log in

Advanced search

Message boards : Number crunching : Venus beta testing

Author Message
Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 567 - 6 Jun 2014, 7:54:47 UTC

Hi all,
This new thread is for those who want to comment about the new program VENUS (still in beta version).
Performance, error messages, calculation time statistics, questions, whatever (you know) ...

Good testing!!
R.

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 568 - 6 Jun 2014, 9:46:04 UTC

After 28,000 seconds, this work unit ended in an Error of Maximum Elapsed Time Exceeded

See this WU

Conan

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 569 - 13 Jun 2014, 23:28:52 UTC

The work units that have been sent out in the last day are working much better and I now have my first validated work units.

Conan

Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 570 - 14 Jun 2014, 9:25:50 UTC - in response to Message 569.

Yeah, its working!

These are for Linux 64bits, now we are working for other plattaforms.

Thanks,

R.

fzs600
Send message
Joined: 2 May 12
Posts: 1
Credit: 112,172
RAC: 3
France
Message 571 - 14 Jun 2014, 17:22:38 UTC - in response to Message 570.

14 Jun 2014, 10:34:49 UTC 14 Jun 2014, 16:34:10 UTC Terminé et validé 21,171.51 20,416.74 130.70 Polyatomic trajectory calculation beta v1.02

Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 572 - 16 Jun 2014, 12:43:08 UTC

Hi again,
New versions for windows and iOS are under testing, so some feedback would be great.

Thanks,
R.

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 576 - 16 Jun 2014, 22:13:29 UTC
Last modified: 16 Jun 2014, 22:15:47 UTC

I have yet to get any Windows work units myself, but a number of work units sent out to Windows users have all failed with the following error

couldn't start app: CreateProcess() failed - (unknown error)

See This Example

The same work units have then run successfully on my Linux machines.

Conan

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 578 - 17 Jun 2014, 5:31:16 UTC - in response to Message 576.

I have yet to get any Windows work units myself, but a number of work units sent out to Windows users have all failed with the following error

couldn't start app: CreateProcess() failed - (unknown error)

See This Example

The same work units have then run successfully on my Linux machines.

Conan


That work unit received Zero credits due to my Linux WU not being able to validate as all the Windows computers crashed all attempts to run, exceeding the error limit on the WU.

Conan

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 581 - 19 Jun 2014, 7:49:39 UTC

It would appear that Windows and Linux do not validate against each other

See Here

Conan

Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 582 - 19 Jun 2014, 10:31:21 UTC - in response to Message 581.

Hi Conan:

It shouldn't. Each calculation is sent twice to different computer, validation is checked only when both results are exactly equal, so both calculations must be sent to the same platform.
Why this calculation were sent to windows and Linux platforms? BOINC mysteries, (Ok, I don't know ..), but you see that the task was validated only once a pair of them were equal.

Thanks,
R.

Curly
Send message
Joined: 16 May 12
Posts: 1
Credit: 162,883
RAC: 0
Germany
Message 593 - 23 Jul 2014, 13:19:08 UTC - in response to Message 568.

After 28,000 seconds, this work unit ended in an Error of Maximum Elapsed Time Exceeded

See this WU

Conan


Another result that ended in an error Maximum Elapsed Time Exceeded after 30,374 seconds. See this result.
Another result of the wu ended in the same error.

With another result that ended with an error (Error while downloading) the wu was marked with status Error while computing.

Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 594 - 25 Jul 2014, 8:27:55 UTC - in response to Message 593.

Hi Curly,
We are not sure if this problem is happening only in windows versions ...
Looking at it.

Thanks.
R.

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 627 - 26 Oct 2014, 0:39:58 UTC

I have returned a number of Venus work units from my 32 bit AMD Win XP machine over the last days but have noticed that I have only 1 WU that has validated.

I have 6 others that have returned Zero credits due to being marked as 'invalid'.

All the computers that I was paired with on these 'invalid' work units have been Intel CPUs.

The 1 successful WU was when I was paired with another AMD CPU.

So to me there seems to be an Intel/AMD incompatibility on Windows (the mix of 32 bit and 64 bit did not seem to make a difference).

All my Linux work units have been successful running the "Venus" application.

(I still have a few to process and 3 pending, but as each pending is paired with an Intel processor I doubt any will validate, 1 of which has run over 44,000 seconds)

Conan

Profile Conan
Send message
Joined: 20 Jun 12
Posts: 24
Credit: 485,993
RAC: 5
Australia
Message 628 - 27 Oct 2014, 9:52:17 UTC - in response to Message 627.
Last modified: 27 Oct 2014, 9:52:46 UTC

I have returned a number of Venus work units from my 32 bit AMD Win XP machine over the last days but have noticed that I have only 1 WU that has validated.

I have 6 others that have returned Zero credits due to being marked as 'invalid'.

All the computers that I was paired with on these 'invalid' work units have been Intel CPUs.

The 1 successful WU was when I was paired with another AMD CPU.

So to me there seems to be an Intel/AMD incompatibility on Windows (the mix of 32 bit and 64 bit did not seem to make a difference).

All my Linux work units have been successful running the "Venus" application.

(I still have a few to process and 3 pending, but as each pending is paired with an Intel processor I doubt any will validate, 1 of which has run over 44,000 seconds)

Conan


So the ones that have not validated due to my AMD processors being paired with Intel processors are
6354375
6354379
6354470
6354477
6354485
6354527

And you can add this one to the list as well 6354371 it had 2 AMD and 2 Intel processors, but the AMD lucked out with Zero whilst the Intel got the points.

I have 3 more pending and I hold no doubts that they will all be given Zero points due to being paired with Intel CPUs.
It is a lot of processing (many, many hours) time that I will get no credit for, please make the time to investigate this issue.

Thanks

Conan

Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 629 - 27 Oct 2014, 15:54:39 UTC - in response to Message 628.

Dear Conan,
thank you for the info, there are no good news ...

Sorry and working on it,
R.

BobCat13
Send message
Joined: 4 May 12
Posts: 4
Credit: 79,901
RAC: 2
United States
Message 630 - 18 Nov 2014, 16:11:31 UTC

You are sending out tasks to pure 64-bit Linux systems with a 32-bit wrapper, which of course fails to run.

file wrapper_x86_64-pc-linux-gnu_73200
wrapper_x86_64-pc-linux-gnu_73200: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.9, not stripped


ldd wrapper_x86_64-pc-linux-gnu_73200
not a dynamic executable


Stderr output

<core_client_version>7.2.42</core_client_version>
<![CDATA[
<message>
process exited with code 2 (0x2, -254)
</message>
<stderr_txt>
Process creation (../../projects/boinc.riojascience.com/wrapper_x86_64-pc-linux-gnu_73200) failed: Error -1, errno=2
execv: No such file or directory

</stderr_txt>
]]>


Could you please compile the wrapper in 64-bit as well as 32-bit and configure it so pure 64-bit systems will work?

Administrador
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 17 Oct 13
Posts: 57
Credit: 0
RAC: 0
Spain
Message 631 - 20 Nov 2014, 8:35:05 UTC - in response to Message 630.
Last modified: 20 Nov 2014, 8:35:23 UTC

You are right, but in some 64 bits systems it works ...
Anyway, it is in things-to-do list.

Best,
R.


Post to thread

Message boards : Number crunching : Venus beta testing


Main page · Your account · Message boards


Copyright © 2017 Knet comunicaciones S.L.