http://www.ssl.berkeley.edu/pipermail/boinc_dev/2004-June.txt

just stumbled accross this conversation, given that you see the error as well, and as it seems on finishing/switching the wu, might be related?

---

--
Florian
www.domplatz1.de

Just got this error a minute or so ago -- wonder if it's related

***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x77FCC022 write attempt to address 0x00000000

---

got this tonight after switch to primegrid:

07.10.2005 03:57:54|rosetta@home|Pausing result 1cfyA_abrelax_08683_0 (removed from memory)
07.10.2005 03:57:54|PrimeGrid|Resuming result t-rsa640_97421_0 using rsa640 version 404
07.10.2005 03:57:55|rosetta@home|Unrecoverable error for result 1cfyA_abrelax_08683_0 ( - exit code -1073741819 (0xc0000005))

---

I have decided that for some as yet unfathomable reason, Rosetta does not play well with other projects - that is to say, it doesn't like to be stopped and restarted by BOINCmanager. If I run Rosetta exclusively 24/7, I have no problem in churning out completed work, even on machines that are on the low edge of capability. I can even totally shutdown and reboot the machines (powerwize) and the project will pick back up and run faultlessly.

The problems appear when one attempts to run this program in a shared environment, and BOINCmanager tries to sequence from Rosetta to another project. Some people have reported that this problem can be minimized by opting for the setting which would leave the program resident in memory when not running, but if the other projects don't need that setting, why should Rosetta?

---

The problems appear when one attempts to run this program in a shared environment, and BOINCmanager tries to sequence from Rosetta to another project.

Not only other projects, but also when boinc does its' weekly(?) benchmarks. Even with the setting to leave in memory set to yes. It appears that boinc removes wu from memory to do benchmarks, even if leave in memory is yes.

---

"Be all that you can be...considering." Harold Green

Hi all,

I am working on this issue. It is a high priority one, along with making the mac app compatible w/ 10.3+

Hi all,

I am working on this issue. It is a high priority one, along with making the mac app compatible w/ 10.3+

We know David, ... :)

We be trying to help ... if we could we would do more ...

One of the reasons that Open Source is so, um, "friendly" sometimes ...

Heck, I would be pawing through it as I am trying to "flesh-out" the developer documentation which is my current main rant this year ... this is the kind of problem that MAY be really an indication that the API documentation, for example, is not clear ...

---

I broke it with my first WU crunched.

https://boinc.bakerlab.org/rosetta/result.php?resultid=298170

<core_client_version>4.19</core_client_version>
<message> - exit code -1073741819 (0xc0000005)
</message>
<active_task_state>1</active_task_state>
<signal>0</signal>
<stderr_txt>

***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x004FF4E3 read attempt to address 0x0A40CFE0

Exiting...

***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x7C911F6C read attempt to address 0x00000001

Exiting...

</stderr_txt>

spidey sense says it's a pointer issue.
maybe there's an issue with saving pointers.. switching projects.. switching back.. and the pointers not pointing to valid locations.
just a wild guess..

I did have to kill boinc_gui the other day to burn a DVD without interrruption.

---

>>> if we could we would do more ...

I'd already offered in this thread.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

On my first computer, Athlon 64 3000+, XP SP2 uptodate, Boinc 5.2.2, Rosetta 4.77, RAM 512Mo, Disk 3Go Free, 3 out of 4 WUs has stopped with this error.

On my second computer, Celeron 450, XP SP2 uptodate, Boinc 4.72, Rosetta 4.77, RAM 384Mo, Disk 1Go Free, 1 out of 1 WU has stopped with the same error.

So I'm stopping crunching until the problem is solved.

---

Maybe related..?

I run BOINC 4.19 and Rosetta 4.78 under Windows 2000 on an old dual Pentium 4 Xeon.

I get computation errors immediately when BOINC starts most Rosetta WUs (in fact I think only one or two ever ran OK) - currently it keeps trying and failing with the same WU. Resetting/updating the project mostly results in the same problem, just with a different WU.

Edit: It now tries a different WU - a new one approx every minute.

More specifically the unrecoverable error is "Incorrect function. (0x1) - exit code 1 (0x1))"

I also run climateprediction, which seems to be a disk space hog - don't know if that's related.

---

If you can, try the new 5+ client.

If you can, try the new 5+ client.

Tried that (with some assorted associated problems - now solved). After aborting the last of the WUs failing under BOINC 4.19 it seems to work - it has now completed 3 in a row.

Thank you!

---

This happend when the kids started the game Black Hawk Down.

https://boinc.bakerlab.org/rosetta/result.php?resultid=373341

Anders n

---

Are we talking about "Wargames" here :) ?

---

Are we talking about "Wargames" here :) ?

Soohting in first person game - Yes.

But it was something in the game that made it happen.

Anders n

---