Page 3 of 3

Re: Computation errors

Posted: Wed Nov 03, 2010 6:59 pm
by matt
Right now, we only support Snow Leopard.

Re: Computation errors

Posted: Wed Nov 03, 2010 8:36 pm
by Slave_Mac
I am running a completely updated 10.6.4 Snow Leopard MacBook Pro.
That would now be 10.6.5 and with the new client have seen no errors. Haven't seen any 139#####'s either.

Re: Computation errors

Posted: Wed Nov 03, 2010 9:33 pm
by matt
Those workunits should now work :) Thanks.

Re: Computation errors

Posted: Wed Nov 03, 2010 9:46 pm
by mitrichr
I just went, on 10.28 and 10.29, a couple of days of a huge load of small WU's, about which I understand. I just want to say, in 24 hours of sub ten minute WU's, I did 2006 credits, I had zero errors or unvalidated WU's. I am on a Win 7 i7-840QM with 8 gigs of DDR3 DRAM.

>>RSM

Re: Computation errors

Posted: Thu Nov 04, 2010 6:27 pm
by nullcoding
Right...it's not a Windows problem with those 139######## WUs. I too mentioned that my XP machine crunched (crunches) them fine.

No it is not fixed. I still get computation errors.

Code: Select all

Thu Nov  4 14:24:24 2010	eon2	Starting 1391006163_633_10415275_0
Thu Nov  4 14:24:24 2010	eon2	Starting task 1391006163_633_10415275_0 using eonclient version 100
Thu Nov  4 14:24:25 2010	eon2	Computation for task 1391006163_633_10415275_0 finished
Thu Nov  4 14:24:25 2010	eon2	Output file 1391006163_633_10415275_0_0 for task 1391006163_633_10415275_0 absent
Still running Snow Leopard. If its just my computer I wouldn't really be surprised, but this particular machine I've actually not done anything to in the ways of customization (which inevitably leads to incompatibility) ...

Re: Computation errors

Posted: Thu Nov 04, 2010 6:39 pm
by Slave_Mac
matt wrote:Those workunits should now work :) Thanks.
FYI, after reading your post yesterday I requested no new work, reset the project and held of for a couple of hours to allow new work. Since resuming work 1391006163 are still failing as in my previous post. Thank you for responding.


The attached is my message log if you can parse it for something useful.

Re: Computation errors

Posted: Thu Nov 11, 2010 4:42 pm
by Saenger
Most of my WUS work fine, I just got one that errored out with this message:

Code: Select all

<core_client_version>6.10.17</core_client_version>
<![CDATA[
<message>
process exited with code 193 (0xc1, -63)
</message>
<stderr_txt>
Error: size of hessian is zero. Try with smaller min_Displacement_Hessian
SIGSEGV: segmentation violation
Stack trace (2 frames):
[0x46477d]
[0x473c50]

Exiting...

</stderr_txt>
What's a "Hessian"? I'm from northern Germany, not Hessen ;)

Re: Computation errors

Posted: Thu Nov 11, 2010 11:16 pm
by davisb
139######## WUs are throwing computation errors for me as well. I am running OS X 10.6.5 and OS X Server 10.6.5 (fails on both).

http://eon.ices.utexas.edu/eon2/results ... =0&state=5

Re: Computation errors

Posted: Mon Nov 22, 2010 12:32 am
by chill
davisb wrote:139######## WUs are throwing computation errors for me as well. I am running OS X 10.6.5 and OS X Server 10.6.5 (fails on both).

http://eon.ices.utexas.edu/eon2/results ... =0&state=5
These issues have been fixed in the latest version of our client. In addition we also support OS X 10.5.

Re: Computation errors

Posted: Mon Nov 22, 2010 9:19 pm
by ScouseSandwich
Saenger wrote:What's a "Hessian"? I'm from northern Germany, not Hessen ;)
#

I cant help you with the direct problem but i can give you a little information on what a 'hessian' is. I believe it is reffering to the Hessian Matrix, this is used in the calculation of the molecular vibrational frequencies, I think a value of zero is the problem as in the calculation i THINK the square root of the Hessian matrix is taken so if there is a zero value in there then that is bad news!

Re: Computation errors

Posted: Fri Dec 03, 2010 8:15 pm
by Saenger
I just had 2 that crunched too long and errored out after 13,455 seconds.

Code: Select all

<core_client_version>6.10.17</core_client_version>
<![CDATA[
<message>
Maximum elapsed time exceeded
</message>
<stderr_txt>

</stderr_txt>
]]>
Usually the WUs take just a few minutes. The current one is at 40 minutes and counting, no progress, but I think there never was any, and I consider aborting it.

It doesn't seem to have a big memory footprint, at least not after 3h, just around 200MB, so virtually nothing on my 4GB machine. So it doesn't seem to have anything in common with the error reported in this thread.

Re: Computation errors

Posted: Fri Dec 03, 2010 10:01 pm
by Phil P
My 1.21 client 605860314_19_45810 task has been running for 100 minutes with no movement of the progress bar and has written the following into std_0 over 200 times:


Dephasing Warning: in a new state,now inverse the momentum and restart from step 0
Starting search for transition step with accuracy of 10 steps
refined step = 3
Dephasing Warning: in a new state,now inverse the momentum and restart from step 0
Starting search for transition step with accuracy of 10 steps
refined step = 3