New BOINC-project launched, the BBC/Climateprediction "Climate Change Experiment".

[Rattledagger]

News from http://climateprediction.net/

Climateprediction.net, a project involving a group of universities and academic institutions led by Oxford University, in partnership with the BBC, today launches its new climate change experiment. This investigates the uncertainty in the climate of 1920-2080, with a view to understanding the risk of faster or slower future climate change than the current best estimate. Participants in the U.K. will be able to watch a series of documentaries, starting with 'Meltdown' at 2100 on Monday, 20th Feb. on BBC4, in which Myles Allen describes the new experiment. The first results of the experiment will be shown in a documentary in about May. Thanks to the hard work of the climateprediction.net team, people joining the project having seen the BBC publicity will find a much more user friendly download package, graphics, etc. Welcome to new participants and ongoing thanks to existing ones!

News from BOINC:

February 14, 2006
BOINC and Climateprediction.net have joined forces with the BBC to launch a new experiment -- a full simulation of climate change from 1920 to 2080. The experiment is described on a BBC television documentary Meltdown (BBC-4, February 20th, for UK BOINCers). Simplified client software is available at http://bbc.co.uk/climatechange. Experienced BOINCers can attach to http://bbc.cpdn.org (check out the new graphics!). Be warned that a 160-year climate simulation is a long workunit even by Climateprediction.net's extravagant standards. But if you have a fast machine and are prepared to give us first call on your CPU, just for the next couple of months, then you could see your results in a second BBC television programme scheduled for May.

For already existing BOINC-users, just "attach" to http://bbc.cpdn.org/, and make a new account/password as normal. Afterwards, join TeAm AnandTech.
Also, there's a chance BBC/CPDN changes your global preferences, so make sure to check these by going here or change them on another project-page...

For new users, installing the "simplified" BOINC-client will automatically try to "attach" to BBC/CPDN, old-time users can also use this client but remember some features like service-install doesn't work so my recommended is to keep using v5.2.xx or v5.3.16/v5.3.17 if runs alpha-builds.
Remember to join TeAm AnandTech

The BBC/CPDN-model is a 160-year model that will trickle after each year, and if too far off can also be killed by CPDN. Progress is saved each 432 timesteps, and a Trickle is 25920 timesteps, and is worth 226.80 Cobblestones.
A larger file-upload happens each 10 timesteps.
Trickle-speed is comparable to the initial "Slab"-model, so is faster than the Sulphur Cycle Model. Still, with 2.4x longer till each Trickle and 160 instead of 72 Trickles, it's 5x longer than the "slab"model was.

System-requirements, Windows 2000, XP or later, or Linux, with 256 MB free memory, 0.6 GB free hd-space, and a broadband-connection. Deadline is roughly 1 year. The model should now clean-up disk-usage after it runs, and therefore uses less disk-space than the Sulphur Cycle Model.

 

[CupCak3]

pardon my ignorance but how is this different from the other climate DC project?


<--- has no clue

 

[Rattledagger]

The BBC/CPDN-experiment uses the HadCM3L, Transient Coupled Model.

While the old "Slab"-model and Sulphur Cycle Model is using a "static" ocean, a Coupled Model also uses a dynamic ocean. Since ocean responds much slower to changes than the atmosphere, some testers with very fast computers has generated 200 year Spinup-models.

The BBC/CPDN-experiment uses these 200-year Spinup-models + AFAIK the "best" parameters from Slab and Sulphur Cycle, to 1st do a "Hindcast", 1920-2000, and if this is fairly close to historical data the model continues with a "Forecast" of 2000-2080.


Anyway, the "normal" CPDN-project will continue to run generating more Sulphur Cycle models, and possibly even more detailed models than the BBC/CPDN-model will be added here later...

 

[petrusbroder]

Hmm, looks very interesting - deadline 1 year, continuous crunching ...

 

[petrusbroder]

Have joined. The team has now two members: Ingleside and Petrusbroder.
But only Ingleside has delivered one trickle ... more to come soon!
How soon? I do not know - a day or a week?

 

[Rattledagger]

Having just featured on the 6pm BBC news, we are experiencing a lot of activity on the site, please have patience and try again if a web page or download doesnt work! We expect things to continue to be busy for the next few hours.

 

[petrusbroder]

Great - I got in just before the crush!

The site looks great! Somewhat simple - but it is for the masses and not only for the DC-experts. Just great that BBC joins. That is a lot of free marketing - a real coup for CPDN.

The one small complaint I have is the place where all the advanced options are (more or less ) hidden: on the "My details"-page, at the very bottom, just below the area with the map.

The graphics are stunning. If you put in the latitude and longitude of your location you can calculate the model for your grid-call (that is for your area!) that is soo cool!

Edit: If you use the BOINC v. 5.2.xx just attach the usual way (within BOINC), Join the team by looking up the Team Anandtech and clicking on the join-link.
For those who use BOINC v. 4.xx please use the link! This is the link to get a TeAm Anandech account!

 

[TAandy]

Weird!! Don't know if this is related to CPDN??

Thought I'd give it a try, suspended Rosetta(temporarily), attached, ran for a while when boinc suddenly announces its lost its connection to localhost.
Tried to connect again, no joy, tried again, no joy, checked gui_rpc_auth.cfg thingy, checked firewall program settings, disabled firewall (NIS 2005), no joy, uninstalled 2 recently installed programs, nope, reinstalled official boinc, nope, tried system restore, nope again :|
Had to uninstall and then reinstall to get it working again, luckily I had nothing waiting to be uploaded

 

[petrusbroder]

I have run this projects for 9 hours straight - still no trickle - how long does it take?
I like it very much but it is slow going ...
No probelms on my comp at all. Rund well.

 

[TAandy]

21 hours now for me, no trickle here yet either
And, since I reinstalled, no more problems

 

[rise]

is this the only project you all are running on the machine you put it on? or do you have other projects running as well?

 

[petrusbroder]

Well, I run SIMAP on the comp - so it runs SIMAP 12 hours, BBC-CPDN-12 hours. As soon as SIMAP-WUs are done it will 100% BBC-CPDN!

 

[TAandy]

Originally posted by: rise4310
is this the only project you all are running on the machine you put it on? or do you have other projects running as well?

Only one running at the moment.

 

[rise]

well, i'd really like to help this project. i'm in, splitting an x2 between this and simap. should allow me to run 1 core on each after i clear my rosetta.

 

[Assimilator1]

Originally posted by: CupCak3
pardon my ignorance but how is this different from the other climate DC project?

<--- has no clue

Originally posted by: Rattledagger
The BBC/CPDN-experiment uses the HadCM3L, Transient Coupled Model.

While the old "Slab"-model and Sulphur Cycle Model is using a "static" ocean, a Coupled Model also uses a dynamic ocean. Since ocean responds much slower to changes than the atmosphere, some testers with very fast computers has generated 200 year Spinup-models.

The BBC/CPDN-experiment uses these 200-year Spinup-models + AFAIK the "best" parameters from Slab and Sulphur Cycle, to 1st do a "Hindcast", 1920-2000, and if this is fairly close to historical data the model continues with a "Forecast" of 2000-2080.


Anyway, the "normal" CPDN-project will continue to run generating more Sulphur Cycle models, and possibly even more detailed models than the BBC/CPDN-model will be added here later...

I was wondering the same when I heard about that on the BBC earlier ,in fact I thought it was the same project at 1st ,but realised it wasn't.
Thanks for the info RD

Is it run by the same people as CPDN?

 

[petrusbroder]

Originally posted by: Assimilator1
snippet ...

Is it run by the same people as CPDN?

Yes. More fun, better site, nice graphics!

 

[Insidious]

I'm confused.

I don't see any link to create an account in the link RD gave. http://bbc.cpdn.org/

I also don't understand if I am supposed to download using the BIG green button or not. I'm an existing boinc user and just want to create my account and attach.. (Is this wanting to re-install my Boinc?... )

TIA

-Sid

 

[Rattledagger]

Originally posted by: Assimilator1
I was wondering the same when I heard about that on the BBC earlier ,in fact I thought it was the same project at 1st ,but realised it wasn't.
Thanks for the info RD

Is it run by the same people as CPDN?

Yes, it's run by the same people.

The main CPDN-site is currently unreachable, but if remembers correctly that has been done up until now by CPDN is "Stage 1" in their plan, while the coupled model is stage 2 + 3, "hindcast" is 2 and "forecast" is 3.


Now, haven't been keeping-up with CPDN for the last months, but now after BBC/CPDN was launched have seen some hints that the "normal" CPDN-project will also "very soon" start supplying the coupled model...

Haven't seen any definite future-plans, but it's also possible the BBC-project will be fairly "short"-lived by only supplying the current coupled-model for 8 months or something, while CPDN will continue to get more improvements...

 

[Rattledagger]

Originally posted by: Insidious
I'm confused.

I don't see any link to create an account in the link RD gave. http://bbc.cpdn.org/

I also don't understand if I am supposed to download using the BIG green button or not. I'm an existing boinc user and just want to create my account and attach.. (Is this wanting to re-install my Boinc?... )

As an existing BOINC-user, you don't need to download the customized BBC/CPDN-version.

To create an account, you needs to run BOINC v5.2.x or later, if you're still running v4.xx you needs to upgrade 1st. Just download v5.2.13 and install over top, you don't need to uninstall old BOINC-client except if you're still running v4.19 (or earlier)
In BOINC Manager choose "attach to project", and use http://bbc.cpdn.org as url.
When asked, just create new account with email + password of your choise.

When successfully created account, remember to join TeAm AnandTech

 

[Insidious]

Thanks RD

I am such an idiot....

I'm so used to attaching 'using an existing account' that I forgot that option was my very first question when attaching....



-Sid

 

[Insidious]

Originally posted by: rise4310
well, i'd really like to help this project. i'm in, splitting an x2 between this and simap. should allow me to run 1 core on each after i clear my rosetta.

I'm doing the same thing Rise (well, except that I'm not quitting Rosetta )

I am wondering if the time sharing, rather than only using a single core for each all the time will work right since this "WU" goes on for months...

fun to try something this different!

 

[caferace]

Attached on the FX-55. Once I clear the remaining SETI and E@H from this box, we'll see what "happens".

 

[TAandy]

Got some credit eventually

 

[TAandy]

Petrusbroder 226.80 22.15
You're now on the credit map

 

[Insidious]

I've struggled to figure out how to word this question......

ANDY.... be good


How do I tell when I've had a trickle? :shocked:

-Sid