I seem to constantly loose work units due to computation errors or model crashes on one of my computers, an Athlon XP 3000+. At first I thought the problem was overheating, but since everything else runs stable I'm beginning to doubt that. I've checked the ram with Memtest86.
It has been running stable for the past months, but about one month ago it started crashing on cpdn. Now I almost can't run for an hour before the model crashes.
Is this to be expected with such a complex model, or is it just my pc that is weird?

Happy new year :-)

---

Proud owner of the CPDN Wow-Mug!

> Is this to be expected with such a complex model, or is it just my pc that is
> weird?

There are other tests you can try - UK_Nick suggests some <a href="http://www.climateprediction.net/board/viewtopic.php?t=2126">here</a>. I would not suggest that it is 'to be expected'; I've been running an XP 3200+ without problems with CPDN in BOINC since August, and before that with the classic version. Plenty of others have been using Athlons for CPDN without difficulty, though some people have had problems with almost anything.

If the machine passes the other tests, then was there any pattern to the failure under CPDN? Were you using the graphics/screensaver?

---

&gt; I seem to constantly loose work units due to computation errors or model
&gt; crashes on one of my computers, an Athlon XP 3000+. At first I thought the
&gt; problem was overheating, but since everything else runs stable I'm beginning
&gt; to doubt that. I've checked the ram with Memtest86.
&gt; It has been running stable for the past months, but about one month ago it
&gt; started crashing on cpdn. Now I almost can't run for an hour before the model
&gt; crashes.
&gt; Is this to be expected with such a complex model, or is it just my pc that is
&gt; weird?
&gt;
&gt; Happy new year :-)
&gt;
You also might want to try the basic hardware maintenance thread suggestions here

http://www.climateprediction.net/board/viewtopic.php?t=2124

I have followed all the suggestions from both tests, and have run Prime95 for a few hours without errors. I will tr and let it run for a full day soon and see if it changes anything.
I've not been able to see any pattern in the cpdn crashes, most of the day the pc runs with blank screensaver and when I check it, the model has crashed.

@Andrew Hingston, how warm is your cpu when running cpdn? Perhaps overheating is indeed the problem here, since it gets about 54C when running.

The only time i've had one model crash on me was when i was running a few other DC projects at the same time. Since then i run cpdn and 1 other and have had no problems.

AMD64@2.3

---

&gt; @Andrew Hingston, how warm is your cpu when running cpdn? Perhaps overheating
&gt; is indeed the problem here, since it gets about 54C when running.
&gt;

This doesn't seem likely to be the problem. My Abit monitor is reporting a temp of 49 degrees with a big Coolermaster CPU cooler in place of the standard one. But these figures are not very reliable, so a difference of 5 degrees may be within the margin of error, and Athlons are reasonably tolerant of high temperatures.

---

This is indeed what I've been told too. Some people have their cpu (same type) running at close to 60C without problems. I'm beginning to think this is either a windows problem or a bad cpu...
Also, the computer has a bad habit of spontaneously restarting every 3-8 hours or so. I'm suspecting this is only happening when running cpdn, I'll have to investigate this further.

Hi Kenneth

Have a look at the boinc Questions and Problems message board, the Windows section. Go to the 'I've had it' thread, and also look at the other thread further down that is referred to.

---

__________________________________________________

After having run some tests and read the Questions and Problems board, I've come to the conclusion that the problem IS overheating. I've now had the pc running with the case plates off for several days without a single error or restart. But as soon as I put them on, it runs from 1 minute to a few hours and then crashes. Now I just have to find out how to get more airflow and still make it run silently (and not spend any money ;-) ).

Thanks for your help everyone!

Then the question is what is overheating? It isn't neccessarily your CPU. It could be one of the chipset chips on the motherboard or your RAM. I was having similar problems to yours. I could run seti@home and LHC all I wanted with no ill effects but as soon as I started CPDN I had to severely underclock the system to keep it stable. In my case it seems to have turned out to be a bad stick of RAM. Both memtest86 and prime95 gave me clean bills of health but one day the stick of RAM finally gave up the ghost and refused to work at all. Since taking it out, the system has become more stable and a few other small quirks have gone away. Now if I can just get newegg to take back the RAM...

---

Today I bought some ram cooling sinks, I'll see if that changes anything.

&gt; Today I bought some ram cooling sinks, I'll see if that changes anything.
&gt;
Maybe a few of the tips here could help keep your case cooler. There may be better articles out there on how to do it and you could find them with google.

http://www.endpcnoise.com/cgi-bin/e/computercooling.html

Finally my computer seems to be running stable and is happily crunching on a cpdn work unit, so far without any errors or crashes. The solution was to mount a cooling fan inside the case to maintain flow throughout, and perhaps also the ram cooling I bought.
To top it all, I just ordered some Arctic Silver thermal paste.

I thank you all for your suggestions and info, it was a great help! :-D

---

Proud owner of the CPDN Wow-Mug!

CPDN goes down as the toughest DC project on pc's I've run with the possible exception of Folding. CPU temps run 4 - 5C warmer than the easiest of the lot which seems to be SETI. It also begs to ask the question whether we're contributing to global warming :) Does the model account for it? :)

---