Computation Error?

Twice now, after 87 hours of astropulse crunching. Am I wasting my time, and my machine's time?

John W.

If you haven't reported these yet, can you please do so so we can see the actual error message on it? At this moment neither of your two computers show any problems with any tasks.

To report the work ahead of time, open BOINC Manager->Advanced view->Projects tab->Select Seti->Click Update.

---

ap_27au09ae_B1_P1_00188_20090921_19229.wu_0, for instance, under TASKS shows the Status "Computation Error" on my machine.

Yes, but we cannot see that at this time. You will have to report that information back to the database first before we can see the task at hand with the error that it has given.

You could check in the Messages tab and find the actual error message. Copy that out of there, with the messages around it (10 lines, no more please). But really the easiest way is to make BOINC report the task to the database, as then it will show up in that computer's task list, which we can see here for your Windows machine and here for your Linux machine.

---

Yes, on your machine, but not on the web page. For it to appear there, you must report it, as Ageless wrote.

Gruß,
Gundolf
[edit]too late :-)[/edit]

Hmmmm. Strange... the Messages show the task finished, and the upload completed. Only the "Status" under "Tasks" seems to have this problem with "Computation Error".

If there's no loss of data I'm happy. Thanks.

There's a difference between uploading and reporting.

[edit]If you just pressed "Update", we'll have to wait for 2,414 seconds before we can see the result :-) [/edit]

Uploading just sends a bit of data back to a place on a hard drive on one of the servers. Reporting will send the output of the task to the database. Since using the database uses more overhead than just uploading does, BOINC tries to report as many tasks at the same time as possible.

Reporting one task takes as much overhead as reporting 20 does. Therefore the preference is to report once a day to get as many as possible in the same report.

But in this case, since you seek help, we cannot help you without either the error message you got on the task or the information in the database. Which is why I asked you to report the task manually.

(and then wait the amount of time Gundolf quoted. This is the amount of time that the replica database is behind on the master database. Each time we look at a person's computers and his tasks, we are checking the database. All of the information about tasks is taken from the replica database, so as to lift the burden on the master database).

---

You're losing me. I'm just a button-pusher. Just tell me that this "Computation Error" message is an error, and the problem goes away. :- )

Well, of course the Computation Error is an error. But we don't know what is causing it without further information from your computer on what caused it. So I will wait for that error to be reported back to the database and if you want me to tell you if it is something caused by your computer or something in the task.

---

Computation Error?

Twice now, after 87 hours of astropulse crunching. Am I wasting my time, and my machine's time?

John W.

Yes, it seems so, at least with AstroPulse on your linux system.

The last erroneous task is now visible in your online task list and shows a segmentation violation error (sigsegv). With that, you seemingly are not alone, see here.

For an explanation of the error, see this message. (I did a forum search for SIGSEGV.)

So, the easiest way to avoid this error would be to disable AstroPulse in your SETI@home preferences. If you do want to get AstroPulse tasks for your windows host, you can Add separate preferences for home, school or work on that same page and move one of your machines there (on the Details page).

Gruß,
Gundolf

---

Computer sind nicht alles im Leben. (Kleiner Scherz)

SETI@home classic workunits 3,758
SETI@home classic CPU time 66,520 hours

If it's any assistance in this error, I also get this, on linux.
Mine happens when it suspends AP work for benchmarks...

Sun 25 Oct 2009 08:31:51 AM CDT Running CPU benchmarks
Sun 25 Oct 2009 08:31:51 AM CDT Suspending computation - running CPU benchmarks
Sun 25 Oct 2009 08:31:52 AM CDT SETI@home Computation for task ap_09se09af_B1_P1_00326_20091019_17933.wu_1 finished
Sun 25 Oct 2009 08:31:53 AM CDT SETI@home Computation for task ap_09se09ae_B6_P0_00210_20091019_30100.wu_3 finished
Sun 25 Oct 2009 08:31:54 AM CDT SETI@home Started upload of ap_09se09af_B1_P1_00326_20091019_17933.wu_1_0
Sun 25 Oct 2009 08:31:55 AM CDT SETI@home Started upload of ap_09se09ae_B6_P0_00210_20091019_30100.wu_3_0
Sun 25 Oct 2009 08:31:58 AM CDT SETI@home Finished upload of ap_09se09af_B1_P1_00326_20091019_17933.wu_1_0
Sun 25 Oct 2009 08:31:58 AM CDT SETI@home Finished upload of ap_09se09ae_B6_P0_00210_20091019_30100.wu_3_0
Sun 25 Oct 2009 08:32:24 AM CDT Benchmark results:
Sun 25 Oct 2009 08:32:24 AM CDT Number of CPUs: 2
Sun 25 Oct 2009 08:32:24 AM CDT 1061 floating point MIPS (Whetstone) per CPU
Sun 25 Oct 2009 08:32:24 AM CDT 2079 integer MIPS (Dhrystone) per CPU
Sun 25 Oct 2009 08:32:26 AM CDT Resuming computation

Those APs show in the task list with computation error as they are obviously incomplete. Hope this helps.

---