I had a question about the new Attribution project, but haven\'t joined the project yet as I needed an answer before I did!

No worries, I\'ll jump into the forum here and ask - no trouble - errrm, wrong!

I tried to log in with my username and password and it said I couldn\'t. So I guessed I\'d forgotten my password, (unlikely), and tried the password recovery. This failed and gave me this...

Sorry, but your password cannot be retrieved because your account is currently inactive. Please contact the forum administrator for more information.

... so I am!

In case anyone was curious, I wanted to now the deadlines on the Attribution wu\'s. I see they take 3-4 weeks on a P-IV but am guessing that this figure is with 100% CPU quota, this will not be possible here, I have 10 projects running on this machine and 4 on my other. I have no wish to download work and then have to junk it.

Oddly, I didn\'t have any trouble getting into this board, which I assume uses the same login as the other?

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

Deadline is 1 year. I guesstimate your PC will need about 550 hours of crunching to complete the work unit (~3 weeks of solid crunching)?

I should point out that the model takes much more memory than the normal CPDN models, moving between 200mb and 450mb as it cycles through each day, so you should take that into account when looking at the balance of projects running on your PC.


I believe the forum on that site is linked to the users attached to the project, so you\'d need to attach before you can post.

---

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

Thank you for that. I\'ll find an envelope to do some calculating on the back of, and see if I can contribute a little!

You misunderstand, the forum I could not get into was the one here, not at the new project.

This one.

I should add, I wondered if I needed to register for that forum seperately, but trying to create an account told me the name and e-mail address were already in use.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

Adrian
These \'help boards\' belong to the cpdn project. The \'Message board\' that you tried to get to, whilst being associated with these help boards and cpdn, is a totally separate entity, hosted by a different university, in a different city. It has it\'s own user name / password setup. (Very handy when this uni is down for maintenance, and vice versa.)

And the new project is separate again.
I think they want results asap for this one, not in a couple of years.

What user name did you try to use on the Message board?
I\'ll see if it can be activated.

What user name did you try to use on the Message board?
I\'ll see if it can be activated.

The usual, \"adrianxw\". I use the same handle on all the DC projects, plus all but one of the other fora I use on the net. No rush. Thanks Les.

And the new project is separate again.
I think they want results asap for this one, not in a couple of years.

Yeah, I simply don\'t have the resources at this time.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

OK Adrian, you should be in business on the message board.
I have a problem fixing this sort of thing, (very strange), but crandles took care of it.

From the Seasonal Attribution project\'s FAQ:

Quote
How long will this project be active for?

This is is relatively short term project and we expect it to have a lifetime of only a few months. This is because we are investigating the change in a seasonal timescale weather event, as opposed to the change in century timescale climate (see – How is this project different from the other climateprediction.net experiments?).

Our current target number of total simulations is 10,800. You can see how many simulations are left to send out in the database on the server status page.

In the light of the Les\'s finding (and also taking into account the large amount of memory the attribution model uses) I feel there are two main choices:

* Run the attribution project exclusively (or at least a 30% share) for the 3 weeks it\'ll take to run, then concentrate on your other projects to make up. The advantage to running at 100% for a shorter period is that if the other projects aren\'t in memory when the attribution project is running you shouldn\'t have an memory issue.
Also note that the checkpoints are relatively infrequent (once every 80-90 minutes at your processor speed) - if you select \'remove from memory\' in order to save memory while it\'s running, you\'ll lose on average 40 minutes of processing time each time the process is discarded from memory. In fact, if you switch away every hour, it\'ll never reach a checkpoint at all if the discard setting is active.

or

* Stick with a more long lived climate project

---

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