Just finished my first workunit which took my computer 73 hours to complete, and now I see i recieved the "glorious" amount of 2057 credits for that. Well, I can definately understand why participants on this project needs to be dedicated...

Some of the fun in participating in BOINC projects is to see if you can climb the charts by chrunshing away, but the credit-policy in CPDN is simply lousi, and there are many other interesting project I rather will use my cpu-time on. Detaching!!

Be patient, the cpdn servers only run some of the scripts about once a week. Besides,most contribute to this project because they believe in it, not because of credits.

Dave

With GPU projects offering the possibility of a million credits per day, it really is futile for a CPU project to attempt to compete BOINC-wide.

You can, of course, still climb the charts within CPDN by continuing to crunch, increasing and optimising your hardware. However, as Dave suggests, this is a project with a dimension other than credits. Some of the models are quite long and require a substantial commitment; other models may need a bit of attention for one reason or another. So the project may not suit someone who's out to make a good impression as rapidly as possible rather than make a good contribution to a long-running science project.

I run CPDN because I want help in a small way. There's no more worthwhile scientific goal than investigating the fate of Earth, and CPDN has made critical contributions to this investigation.

I once mentioned, when no work was available, that this was actually a real achievement for the project in terms of commitment and effort and that how many other research projects could boast of having 'too many' resources.
It's clearly nice to have one's contribution recognised with credits, but I, for one, am not in it for fame or glory. I'd run these programs with or without credits and I suspect that so would the majority of CPDN contributors.

---

But since these are ( at least as far as I know ) the most timeconsuming BOINC work-units, it would, in my opinion be evident to do a little "extra" to attract more volounteers especially since it is not possible to develop gpu applications for the project.

It is not a matter of "to make a good impression as rapidly as possible" - this would be for those with several computers which I don't have.

There are other projects out there which applies to the "fate of the earth" and are as equally long-running as this, and for me seeing af finished workunit provides a lot of the feeling of having contributed something.

kibidk:
I hid your duplicate post.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Not 100% sure it is impossible to develop GPU capable tasks for CPDN, I thought it more that without a substantial amount of money to pay for people with the coding and modeling experience required it is just not going to happen. Now if anyone has ? in the same sort of numbers as BOINC credits to donate perhaps we could see that.

In the meantime while CPDN is run on such a shoestring budget given the importance of the subject matter I will keep on donating my spare cpu cycles.

Dave

Dave, in addition to money, my understanding is that the model code is owned by the UK Met. office and not open source. So there might be other barriers.

And the problem (simulating climate) is not as inherently parallel as people seem to think, so the speed-up might not be very great.

Finally, until recently no GPU could do 80-bit floating point, which, AIUI, the code relies on. Nearly all the GPUs that people own still can't.

IMHO a GPU conversion might start to be worthwhile in five or six years, if the money can not be better spent on more sys. admins and more storage and more science. As you say, in the meantime, keep on crunching!

Greg, Dave

It would take more than my expected lifetime for me to understand the hadley Fortran code.

How to parallelize the code for GPUS -- no idea.

We work with what we have.

Thanks for your useful comments

keep crunching

E

---