen 2.6.9-gentoo-r4 #1 Sun Nov 21 17:17:00 CST 2004 i686 Pentium II (Deschutes) GenuineIntel GNU/Linux BOINC v4.13 https://youtu.be/iY57ErBkFFE #Texit Don't blame me, I voted for Johnson(L) in 2016. Truth is dangerous... especially when it challenges those in power. |
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
I saw this again today, but with mfold and seti.. Mfold started to crunch but seti tried to keep going too. Still redhat 9 and boinc 4.13. |
Jean-David Beyer Send message Joined: 10 Jun 99 Posts: 60 Credit: 1,301,105 RAC: 1 |
> I saw this again today, but with mfold and seti.. > > Mfold started to crunch but seti tried to keep going too. > > Still redhat 9 and boinc 4.13. > I get the same kind of problems on my Red Hat Enterprise Linux 3 ES system. This system has two hyperthreaded Intel Xeon processors, so Linux thinks there are 4 processors. Furthermore, I told BOINC to use up to 4 processors in SMP systems. Right now I have 4 setiathome tasks available and three hadsm3 tasks available. Only three tasks are running: all are setiathome. The three hadsm3 tasks are in the process table, but in stopped state. Surely the BOINC client should start one of them. Earlier today, it was running 5 processes, at least on of each: climateprediction, proteinfolding, and setiathome. It is not too concientious about running as much work as it could, and sometimes it does not respect the proper number of tasks to run. Running too many processes serves only to lower the total throughput, so it is not an effective thing to do. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.