I believe in this project, I really do. I think Global Warming is the #1 threat to civilisation.

Therefore I\'m really disappointed with the result of all the time I have spent crunching for CPDN. The model I had been working on since last year failed with an error a few days ago. It was only 20% done but I had just built a new PC which was going faster so the failure was a bit gutting.

Here is the \'result\' including the error message:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=5695358

Another user told me on the Seti board that my work will not be wasted. Can anyone confirm this is the case? It will make me feel a little better.

Also, is there any other way of contributing to the project that doesn\'t take months of processing? My PC is not on 24/7 so I need shorter work.

I really want a way to help this project.

---

The improvement in the last couple of trickles is very obvious, as it was running at pretty much double the speed after the rebuild (2 sec/timestep).

The simulation uploads some data every model year and more every decade (and even more every 40 years). Your model had reached 1953, so it will have made three decade uploads, which will be used by the project.

The models are rather long and the best way to improve your completion chances is to read the posts on avoiding crashes (e.g. this one) and to take backups occasionally (e.g. how to).

The project is developing new versions of the simulator that will run shorter work units, but they\'re not available yet.

I hope you stick with it, despite the annoyances: it is a worthwhile effort.

It\'s good to know that all the effort was not wasted.

I had backed up a few days ago prior to transitioning to my new machine. I have read the posts about restoring back-ups but I don\'t feel too confident as I run multiple projects and have since attached to another project since the crash.

Since my client has reported the work back now doesn\'t it mean it\'s too late anyway?

I\'ll research BOINC too get meself more informed then I\'ll probably try another model when I have got some enthusiasm back..

Thanks for the info.

BTW, how do I get my timestep info? I could only see my average timestep time.

---

1) Restoring a backup: BOINC Wiki.

Basically:
Set all projects for no new work.
Suspend the faulty climate model. (In the menu.)
Run down all other projects.

Restore backup.
Re-boot to get rid of anything in memory. (Memory data for cpdn will be faulty.)

Start running again. The uploads for the other projects should be ignored by their servers. (You may get error messages.)

2) The error messages on the cpdn server page for the model will stay there, and you\'ll get a nasty rejection message when/if you complete the model, but the model\'s data will still upload OK, as this project ignores the server messages.

3) The s/TS (seconds per TIME STEP), is an average for that period of time that BOINC trys to calculate as it goes.
Both this and the timestep info is on the graphic display for the model, and on the page for the model in the trickle info listing.

Timestep ......CPU Time (sec) ......Average (sec/TS)
1 855,360 .....3,403,888 .............3.9795
1 829,440 .....3,351,879 .............4.0411

I think Iain must have done some quick mental arithmetic. Here are your last two trickles. Rather more than 25,000 timesteps between the trickles, done in rather more than 50,000 seconds. Equals about 2.

If you don\'t restore the backup for a while, it would be worth checking before you do so that the model hasn\'t been given out to another computer (you wouldn\'t want to duplicate efforts). If you click on the workunit number, you see which computers it\'s been sent to ie whether your computer is still the only one to have received it.

Check this with a couple of your previous models which were sent out again; another person\'s computer is also listed.

---

Cpdn news

I believe in this project, I really do. I think Global Warming is the #1 threat to civilisation.

Therefore I\'m really disappointed with the result of all the time I have spent crunching for CPDN. The model I had been working on since last year failed with an error a few days ago. It was only 20% done but I had just built a new PC which was going faster so the failure was a bit gutting.

Here is the \'result\' including the error message:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=5695358

Another user told me on the Seti board that my work will not be wasted. Can anyone confirm this is the case? It will make me feel a little better.

Also, is there any other way of contributing to the project that doesn\'t take months of processing? My PC is not on 24/7 so I need shorter work.

I really want a way to help this project.

Can understand how frustrated you feel but please don\'t give up, the first few years are important. I\'m sure that soon someone much more knowledgeable than me will be along to chat to you:)I\'ve had a number of models crash over in the Beeb CCE and it\'s disappointing but you kind of get used to it and keep rooting for the ones that look like making the finish line:)

A team mate of mine ( The Greenies - that\'s us) has started a shorter model on
the World Community Grid. There\'s a lot of talk about shorter models over here
but from what I\'ve seen there\'s nothing definite yet....


Hang in there and good luck.

Chrissy

---

IainsstatspageforTGP

...Since my client has reported the work back now doesn\'t it mean it\'s too late anyway? ...

The server will keep reporting the first error it receives and may give some rather alarming error messages when the model finishes, but it doesn\'t matter: any repeated trickles from a restored model are just ignored, and new trickles are accepted as if there had been no crash.

I restored my back-up this morning. My stats tab on the BOINC app has gone wonky but I\'m hoping it will sort itself out once all the projects communicate with the servers.

I\'m keeping my fingers crossed...

---