Impossibly high at save 8975989065225897796,11633537,500,14,5; returning 204 error
log in

Advanced search

Message boards : Number crunching : Impossibly high at save 8975989065225897796,11633537,500,14,5; returning 204 error

Author Message
Conan
Send message
Joined: 10 Mar 15
Posts: 8
Credit: 48,848
RAC: 0
Message 9 - Posted: 11 Mar 2015, 0:12:47 UTC

I am getting this error on all Windows work units so far, am I missing something?

See this example

Conan

ChertseyAl
Send message
Joined: 9 Mar 15
Posts: 4
Credit: 113
RAC: 0
Message 13 - Posted: 11 Mar 2015, 8:32:00 UTC - in response to Message 9.

Same here - Every WU fails in the same way.

Conan
Send message
Joined: 10 Mar 15
Posts: 8
Credit: 48,848
RAC: 0
Message 15 - Posted: 11 Mar 2015, 9:49:53 UTC
Last modified: 11 Mar 2015, 9:51:06 UTC

Also all my Linux work units get marked invalid against two Windows machines even though the Linux work unit processes and is returned successfully but does not validate.

Set to No New Work as nothing is working for me.

Conan

Profile Daniel
Project administrator
Send message
Joined: 5 Mar 15
Posts: 73
Credit: 162,134
RAC: 0
Message 17 - Posted: 12 Mar 2015, 0:44:57 UTC - in response to Message 9.

I think the problem might be that you need a lot of memory to run this program. The lengths of the arrays needed to be stored are the numbers in parentheses. Your computers might not have enough free space to use these arrays, which are needed for this program to run properly.

Conan
Send message
Joined: 10 Mar 15
Posts: 8
Credit: 48,848
RAC: 0
Message 20 - Posted: 12 Mar 2015, 7:22:48 UTC

That maybe true but I don't know the memory requirements for this application so can't say.
I have 4 computers, 2 Windows XP and 2 Fedora Linux . The Windows computers are 32 bit so have less than 4 GB of RAM available for use. One of the Linux computers also has 4 GB but is 64 bit so is handling the memory requirements, the other Linux computer is also 64 bit and has 8 GB of RAM.
The Linux tasks complete fine with no errors but they do not validate against the Windows computers they are teamed with.

As an aside the above issue it appears that the same work unit is often resent back the computer that has already processed it instead of to another computer.

Thanks

Conan

Dr Who Fan
Avatar
Send message
Joined: 12 Mar 15
Posts: 18
Credit: 4,613
RAC: 0
Message 23 - Posted: 12 Mar 2015, 17:35:57 UTC

I had one with a similar error message:
Task 628882

Stderr output

<core_client_version>7.2.42</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 204 (0xcc)
</message>
<stderr_txt>
Impossibly high at save 8975989065225897796,54845173,500,14,5; returning 204 error

</stderr_txt>
]]>


Run Time: 34 min 27 sec
CPU Time: 33 Min 01 Sec
Peak Virtual Memory Use: 420.16 MB
Peak RAM Memory Use: 421.63 MB

I think we can rule OUT excessive RAM use


Without seeing the source code, my next guess is the array size(s) may not set up correctly - maybe they need to be larger.
____________

Profile Daniel
Project administrator
Send message
Joined: 5 Mar 15
Posts: 73
Credit: 162,134
RAC: 0
Message 24 - Posted: 12 Mar 2015, 19:47:02 UTC - in response to Message 23.
Last modified: 12 Mar 2015, 20:14:41 UTC

I've found that the array are too large. They are of length 16 quintillion and it would be helpful to change them to length 4 billion. We only need length 60 million right now. There are two solutions to this problem: to edit the program to make it memory-efficient or to change the array's length.
____________
Daniel Monroe
vdwnumbers.org Project Administrator

Profile Daniel
Project administrator
Send message
Joined: 5 Mar 15
Posts: 73
Credit: 162,134
RAC: 0
Message 28 - Posted: 13 Mar 2015, 2:31:38 UTC - in response to Message 24.

Version 2.1 is released which uses half the memory of earlier versions.

Please check whether the "Impossibly high ... 204 error" goes away.

ChertseyAl
Send message
Joined: 9 Mar 15
Posts: 4
Credit: 113
RAC: 0
Message 29 - Posted: 13 Mar 2015, 11:56:11 UTC - in response to Message 28.

Still all failing for me, e.g.

http://www.vdwnumbers.org/vdwnumbers/result.php?resultid=678964

<message>
- exit code 204 (0xcc)
</message>
<stderr_txt>
Impossibly high at save 8976916934985515008,63936569,500,63937069


2GB RAM, only one task running at a time.

Alez
Send message
Joined: 13 Mar 15
Posts: 3
Credit: 33,169
RAC: 0
Message 30 - Posted: 13 Mar 2015, 14:40:01 UTC

my error is

Stderr output

<core_client_version>7.4.34</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code -1073741819 (0xc0000005)
</message>
<stderr_txt>


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x0035BCF2 write attempt to address 0xFAF750C8

Engaging BOINC Windows Runtime Debugger...

zombie67 [MM]
Avatar
Send message
Joined: 9 Mar 15
Posts: 15
Credit: 1,283,842
RAC: 0
Message 31 - Posted: 13 Mar 2015, 15:22:56 UTC

I am getting the same, but on only one machine.

Is there a particular requirement for this app? A version of .Net or Java or something?
____________
Dublin, California
Team: SETI.USA

Chris Granger
Send message
Joined: 12 Mar 15
Posts: 1
Credit: 144
RAC: 0
Message 32 - Posted: 13 Mar 2015, 22:22:37 UTC
Last modified: 13 Mar 2015, 22:27:29 UTC

Three 260.01 WU's have failed for me:

http://www.vdwnumbers.org/vdwnumbers/result.php?resultid=713497
http://www.vdwnumbers.org/vdwnumbers/result.php?resultid=713602
http://www.vdwnumbers.org/vdwnumbers/result.php?resultid=709247

Conan
Send message
Joined: 10 Mar 15
Posts: 8
Credit: 48,848
RAC: 0
Message 34 - Posted: 14 Mar 2015, 1:46:21 UTC

I have re-enabled work fetch and the new tasks are doing the same as the old ones.

The work units appear to use from 380 MB to 570 MB per task so memory should not be an issue.

Same error on Windows.

Same problem on Linux, run to a successful completion but the results are not compatible with Windows results.

With there being far more Windows computers running this project my Linux results always have a Windows wingman, the results don't agree and my Linux results are then marked invalid when the third computer (always a Windows one) agrees with the other Windows computer.

This then leads to the conclusion that I am unable to participate in this Project as my Windows 32 bit XP computers seem to be too old and my Linux 64 bit computers don't generate the same results as the Windows computers so I am just wasting my time, computer usage and electricity.

Good luck with your project Daniel, you have done very well setting up a new project that runs quite well from the start, it is just apparently not for me.

Conan

Alez
Send message
Joined: 13 Mar 15
Posts: 3
Credit: 33,169
RAC: 0
Message 35 - Posted: 14 Mar 2015, 1:56:39 UTC
Last modified: 14 Mar 2015, 2:04:57 UTC

Conan,
I am getting the same error on win xp 32 bit but not on win 7 64 bit and 32 bit, so for me it seems the impossibly high error is xp related.
It's a pity for testing new projects that wingmen can't be designated as I could match my Linux systems with yours and we might get some valid work. At the moment I only have my win systems attached.

Profile rebirther
Avatar
Send message
Joined: 9 Mar 15
Posts: 3
Credit: 137
RAC: 0
Message 38 - Posted: 14 Mar 2015, 7:58:08 UTC

My error rate is very high, around 90% with computation errors on win7 64bit.

Werinbert
Send message
Joined: 11 Mar 15
Posts: 4
Credit: 100,186
RAC: 0
Message 39 - Posted: 14 Mar 2015, 8:31:00 UTC

The v250.xx tasks ran fine previously. The v260.1 tasks, however, are generating errors but only about 40% error the rest are valid. I am running a Win7 64bit machine.

fzs600
Send message
Joined: 10 Mar 15
Posts: 4
Credit: 267,826
RAC: 0
Message 40 - Posted: 14 Mar 2015, 9:06:17 UTC - in response to Message 39.

I temporarily stopped the project

Sorry

Profile Daniel
Project administrator
Send message
Joined: 5 Mar 15
Posts: 73
Credit: 162,134
RAC: 0
Message 45 - Posted: 14 Mar 2015, 21:51:14 UTC - in response to Message 40.

I am sorry you are having to detach from the project. I need your help to sort out these errors. On the machines where you are getting errors, please download the boinc code from git and replace samples/example_app/uc2.cpp with either of the two files below. The first version is the older one (25xxx) that uses twice as much memory. At line 266, the problem is that the values in array bestPrimes may be corrupted. If you are able to change the code so this does not happen, the problem will be solved.

Daniel


Version with powers array as "long long"

https://raw.githubusercontent.com/hmonroe/vdw/master/VDWDaniel/src/uc2.cpp.longlong

Version with powers array as "long"

https://raw.githubusercontent.com/hmonroe/vdw/master/VDWDaniel/src/uc2.cpp.long

Conan
Send message
Joined: 10 Mar 15
Posts: 8
Credit: 48,848
RAC: 0
Message 47 - Posted: 14 Mar 2015, 23:33:06 UTC

Yes alez [TSBT's Pirate], it would appear that Windows XP is not supported, perhaps it is a .NET 4.5 thing as on physiome, where the newer version of .NET is required and is not compatible with XP.

As for my Linux computers they run all of the versions of the applications successfully but get marked as invalid as they do not validate with Windows computers.

With the new 260.01 application my Linux computers run this fine but with all other computers being Windows and most of them failing this application the error rate (or 2 successful Windows results) will again mark me as invalid when I am not.

Conan


Post to thread

Message boards : Number crunching : Impossibly high at save 8975989065225897796,11633537,500,14,5; returning 204 error


Main page · Your account · Message boards


Code and content created by Daniel Monroe © 2018.