Recently, following the latest round of Microsoft security updates I think, I have been getting errors with hadsm3 4.12 runnung through Boinc 4.45. I use Windows 2000 Pro Sp4 + all security and hotfix updates applied.

The following message was in stderrgui.txt:

***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x0042F342 read attempt to address 0x00000010

1: 05/25/05 13:43:02


Unfortunately these errors have ended the model I was 40% through with a client error.

Any thoughts on how I may correct this? or is a modified version of the software required?

Thanks

Danny

BOINC 4.19 runs very smooth for me on several boxes with Win2000 SP4 with (nearly) all patches.

You will be missing some features of 4.45 but a stable version is worth gold especially for the long running CPDN models.

> BOINC 4.19 runs very smooth for me on several boxes with Win2000 SP4 with
> (nearly) all patches.
>
> You will be missing some features of 4.45 but a stable version is worth gold
> especially for the long running CPDN models.
>
>

I have reverted to 4.25, which I was running OK, but the problem persists:

2005-06-19 12:48:38 [climateprediction.net] Unrecoverable error for result 3vxf_100203167_0 (Incorrect function. (0x1) - exit code 1 (0x1))
2005-06-19 12:48:38 [climateprediction.net] Deferring communication with project for 1 minutes and 0 seconds
2005-06-19 12:48:47 [climateprediction.net] Unrecoverable error for result 35li_200168705_1 (Incorrect function. (0x1) - exit code 1 (0x1))
2005-06-19 12:48:47 [climateprediction.net] Deferring communication with project for 1 minutes and 0 seconds

I also have 4.45 running SETI on a Win2K laptop alright.

Is this a BOINC issue or CPDN?

Danny

You have completed a full model so system appeared stable. Problems started on 19th June. Since you are listed from UK, I am wondering if this could be heat related?

Do you overclock at all? Also have you looked at CPU temp under full load in the current heat?

The only time I've ever seen exit code 1 is after the hadsm3_* controller process terminates and the hadsm3um_* worker process is left running in isolation, so it's worth checking if you've got an orphaned hadsm3um_* process.
<br><a href="http://www.teampicard.net/"><img src="http://www.teampicard.net/images/picardmini.gif"></a><a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3">Join us here</a>

&gt; You have completed a full model so system appeared stable. Problems started on
&gt; 19th June. Since you are listed from UK, I am wondering if this could be heat
&gt; related?
&gt;
&gt; Do you overclock at all? Also have you looked at CPU temp under full load in
&gt; the current heat?
&gt;
I do not overclock the system at all, Athlon XP 3000+, and run a suitable CPU fan, power supply fan and two case fans. However the air temperature is reasonably warm over here at the moment.

Currently the CPU appears to be stable at 70 degrees C (158 F), room temperature around 25-30 degrees C.

The model I am running at the moment, under BOINC 4.25, has not caused any problems, but is under 2% complete, so lets hope it has sorted itself out.

&gt; The only time I've ever seen exit code 1 is after the hadsm3_* controller
&gt; process terminates and the hadsm3um_* worker process is left running in
&gt; isolation, so it's worth checking if you've got an orphaned hadsm3um_*
&gt; process.
&gt; <br><a href="http://www.teampicard.net/"><img> src="http://www.teampicard.net/images/picardmini.gif"&gt;</a><a> href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"&gt;Join
&gt; us here</a>
&gt;

Thanks for this possibility,

task manager shows things running OK at the moment but I have seen one of the processes die in the past - ...is generating a log file... but I'm not sure which one.

The current model is still running OK under BOINC 4.25.

Danny

&gt; &gt; The only time I've ever seen exit code 1 is after the hadsm3_*
&gt; controller
&gt; &gt; process terminates and the hadsm3um_* worker process is left running in
&gt; &gt; isolation, so it's worth checking if you've got an orphaned hadsm3um_*
&gt; &gt; process.
&gt; &gt; <br><a href="http://www.teampicard.net/"><img>
&gt; src="http://www.teampicard.net/images/picardmini.gif"&gt;</a><a>
&gt; href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"&gt;Join
&gt; &gt; us here</a>
&gt; &gt;
&gt;
&gt; Thanks for this possibility,
&gt;
&gt; task manager shows things running OK at the moment but I have seen one of the
&gt; processes die in the past - ...is generating a log file... but I'm not sure
&gt; which one.
&gt;
&gt; The current model is still running OK under BOINC 4.25.
&gt;
&gt; Danny
&gt;

This model has just stopped working,it will probably continue OK after a restart:

2005-06-20 14:50:33 [climateprediction.net] Unrecoverable error for result 0ku8_100047303_0 ( - exit code -5 (0xfffffffb))
2005-06-20 14:50:33 [climateprediction.net] Deferring communication with project for 1 minutes and 0 seconds

The task manager shows boincmgr.exe and hadsm3_4.12_win but it looks like hadsm3um_4.12_w has died.

Danny

I thought Unrecoverable errors usually turned out to be umm Unrecoverable. However, a backup of the BOINC folder from prior to the message can save such situations.

70 Degrees C sounds dangerously hot for a processor to me.

Dust on the heatsink?
Cables blocking the airflow to the cpu?

70C is quite hot. What brand motherboard do you have? Some read higher than others for the same processor, but no matter what, that is very high.

As a system builder, I would suggest that you seriously take another look at the cooling solution you have on that CPU! 70C is way way too high. Try running the machine with the case open and look for snagging cables, dust &amp; detritus. I can guarantee here that this temperature issue, while possibly not being the cause of the problem, is greatly exasipating it!


<img src="http://boinc.mundayweb.com/one/stats.php?userID=444&amp;trans=off">

&gt; 70C is quite hot. What brand motherboard do you have? Some read higher than
&gt; others for the same processor, but no matter what, that is very high.
&gt;

I've given it a good clean out now and is is still running about 60-66 degrees C at a constant 100% CPU, a few degrees cooler when not running CDPN. I'm currently monitoring the CPU temperature.

I have a Gigabyte motherboard (Via KT880), if that makes a difference.

The unrecoverable error was, of course, unrecoverable - My brain must have been on holiday for that post.

Thanks everyone for all your help and suggestions.

Danny

&gt; I've given it a good clean out now and is is still running about 60-66 degrees
&gt; C at a constant 100% CPU, a few degrees cooler when not running CDPN. I'm
&gt; currently monitoring the CPU temperature.
&gt;
&gt; I have a Gigabyte motherboard (Via KT880), if that makes a difference.
&gt;
Gigabyte's do read hotter than most other motherboards. 60-66C is obviously better than 70C, but still warm, even for a Gigabyte. Try Gareth's suggestion of running with the case open and see if that helps. If it does, you can try to see if you can get better airflow through the case.

&gt; &gt; I've given it a good clean out now and is is still running about 60-66
&gt; degrees
&gt; &gt; C at a constant 100% CPU, a few degrees cooler when not running CDPN.
&gt; I'm
&gt; &gt; currently monitoring the CPU temperature.
&gt; &gt;
&gt; &gt; I have a Gigabyte motherboard (Via KT880), if that makes a difference.
&gt; &gt;
&gt; Gigabyte's do read hotter than most other motherboards. 60-66C is obviously
&gt; better than 70C, but still warm, even for a Gigabyte. Try Gareth's suggestion
&gt; of running with the case open and see if that helps. If it does, you can try
&gt; to see if you can get better airflow through the case.
&gt;

My 1900+ is running at about 54-57C at the moment, I consider that on the warm side and indeed, it has locked up on me recently a number of times. If you can, I would consider adding a couple of chassis fans to your setup to try and keep the air inside the case moving. See <a href="http://www.garethlock.shorturl.com/index.htm?boinc">my entry here on keeping your machine cool</a> for more of my tips.


<img src="http://boinc.mundayweb.com/one/stats.php?userID=444&amp;trans=off">

&gt; &gt; &gt; I've given it a good clean out now and is is still running about
&gt; 60-66
&gt; &gt; degrees
&gt; &gt; &gt; C at a constant 100% CPU, a few degrees cooler when not running
&gt; CDPN.
&gt; &gt; I'm
&gt; &gt; &gt; currently monitoring the CPU temperature.
&gt; &gt; &gt;
&gt; &gt; &gt; I have a Gigabyte motherboard (Via KT880), if that makes a
&gt; difference.
&gt; &gt; &gt;
&gt; &gt; Gigabyte's do read hotter than most other motherboards. 60-66C is
&gt; obviously
&gt; &gt; better than 70C, but still warm, even for a Gigabyte. Try Gareth's
&gt; suggestion
&gt; &gt; of running with the case open and see if that helps. If it does, you can
&gt; try
&gt; &gt; to see if you can get better airflow through the case.
&gt; &gt;

Thanks,

I currently have the machine running at 61-63 C and opening the case has no effect on the temperature, so I don't think increased air flow will help.

The current CDPN model I have running appears OK, as I am about 6.5% through under BOINC 4.45.

So I guess this thread should be closed.

Thanks again to everyone who has responed and helped me with this problem.

Danny
&gt;
&gt; My 1900+ is running at about 54-57C at the moment, I consider that on the warm
&gt; side and indeed, it has locked up on me recently a number of times. If you
&gt; can, I would consider adding a couple of chassis fans to your setup to try and
&gt; keep the air inside the case moving. See <a> href="http://www.garethlock.shorturl.com/index.htm?boinc"&gt;my entry here on
&gt; keeping your machine cool</a> for more of my tips.
&gt;
&gt;
&gt; <img src="http://boinc.mundayweb.com/one/stats.php?userID=444&amp;trans=off">
&gt;

If you're still coming up warm, then try going for a bigger heatsink on your processor. I have a block rated for a 3000+ on my 1900+ CPU and it still ran in the mid 60s today. Even my laptop is running hot!! If you're considering swapping the heatsink, go for an all copper model and use a decent silver-oxide paste rather than the cheap pads.

During the heatwave/sticky weather we are having in the UK at the moment, you need to take precautions against trouble like this and early, especially on a normally unattended BOINC machine.

I can't testify as to the stability of v4.12 hadsm, because, at the moment, my 1900+ is still soldiering away on a previous model for v4.10.




<img src="http://boinc.mundayweb.com/one/stats.php?userID=444&amp;trans=off">

Thanks for all these most useful posts + the link to Gareth's page about keeping cool. This has made me wonder whether my machine's repeated failures with boinc cpdn (I reverted to classic) might not, as I thought, be due to the Athlon's way of doing the calculations, but instead be caused by overheating.

I also have a Gigabyte and I dare not publicly reveal what the temp was when it unexpectedly appeared on the screen. Don't think I'm capable of personally doing anything about it, but I'm keeping note of all of this for the future rebuild........
__________________________________________________

FWIIW Danny, the recent Windows security &amp; other updates for my 2000Pro haven't created any problems that weren't already there.
__________________________________________________