Message boards : Sixtrack Application : Inconclusive, valid/invalid results
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 4 · 5 · 6 · 7 · 8 · 9 · Next

AuthorMessage
Eric Mcintosh
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 12 Jul 11
Posts: 857
Credit: 1,619,050
RAC: 0
Message 31504 - Posted: 19 Jul 2017, 19:01:40 UTC

Until two weeks ago NULL results, empty results, files nbytes 0 in stderr, were being
validated against each other. The unlucky volunteer who computed
for a few hours would then be invalidated! NULL results are now invalidated
and this is much better. I believe Intel family 6 Kernel 5.8.* is a major source
of the NULL results as SixTrack crashes sometimes with a SIGSEV. Eric.
ID: 31504 · Report as offensive     Reply Quote
James Molson
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 25 Jan 17
Posts: 27
Credit: 3,258,853
RAC: 0
Message 31506 - Posted: 19 Jul 2017, 19:11:31 UTC

Is anybody with one of the above CPUs willing to run the sixtrack testing suite? It will involve building your own version of sixtrack and about 10 hours of cpu time.
ID: 31506 · Report as offensive     Reply Quote
Toby Broom
Volunteer moderator

Send message
Joined: 27 Sep 08
Posts: 798
Credit: 644,705,562
RAC: 234,921
Message 31508 - Posted: 19 Jul 2017, 19:37:29 UTC

Sure James, I have Xeon v3 this is family 6, athough not Model 94, etc like the Kabbey/Sklakes
ID: 31508 · Report as offensive     Reply Quote
Profile Yeti
Volunteer moderator
Avatar

Send message
Joined: 2 Sep 04
Posts: 453
Credit: 193,369,412
RAC: 10,065
Message 31527 - Posted: 20 Jul 2017, 21:14:26 UTC - in response to Message 31506.  
Last modified: 20 Jul 2017, 21:15:16 UTC

Is anybody with one of the above CPUs willing to run the sixtrack testing suite? It will involve building your own version of sixtrack and about 10 hours of cpu time.

and

Intel family 6 Kernel 5.8.*

Could you tell the usual Prozessor-Name ? I don't know how to "translate" Family 6 Kernel 5.8.*


Supporting BOINC, a great concept !
ID: 31527 · Report as offensive     Reply Quote
Toby Broom
Volunteer moderator

Send message
Joined: 27 Sep 08
Posts: 798
Credit: 644,705,562
RAC: 234,921
Message 31528 - Posted: 20 Jul 2017, 22:36:55 UTC

Family 6, is basically all intel cpu, from Pentium Pro!


4.8.x is Linux kernel number fore Debian based operating systems, like mint ubuntu etc
ID: 31528 · Report as offensive     Reply Quote
Eric Mcintosh
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 12 Jul 11
Posts: 857
Credit: 1,619,050
RAC: 0
Message 31529 - Posted: 21 Jul 2017, 4:12:17 UTC - in response to Message 31528.  

I understand that the Intel "Family 6" was introduced/announced on
1st September, 2015!

I am using the terms "Family 6" and "Kernel 4.8.*" because that
is what I find in the Database. (Sadly there is no information about
motherboards/BIOS versions etc.) . Eric. :-)


Family 6, is basically all intel cpu, from Pentium Pro!


4.8.x is Linux kernel number fore Debian based operating systems, like mint ubuntu etc

ID: 31529 · Report as offensive     Reply Quote
boinc127

Send message
Joined: 5 Feb 12
Posts: 19
Credit: 700,079
RAC: 0
Message 31542 - Posted: 21 Jul 2017, 18:33:06 UTC

I understand that the Intel "Family 6" was introduced/announced on
1st September, 2015!


So based on the introduction date the 6th generation Intel Core processors... skylake.

Intel calls all processors from Pentium pro to present Family 6. The model number tells people what the actual processor is. (I don't know how much rhyme or reason is used to assign model numbers, maybe based on date of manufacture).
ID: 31542 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Nov 14
Posts: 602
Credit: 24,371,321
RAC: 0
Message 31543 - Posted: 21 Jul 2017, 21:37:08 UTC - in response to Message 31542.  
Last modified: 21 Jul 2017, 21:40:30 UTC

So based on the introduction date the 6th generation Intel Core processors... skylake.

Intel calls all processors from Pentium pro to present Family 6. The model number tells people what the actual processor is. (I don't know how much rhyme or reason is used to assign model numbers, maybe based on date of manufacture).


To further confuse matters, Intel also defines "generations" within "families". So the Kaby Lake is 7th generation, being i7-7700 for example, Skylake is 6th generation as you note, and a Haswell is 4th generation, being i7-4770 for example. I think.

It does not have much to do with the technology level either, since both Skylake and Kaby Lake are 14nm, and not much different.
ID: 31543 · Report as offensive     Reply Quote
tullio

Send message
Joined: 19 Feb 08
Posts: 708
Credit: 4,336,250
RAC: 0
Message 31549 - Posted: 22 Jul 2017, 9:13:34 UTC

I have a result on my Linux laptop, with kernel 4.4. The wingman failed with a kernel 4.10 and the task is unsent.Why?
Tullio
ID: 31549 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2071
Credit: 156,088,206
RAC: 104,040
Message 31550 - Posted: 22 Jul 2017, 9:57:21 UTC

You need more time to wait for finishing the wingman.
For me, sometime more than seven days. But it works well.
The LHC-Server say 900k sixtrack work is for doing.
ID: 31550 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 31556 - Posted: 22 Jul 2017, 16:35:43 UTC - in response to Message 31549.  

I have a result on my Linux laptop, with kernel 4.4. The wingman failed with a kernel 4.10 and the task is unsent.Why?
Tullio

I think that is one of the major scheduling problems LHC@home has.
To me it looks like that all SixTrack-tasks are stored and waiting in memory.

I think a better setup would be to have about 5000 SixTrack tasks in memory and all the rest prepared in a queue on disk to fill up the memory queue when's going below the minimum of 5000 with e.g. 1000 tasks.

When a resend is needed it would/should be placed in front of the disk-queue, but I think it's now placed to the end of the huge memory queue.
ID: 31556 · Report as offensive     Reply Quote
Eric Mcintosh
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 12 Jul 11
Posts: 857
Credit: 1,619,050
RAC: 0
Message 31558 - Posted: 22 Jul 2017, 17:16:53 UTC - in response to Message 31556.  

iIndeed I think you are correct. By in "memory" I think you mean in the Database.
Nice idea for prioritising....we need to do it somehow.

I have a result on my Linux laptop, with kernel 4.4. The wingman failed with a kernel 4.10 and the task is unsent.Why?

I would really like to have a hostid(s) or resultid(s).....I can then follow up.
=====================================================
Eric.

Tullio

I think that is one of the major scheduling problems LHC@home has.
To me it looks like that all SixTrack-tasks are stored and waiting in memory.

I think a better setup would be to have about 5000 SixTrack tasks in memory and all the rest prepared in a queue on disk to fill up the memory queue when's going below the minimum of 5000 with e.g. 1000 tasks.

When a resend is needed it would/should be placed in front of the disk-queue, but I think it's now placed to the end of the huge memory queue.

ID: 31558 · Report as offensive     Reply Quote
tullio

Send message
Joined: 19 Feb 08
Posts: 708
Credit: 4,336,250
RAC: 0
Message 31560 - Posted: 23 Jul 2017, 7:17:36 UTC - in response to Message 31558.  

My host id is 10454176 and has now 3 results waiting for validation.
Tullio
ID: 31560 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2071
Credit: 156,088,206
RAC: 104,040
Message 31635 - Posted: 26 Jul 2017, 15:12:58 UTC
Last modified: 26 Jul 2017, 15:14:07 UTC

This task was finished from two PC's. Saw this in Task-list:

forrtl: error (78): process killed (SIGTERM)

Stack trace terminated abnormally.
SIGSEGV: segmentation violation
Stack trace (13 frames):
[0x826992d]
[0xf77d4f70]
[0x82e14aa]
[0x83b2c46]
[0x83b18bf]
[0x837a625]
[0x833eeb6]
[0x8345bc5]
[0xf77d4f70]
[0x8276e9c]
[0x82666f6]
[0x82ac69f]
[0x82f5d1e]

Exiting...
forrtl: error (78): process killed (SIGTERM)

Stack trace terminated abnormally.
11:33:54 (2843): called boinc_finish

https://lhcathome.cern.ch/lhcathome/result.php?resultid=152308720
ID: 31635 · Report as offensive     Reply Quote
Toby Broom
Volunteer moderator

Send message
Joined: 27 Sep 08
Posts: 798
Credit: 644,705,562
RAC: 234,921
Message 31646 - Posted: 27 Jul 2017, 6:11:35 UTC

I got one a little like that

https://lhcathome.cern.ch/lhcathome/result.php?resultid=152246976

forrtl: error (78): process killed (SIGTERM)

Stack trace terminated abnormally.
00:31:52 (1920): called boinc_finish
ID: 31646 · Report as offensive     Reply Quote
Eric Mcintosh
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 12 Jul 11
Posts: 857
Credit: 1,619,050
RAC: 0
Message 31660 - Posted: 27 Jul 2017, 15:01:25 UTC - in response to Message 31646.  

Very very interesting indeed. I cannot find ANY trace of the
workunit nor of the result in the sixtrack_validator logs!
Sad, but In have to leave this to my colleagues. Eric.
(It is an Intel family 6 but that means nothing I suppose
and it is running Windows....



I got one a little like that

https://lhcathome.cern.ch/lhcathome/result.php?resultid=152246976

forrtl: error (78): process killed (SIGTERM)

Stack trace terminated abnormally.
00:31:52 (1920): called boinc_finish

ID: 31660 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 31664 - Posted: 27 Jul 2017, 17:22:02 UTC

I don't trust this one for some reason. Both valid.

https://lhcathome.cern.ch/lhcathome/workunit.php?wuid=72911894

152443261 10450537 26 Jul 2017, 15:02:52 UTC 27 Jul 2017, 12:20:37 UTC Completed and validated 38,566.68 32,227.33 186.53 SixTrack v451.07 (sse2) x86_64-pc-linux-gnu
152443262 10455705 26 Jul 2017, 15:00:19 UTC 26 Jul 2017, 15:06:40 UTC Completed and validated 2.08 0.05 186.53 SixTrack v451.07 (pni) x86_64-pc-linux-gnu

...and also not most or all valids of that second listed machine 10455705
ID: 31664 · Report as offensive     Reply Quote
PFLIEGER Guy

Send message
Joined: 22 Jun 17
Posts: 12
Credit: 272,799
RAC: 0
Message 31687 - Posted: 28 Jul 2017, 22:07:16 UTC

pending 29 tasks
unconcluding 4 tasks
conclusive 45 Tasks
Invalid 1 task
Error 12 Tasks

Detail of error Tasks
6 tasks Atlas simulation
2 Tasks LHCb
2 Tasks Theory simulation
2 tasks CMS simulation
0 tasks Six Track
ID: 31687 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2071
Credit: 156,088,206
RAC: 104,040
Message 31991 - Posted: 18 Aug 2017, 19:56:41 UTC
Last modified: 18 Aug 2017, 20:27:48 UTC

This task finished successful. FYI
Saw this message:
<core_client_version>7.2.42</core_client_version>
<![CDATA[
<stderr_txt>
forrtl: error (78): process killed (SIGTERM)

Stack trace terminated abnormally.
21:49:54 (3739): called boinc_finish

</stderr_txt>

https://lhcathome.cern.ch/lhcathome/result.php?resultid=153757957

A second one, sorry

<core_client_version>7.2.42</core_client_version>
<![CDATA[
<stderr_txt>
forrtl: error (78): process killed (SIGTERM)
Image PC Routine Line Source
Unknown F77E2F89 Unknown Unknown Unknown
sixtrack_lin32_45 08262C15 Unknown Unknown Unknown

Stack trace terminated abnormally.
forrtl: error (78): process killed (SIGTERM)

Stack trace terminated abnormally.
22:15:49 (4325): called boinc_finish

</stderr_txt>

https://lhcathome.cern.ch/lhcathome/result.php?resultid=153724932
ID: 31991 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2386
Credit: 222,911,012
RAC: 138,148
Message 32139 - Posted: 29 Aug 2017, 15:25:46 UTC

Today I got 2 sixtrack WUs but both finished as errors on my host as well as on all wingnmen's hosts:
https://lhcathome.cern.ch/lhcathome/workunit.php?wuid=73809210
https://lhcathome.cern.ch/lhcathome/workunit.php?wuid=73811381
ID: 32139 · Report as offensive     Reply Quote
Previous · 1 . . . 4 · 5 · 6 · 7 · 8 · 9 · Next

Message boards : Sixtrack Application : Inconclusive, valid/invalid results


©2024 CERN