Version 2.30 released & new Project available

Message boards : News : Version 2.30 released & new Project available

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
Thomas Koch
Project administrator
Project developer
Project scientist

Send message
Joined: 17 Feb 12
Posts: 436
Credit: 37,847
RAC: 0
Message 10954 - Posted: 13 Apr 2016, 12:43:45 UTC

Finally the POEM++ Version 2.30 is live.
The test phase a few weeks ago ended successfully, but if you encounter any problems with the new release, please let me know.

Along with the update we have started a new
Peptide project. You can get some information on the linked project page, a visualization of the simulation process will follow.


Die POEM++ Version 2.30 ist jetzt verfügbar.
Die Testperiode vor einigen Wochen war erfolgreich, falls es aber trotzdem Probleme mit dem neuen Release geben sollte, lasst es mich bitte wissen.

Einhergehend mit dem Update haben wir ein neues Peptidprojekt gestartet. Auf der verlinkten Seite gibt es einige Informationen, eine Visualisierung des Simulationsprozesses wird noch folgen.
ID: 10954 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
frederikhk

Send message
Joined: 28 Oct 15
Posts: 6
Credit: 2,030,390
RAC: 0
Message 10955 - Posted: 13 Apr 2016, 15:57:38 UTC
Last modified: 13 Apr 2016, 16:01:50 UTC

EDIT: Seems to only happen with 2dx4cl units others run fine.
All my GPU workunits error out now.

<core_client_version>7.6.22</core_client_version>
<![CDATA[
<message>
The remote server has been paused or is in the process of being started.
(0x46) - exit code 70 (0x46)
</message>
<stderr_txt>
17:54:24 (7000): called boinc_finish(70)

</stderr_txt>
]]>

http://boinc.fzk.de/poem/result.php?resultid=188658989
ID: 10955 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile nenym

Send message
Joined: 16 Oct 09
Posts: 3
Credit: 100,156,703
RAC: 0
Message 10956 - Posted: 13 Apr 2016, 16:17:33 UTC - in response to Message 10955.  
Last modified: 13 Apr 2016, 16:19:09 UTC

EDIT: Seems to only happen with 2dx4cl units others run fine.
All my GPU workunits error out now.

<core_client_version>7.6.22</core_client_version>
<![CDATA[
<message>
The remote server has been paused or is in the process of being started.
(0x46) - exit code 70 (0x46)
</message>
<stderr_txt>
17:54:24 (7000): called boinc_finish(70)

</stderr_txt>
]]>

http://boinc.fzk.de/poem/result.php?resultid=188658989

+1 The same here
ATI R9 280 W7 64bit
2f21 brand runs on 2.30app fine
ID: 10956 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Thomas Koch
Project administrator
Project developer
Project scientist

Send message
Joined: 17 Feb 12
Posts: 436
Credit: 37,847
RAC: 0
Message 10957 - Posted: 13 Apr 2016, 17:14:30 UTC

Apparently there is a bug in the work unit creation for the new jobs.
I've turned off the project for the moment, and try to flush the created workunits from the queue.

This may take a while, please be patient.
ID: 10957 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Detlef Harbach

Send message
Joined: 27 Dec 15
Posts: 2
Credit: 16,424
RAC: 0
Message 10966 - Posted: 15 Apr 2016, 3:14:25 UTC - in response to Message 10954.  

Ich habe heute Nacht 7 Arbeitspakete erhalten, die alle nach kurzer Zeit mit einem BERECHNUNGSFEHLER endeten, Endstatus: -1073741515 (0xffffffffc0000135) Unknown error number. Ich wundere mich nur, weil ich hier noch nie einen Berechnungsfehler hatte.
Gruß, Detlef Harbach.
ID: 10966 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Detlef Harbach

Send message
Joined: 27 Dec 15
Posts: 2
Credit: 16,424
RAC: 0
Message 10967 - Posted: 15 Apr 2016, 6:02:12 UTC - in response to Message 10966.  

Gerade sind weitere 9 Arbeitspakete mit dem gleichen Fehler beendet worden. Ich hab das Projekt jetzt für mich erst mal angehalten, bevor ich hier was kaputt mache...
ID: 10967 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Felix2015

Send message
Joined: 31 Jul 15
Posts: 14
Credit: 109,080,301
RAC: 0
Message 10970 - Posted: 15 Apr 2016, 10:47:04 UTC

Hallo Thomas - vielleicht solltest du zur Information in allen treads reinschreiben, dass der Treiber von Geforce 364.xx immer Poem-WUs killt.
Das ist ein Treiber-Problem.
Funktionieren tut 362.00.

Wenn also user sagen, dies und das hat Berechnungsfehler, dann ist es sinnvoll zu schauen, welche ART von Grafikkarte derjenige hat und eben welchen Treiber!
Sonst gibt es Fehlersuchen - Aufregungen - Beschwerden, die völlig überflüssig sind.
ID: 10970 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
BlisteringSheep
Avatar

Send message
Joined: 26 Dec 07
Posts: 1
Credit: 51,864,425
RAC: 0
Message 10972 - Posted: 16 Apr 2016, 2:51:40 UTC

Most CPU tasks are failing immediately. The Linux CPU executable was linked with OpenCL. Linux hosts that have OpenCL installed seem to be running the WUs okay.

Linux 64-bit host:
<core_client_version>7.2.33</core_client_version>
<![CDATA[
<message>
process exited with code 127 (0x7f, -129)
</message>
<stderr_txt>
../../projects/boinc.fzk.de_poem/poempp_2.30_x86_64-pc-linux-gnu: error while loading shared libraries: libOpenCL.so.1: cannot open shared object file: No such file or directory

</stderr_txt>
]]>


Windows 7 host (same for 64-bit and 32-bit):
<core_client_version>7.6.22</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code -1073741515 (0xc0000135)
</message>
]]>
ID: 10972 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile robertmiles

Send message
Joined: 13 Jul 08
Posts: 68
Credit: 46,353,581
RAC: 0
Message 10973 - Posted: 16 Apr 2016, 19:50:47 UTC

I've had several opencl_nvidia_101 2.30 tasks give an Error Computing lately, and act like they left BOINC and the graphics driver no longer agreeing on whether the GPU is in use - several GPU tasks from other BOINC projects failed quickly soon afterwards. Windows showed an error message on the screen, asking for permission to close the 2.30 application program.

http://boinc.fzk.de/poem/result.php?resultid=189023997
http://boinc.fzk.de/poem/result.php?resultid=189023998
http://boinc.fzk.de/poem/result.php?resultid=189023776
http://boinc.fzk.de/poem/result.php?resultid=189023782
http://boinc.fzk.de/poem/result.php?resultid=189022297
http://boinc.fzk.de/poem/result.php?resultid=189022313
http://boinc.fzk.de/poem/result.php?resultid=189022151
http://boinc.fzk.de/poem/result.php?resultid=189022202
http://boinc.fzk.de/poem/result.php?resultid=189022087
http://boinc.fzk.de/poem/result.php?resultid=189022110

stderr showed:

(unknown error) - exit code -1001 (0xfffffc17)
</message>
<stderr_txt>
An OpenCL error occured in main:
clGetPlatformIDs -1001 Unknown OpenCL error
13:03:51 (8040): called boinc_finish(-1001)

One 2.30 task gave a different error:

http://boinc.fzk.de/poem/result.php?resultid=188999468

(unknown error) - exit code -1001 (0xfffffc17)
</message>
<stderr_txt>
Error while running born radii 2 kernel!
clFinish -36 CL_INVALID_COMMAND_QUEUE
terminate called after throwing an instance of 'cl::Error'
what(): clFinish

Several earlier 2.30 tasks completed and validated on the same computer.

http://boinc.fzk.de/poem/show_host_detail.php?hostid=53474

Looks like you need to add more stderr output near these errors.
ID: 10973 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile PFLIEGER Guy

Send message
Joined: 4 Jan 16
Posts: 28
Credit: 34,550,560
RAC: 0
Message 10974 - Posted: 16 Apr 2016, 20:42:13 UTC

The new project ot peptide present important technik problem: a lot of work unit with Nvidia card make shut down the computer. I like the poem home project but we can not work with technical problem

Best Regards

Guy PFLIEGER
ID: 10974 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
frederikhk

Send message
Joined: 28 Oct 15
Posts: 6
Credit: 2,030,390
RAC: 0
Message 10975 - Posted: 16 Apr 2016, 20:42:46 UTC - in response to Message 10974.  

What driver version are you using?
ID: 10975 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile PFLIEGER Guy

Send message
Joined: 4 Jan 16
Posts: 28
Credit: 34,550,560
RAC: 0
Message 10976 - Posted: 16 Apr 2016, 21:20:25 UTC - in response to Message 10975.  

10.18.13.6472
Nvidia
ID: 10976 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile robertmiles

Send message
Joined: 13 Jul 08
Posts: 68
Credit: 46,353,581
RAC: 0
Message 10977 - Posted: 16 Apr 2016, 21:25:19 UTC - in response to Message 10973.  

I've had several opencl_nvidia_101 2.30 tasks give an Error Computing lately, and act like they left BOINC and the graphics driver no longer agreeing on whether the GPU is in use - several GPU tasks from other BOINC projects failed quickly soon afterwards. Windows showed an error message on the screen, asking for permission to close the 2.30 application program.


Threads on the problems:

https://www.primegrid.com/forum_thread.php?id=6769#94223

http://boinc.fzk.de/poem/forum_thread.php?id=1205#10896

Summary - the problem appears to be part of the 364.* Nvidia driver series. Some people went back to the 362.00 driver, and haven't seen the problem since.
ID: 10977 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile PFLIEGER Guy

Send message
Joined: 4 Jan 16
Posts: 28
Credit: 34,550,560
RAC: 0
Message 10978 - Posted: 16 Apr 2016, 21:35:35 UTC - in response to Message 10976.  

I made update of the nvidia driver. The computer said:I must update the pci driver. I updated now 15 other system driver. Now I want to see what happen. I make now test
ID: 10978 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile PFLIEGER Guy

Send message
Joined: 4 Jan 16
Posts: 28
Credit: 34,550,560
RAC: 0
Message 10979 - Posted: 16 Apr 2016, 21:52:13 UTC - in response to Message 10978.  

The testing of the computer with the new driver shows that the gpu nvidia problem is failing. there must be à bug in the poem program.
The errors are comming only from the GPU work unit.
Every GPU WU are failing
ID: 10979 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
frederikhk

Send message
Joined: 28 Oct 15
Posts: 6
Credit: 2,030,390
RAC: 0
Message 10982 - Posted: 16 Apr 2016, 22:44:15 UTC - in response to Message 10979.  

the bug is in nvidias recent drivers. Downgrade to 362.00 and under.
ID: 10982 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile PFLIEGER Guy

Send message
Joined: 4 Jan 16
Posts: 28
Credit: 34,550,560
RAC: 0
Message 10985 - Posted: 17 Apr 2016, 17:36:30 UTC - in response to Message 10982.  

how did you make to downgrade the gpu?
ID: 10985 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile PFLIEGER Guy

Send message
Joined: 4 Jan 16
Posts: 28
Credit: 34,550,560
RAC: 0
Message 10986 - Posted: 17 Apr 2016, 17:58:08 UTC - in response to Message 10985.  

this was the great answer to the gpu problem! Thank You! Now it is solved!

Guy PFLIEGER
ID: 10986 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
frederikhk

Send message
Joined: 28 Oct 15
Posts: 6
Credit: 2,030,390
RAC: 0
Message 10987 - Posted: 17 Apr 2016, 22:28:28 UTC - in response to Message 10986.  

You are most welcome!
ID: 10987 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Thomas Koch
Project administrator
Project developer
Project scientist

Send message
Joined: 17 Feb 12
Posts: 436
Credit: 37,847
RAC: 0
Message 10995 - Posted: 19 Apr 2016, 17:06:04 UTC
Last modified: 19 Apr 2016, 17:06:56 UTC

As there seem to be problems on multiple hosts which I cannot solve without extra debug information, I've created short runs of the current jobs for testing purposes.

If your host ran with the last poempp version, and does no longer, please download one of these files (depending on your architecture):
Linux
Windows

Extract the folder, and run cpu_test.bat / gpu_test.bat (on Windows) or cpu_test.sh / gpu_test.sh (on Linux) depending on which work unit you want to test. This will take up to a few minutes and create a file "log_*.txt".
Please send this textfile to poem-home@int.kit.edu.

Thanks for your help!
ID: 10995 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
1 · 2 · 3 · Next

Message boards : News : Version 2.30 released & new Project available


Copyright © 2017 KIT-INT