keep getting "Global Correlations S5 HF search #1 " tasks despite having
11 May 2011 8:28:10 UTC
Topic 195785
(moderation:
Hello,
more than 90% of my attempts at "Global Correlations S5 HF search #1 v1.07 (SSE2)" tasks lead to "Error while computing", while the binary pulsar tasks are completed fine. Therefore I changed my options to not receive any of the former. However, I keep getting (and keep wasting cycles on) them regularly. How can I stop this?
Thanks for your help.
Geza
keep getting "Global Correlations S5 HF search #1 " tasks despit
)
1) The error you are getting with the GW tasks look like the ones we know from using a "preemptive kernel" (AFAIK the SuSE "Dektop" kernel belongs to these). Use a different kernel (on SuSE, switch to the "Default" kernel) if possible.
2) Since S5GC1HF the web interface allows to opt-out of the GW search. However we recently discovered a bug in the current scheduler that leads to this setting being ignored. There is nothing you can do about that.
3) If you still feel you need to run only radio pulsar workunits, use an app_info.xml with only the "einsteinbinary_BRP3" application.
For 1) and 3) there are forum threads which explain these in more detail, but I can't find them ATM. Someone else should.
BM
BM
thanks a lot for the quick
)
thanks a lot for the quick and helpful reply; I'll try the suse-default kernel next time I boot up. let's see how it goes.
regards
Geza
For 1), here is the FAQ,
)
For 1), here is the FAQ, which includes links to the threads.
I am using the SuSE pae
)
I am using the SuSE pae kernel in order to use all my 8 GB RAM on 6 BOINC projects.
Tullio
I also ran into this same
)
I also ran into this same problem. I tried updating my application preferences but that did not work. I also tried reset the project but received the following message.
Message from server: Resent lost task...
So I am back to square one. I have decided to set this project 'no new tasks' temporarily until this is fixed.
Hi Jord, Maybe you could
)
Hi Jord,
Maybe you could update that section of the FAQ because the fix that is mentioned at the end obviously doesn't work.
The same behaviour has been observed up to at least kernel version 2.6.34. I don't know about any newer versions because I haven't tried them. But up to 2.6.34 the only solution for this problem was to use a non-preemptible kernel.
Michael
My SuSE kernel is Linux
)
My SuSE kernel is Linux 2.6.27.56-0.1-pae i686 and it works perfectly.
Tullio