What a waste of time and resources!


http://climateapps2.oucs.ox.ac.uk/cpdnboinc/workunit.php?wuid=6135197

---

Different workunits have different amounts of replication, depending on the scientists requirements. In some cases they want to see how different architectures / operating systems compare when running the model. If you look at workunits which have all been processed, you\'ll see that on average only one model completes - this is mostly due to a subset of machines which \'trash\' models.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

There may well be more than 10.
If the researchers saw something interesting in a previous run, that perhaps didn\'t finish, they may have issued that data set again with 10 lots, so as to ensure that at least one got through.

If one isn\'t a physicist, phd student, or other researcher using the data for some purpose, then there\'s no point in trying to guess why models are issued the number of times they are, or with the starting parameter values they have.

We\'re just lab assistants, who have volunteered to crunch whatever models are on offer.

Hi FL4I

Here\'s the WU which unfortunately exemplifies exactly what Mike said about a subset of computers/users who trash models

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/workunit.php?wuid=6135197

#1 is you.
#2 is a new member. New members usually crash a few models while they learn what to do & what not to do.
#3 has previously crashed 48 models.
#5 has had 159 previous models.
#6 is new and has already crashed one model.
#7 is newish.........
#8 has crashed 8 models with curable errors but hasn\'t posted for help.
#9 identical situation to #8
#10 has had 31 curable crashes but AFAIK hasn\'t posted on the forum.

The mods could of course start sending private messages to at least some of these people, offering help. But I reckon that many would consider this intrusive, plus the mods need to have a life.

So CPDN needs you to complete this model if you possibly can.

---

Cpdn news

Hi FL4I

Here\'s the WU which unfortunately exemplifies exactly what Mike said about a subset of computers/users who trash models

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/workunit.php?wuid=6135197

#1 is you.
#2 is a new member. New members usually crash a few models while they learn what to do & what not to do.
#3 has previously crashed 48 models.
#5 has had 159 previous models.
#6 is new and has already crashed one model.
#7 is newish.........
#8 has crashed 8 models with curable errors but hasn\'t posted for help.
#9 identical situation to #8
#10 has had 31 curable crashes but AFAIK hasn\'t posted on the forum.

The mods could of course start sending private messages to at least some of these people, offering help. But I reckon that many would consider this intrusive, plus the mods need to have a life.

So CPDN needs you to complete this model if you possibly can.

Fair enough, I just don\'t like to see waste but, I can see your problem.

Thanks for the positive replies.

---

In the News thread posts I put regular reminders for members to post on a forum if their models keep crashing, or refer to the project READMEs which offer really detailed and CPDN-focussed advice. But members who don\'t visit the forums don\'t see these reminders.

Crashed HADCM models will often be completed on another computer from a restart dump part-way through, so this reduces some otherwise wasted crunching.

---

Cpdn news