climateprediction.net home page
Not finishing phase 1

Not finishing phase 1

Message boards : Number crunching : Not finishing phase 1
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user1098
Avatar

Send message
Joined: 25 Aug 04
Posts: 41
Credit: 34,555
RAC: 0
Message 7016 - Posted: 18 Dec 2004, 17:47:08 UTC

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/trickle.php?hostid=24450

my laptop doesn't mamage to get through phase 1 of any model. The most it got is a 54010 trickle. What can I do about it?
<br>***********************************<br>
WinXP Home, P4 3.2GHz HT, 512Mb RAM. Running SETI, CPDN, Predictor, LHC, Einstein, Orbit, Lattice and Folding
ID: 7016 · Report as offensive     Reply Quote
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2185
Credit: 64,822,615
RAC: 5,275
Message 7017 - Posted: 18 Dec 2004, 18:18:07 UTC - in response to Message 7016.  
Last modified: 18 Dec 2004, 18:20:38 UTC

&gt; http://climateapps2.oucs.ox.ac.uk/cpdnboinc/trickle.php?hostid=24450
&gt;
&gt; my laptop doesn't mamage to get through phase 1 of any model. The most it got
&gt; is a 54010 trickle. What can I do about it?
&gt;
Some laptops can overheat pretty easily. Overheating can cause model instability. Tilting up the back edge to allow more airflow, or buying a laptop cooler can help, as talked about in this thread.

http://www.climateprediction.net/board/viewtopic.php?t=1954

Looking at that PC, it also appears that it only has 128 MB of memory. Running Windows 2000 and the model probably uses all the memory easily, and then has to work a lot with a swap file. More memory would no doubt help.
ID: 7017 · Report as offensive     Reply Quote
old_user23880
Volunteer tester

Send message
Joined: 10 Oct 04
Posts: 223
Credit: 4,664
RAC: 0
Message 7027 - Posted: 20 Dec 2004, 7:34:01 UTC

If you've sorted out the points Geophi mentions, you should then look at the numerous postings on the other Boinc message board about lost models.
__________________________________________________

ID: 7027 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 7044 - Posted: 21 Dec 2004, 9:44:34 UTC

With less than the minimum recommended memory, data will keep getting cached onto the hard disk.
Very frequently! This will lead to more wear than is good for it. The normal data store interval
is 60 seconds, which is a compromise between excessive wear, and possible data loss.
And you are using the disk for caching as well. Software these days is memory hungry, so if
at all possible, add some more. First though, find out how fast what you have is.
It may be better still to replace what you already have, (and add more), with faster ram.
Faster ram gives the cpu quicker access to the data, which will mean faster processing of your models.

Les
ID: 7044 · Report as offensive     Reply Quote
old_user1098
Avatar

Send message
Joined: 25 Aug 04
Posts: 41
Credit: 34,555
RAC: 0
Message 7045 - Posted: 21 Dec 2004, 10:21:17 UTC

Actually, it's a Pentium III 900MHz laptop (a Dell Inspiron 8000 I bought in 2001) sitting there doing nothing but CPDN, SETI (and LHC) work most of the day (I turn it off at night). I am a bit short of cash right now, so I don't know if and when I'll be able to get more RAM for it.
Will increasing the disk space allotted for BOINC help or not? Paging file is currently 125Mb.
Thanks for helping...
<br>***********************************<br>
WinXP Home, P4 3.2GHz HT, 512Mb RAM. Running SETI, CPDN, Predictor, LHC, Einstein, Orbit, Lattice and Folding
ID: 7045 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 7060 - Posted: 24 Dec 2004, 18:13:41 UTC

I'm not an expert, but I don't think extra BOINC space will help. Are you running the GUI version?
If so, look at 'Disk'. This will show you if there is enough space for the 'permanent' data.

The problem is more likely 'temporary' data storage. (More fingers on which to count.)
So what's needed is a larger cache and/or page file (not sure if these are the same thing.
The name seems to change with the operating system.)

And the problem may be elsewhere. But one thing at a time.

Another thought: Do you run the machine continuously? It may not be running for long enough periods.

And have you tried Geophi's suggestion to lift the machine above the support surface?

Les

ID: 7060 · Report as offensive     Reply Quote

Message boards : Number crunching : Not finishing phase 1

©2024 cpdn.org