Boinc seems to be stomping on the same spot all the time... I've had 2 computers running CP and one of them (the slower of the two) is now about 10 days ahead of the other one...

Noticed that the one that is behind said earlier today that it was at 35.54% completed, a few minutes ago when it reactivated it said 35.53% and the CPU-time has backed off about an hour...

---

<img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=978" />

when it reactivated it said 35.53%

Whenever started CP starts from the last multiple of 144 timestep so it is usual to loose an average of 72 timesteps whenever you exit BOINC. If possible just suspend BOINC and set the preference to keep in memory while suspended, to minimise this loss.

Unless you turn off the faster machine very frequently, that probably wont account for being 10 days behind. It is possible that the faster machine encountered some strange climate conditions and decided to rewind a little to make sure whether it is the model or a calculation error. It should rewind to the last 144 multiple timestep then a model month then a model year.

&gt; when it reactivated it said 35.53%

That is correct!

&gt; Whenever started CP starts from the last multiple of 144 timestep so it is
&gt; usual to loose an average of 72 timesteps whenever you exit BOINC. If possible
&gt; just suspend BOINC and set the preference to keep in memory while suspended,
&gt; to minimise this loss.

Both machines have been running 24/7 since the beginning of time B-)

&gt; Unless you turn off the faster machine very frequently, that probably wont
&gt; account for being 10 days behind. It is possible that the faster machine
&gt; encountered some strange climate conditions and decided to rewind a little to
&gt; make sure whether it is the model or a calculation error. It should rewind to
&gt; the last 144 multiple timestep then a model month then a model year.

Well, a little is not the correct description IMHO... Usually the faster one crunches 4 seti WU's while the slower one makes 3.

The slower one is currently at 31 days 8 hours and the faster at 19 days 2 hours, when I wrote my original message it was at 19 days 1 hour, only 1 hour more cpu registered for a whole day, I switch application only once/hour so it should have gained atleast 3 hours by now...

---

<img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=978" />

&gt; &gt; when it reactivated it said 35.53%
&gt;
&gt; That is correct!
&gt;
&gt; &gt; Whenever started CP starts from the last multiple of 144 timestep so it
&gt; is
&gt; &gt; usual to loose an average of 72 timesteps whenever you exit BOINC. If
&gt; possible
&gt; &gt; just suspend BOINC and set the preference to keep in memory while
&gt; suspended,
&gt; &gt; to minimise this loss.
&gt;
&gt; Both machines have been running 24/7 since the beginning of time B-)
&gt;
&gt; &gt; Unless you turn off the faster machine very frequently, that probably
&gt; wont
&gt; &gt; account for being 10 days behind. It is possible that the faster machine
&gt; &gt; encountered some strange climate conditions and decided to rewind a
&gt; little to
&gt; &gt; make sure whether it is the model or a calculation error. It should
&gt; rewind to
&gt; &gt; the last 144 multiple timestep then a model month then a model year.
&gt;
&gt; Well, a little is not the correct description IMHO... Usually the faster one
&gt; crunches 4 seti WU's while the slower one makes 3.
&gt;
&gt; The slower one is currently at 31 days 8 hours and the faster at 19 days 2
&gt; hours, when I wrote my original message it was at 19 days 1 hour, only 1 hour
&gt; more cpu registered for a whole day, I switch application only once/hour so it
&gt; should have gained atleast 3 hours by now...
&gt;

Ohh... Forgot to say, I turned the "leave in ram" feature off since it seems to generate som nasty "fuck up's" in memory, the crashfrequence went from "always" to "almost never" when using other apps at the same time...

---

<img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=978" />