Wish/Question. Do you have any smaller work units besides the 3000 hours needed. I have gotten almost half done (900 hours) one or two time then losing the unit. I know once its lost, can\'t be recover. Even a model with a 1000 hours is 41 days. I have three machines run and have tried maybe 20 models without getting one to the end. I would ready love to finish at least one, but 3000-4000 hours just seems too long.
thanks siifred

---

SIIFRED

No, these are the only ones at present. Which is why we keep telling people to backup their BOINC folders.
See the README fles here.

But data is returned at regular intervals during the crunching, so all is not lost if it doesn\'t make it.

However, a lot of yours have been aborted by the user, some have an exit error 1, and others the very common 107... error.
The error 1 can be exiting Windows without first exiting the program, or a graphics problem. eg Attempting to expand the globe image to full screen.
The 107 error is also a graphics problem.
Both of these are covered by the READMEs.

Given the big uploads at 1960, 2000 and 2040, you can look at the 160 year model as being four smaller models run in turn. If your model crashes just after 2000, for example, a new one can theoretically be generated from 2000-2080. This isn\'t possible at the moment, because the software hasn\'t been written, but the potential is there.

---

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

Given the big uploads at 1960, 2000 and 2040, you can look at the 160 year model as being four smaller models run in turn. If your model crashes just after 2000, for example, a new one can theoretically be generated from 2000-2080. This isn\'t possible at the moment, because the software hasn\'t been written, but the potential is there.

OK, I was given an almost 7000 hour simulation, which isn\'t much less than the one year deadline, so I think I won\'t even try, and will exit your project. Maybe you should rethink and produce 20 year units, uploading the results and starting over with somebody else. That would only be a problem for people on dial-up access, which my guess not very many are.

Maybe you could give people a choice. Especially if your sofware takes advantage of multithreading, somebody with an 8 way processor running 2+ GHz, it would be OK, but for my single threaded 1.4 GHz, that is too much.

Since it will be ~7000 hours and you can\'t restart it after a large upload, I will quiesce this project.

---

I agree that smaller work units would be great. 2000 hours on my machine is just way too long. Reducing the work unit size to 1/10 of the current size would be great. I had started a work unit when my computer was running almost 24/7, but then my environment changed so my computer runs during the day and is turned off at night. Additionally during the summer it got too hot to run the client. I looked at the deadline and realized that the work unit would not complete in time. I was not interested in wasting thousands of hours on a work unit with no chance to finish. If the work unit size stayed the same but the length of time dropped to 1/10, it would still be 200 hours for me before finishing. That is a long time but far more reasonable than a man work year. This would increase bandwidth, but hopefully it would be doable. As of now I am not currently participating in the project because the work units take too long. Please consider this change request as I think the project is excellent and worthy of my computing time.

---

1) Shorter work units are being tested at the moment.

2) There is no deadline. The \"1 year\" is because the server software requires a finish time for the other projects, but it\'s ignored here.

3) The time-to-completion given by BOINC is an estimate, and is only accurate for the shorter projects. The actual time that will be taken is less.

4) And the climate of each model is uploaded to the servers at intervals as described in an earlier post in this thread. Hence it doesn\'t actually matter too much if you don\'t get to the end of the model.

---

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

Honestly, I\'m not looking for an argument here, I have dumped this project, and will not rejoin until the work units change to considerably shorter. I had one dumped on me that would have taken 3 years (not far off from my observations). The answer is quite simply no. Take it for what you will. I see you are testing small work units, that\'s great. If you\'d like a model for a work unit, SETI is good, ABC is short.

---

If you\'d like a model for a work unit, SETI is good, ABC is short.

One cannot compare the work done. Seti searches for that 1 elusive message by ET. ABC looks for triples in the code given to you.

CPDN looks for weather changes over hundred years. How exactly do you expect that to be compressed in a result that runs in an hour? Unless you aren\'t interested in the science done behind whatever you crunch for, can no one come up with a comparison this strange.

Why not add Primegrid into it, which searches for high primes, yet only runs 5 minutes? If it\'s run time results you want.

---

Jord.

Hi Prince

The best idea for you is to bookmark and occasionally look at one of the cpdn News and Announcements threads (eg through my signature). As soon as shorter workunits are available we\'ll announce them there.

---

Cpdn news

Hi Prince

The best idea for you is to bookmark and occasionally look at one of the cpdn News and Announcements threads (eg through my signature). As soon as shorter workunits are available we\'ll announce them there.

This was a very polite and helpful response. Thank you from me, though I did not post this wish. The people that provide the computers that do the crunching have to be pretty dedicated to even reach these discussion boards:
1) Find the project website (not the BBC one, the correct one)
2) Decide to use the left hand menu instead of any of the other text and links, because if there is somewhere to post it\'s probably there
3) Select the correct option from 14 choices in that menu - again guessing that this is the way to go
4) Aha! Select the link on the page \"take me to the boards\"
5) Select the right one of three boards
6) Select the right one of 6 sub-boards
7) Post

For most users this is far far more effort than they have ever put into finding a place to post a simple wish or request. Then when they do go to the effort of making a wish in a place that\'s actually called \"wish list\" half the responses tell them that their wish demonstrates their lack of credentials. The poster is sometimes also reminded that they committed the sin of not reading through thousands of previous posts in case someone else had the same wish and this has already been answered. If they did read the other posts they would find that most of the other \"wishers\" have given up on cpdn as well, and for the same two reasons:
1) Long work units
2) Difficulty in finding information / where to post / find answers / fix issues
and maybe also
3) Dismissal by some of the responders when they actually manage the above

So mo.v, thank you

Idea: How about splitting off 90% of what\'s on this site (or just leaving it where it is), and making a site with the 10% of information that the average user actually wants and can navigate?

Yes I know us average users are not really interested in the science (horrors!), we\'re lazy and impatient, we ask stupid questions, we don\'t listen, and we\'re often ungrateful. But we do have 1,000,000,000 computers between us and that\'s a lot of computing power. And if you can find someone to spend the time and effort (I know this isn\'t growing on trees, but stay with me) to make things clear and easy for us we might not ask quite so many stupid and repetitive questions. That\'s saved time for you.

I guess that over 10,000,000 users are just four clicks and shorter work units away. If we can do it in 4 clicks and it runs in a reasonable time we\'re in. If we can\'t then we\'re running another project or watching YouTube instead. Morons, I know, but there you are.

Perhaps start a thread (or shall I?) on what a redesigned website would need - you might enlist a few \"wishers\" to do half the design and creation for you through a discussion or wiki.

I know you mentioned that there\'s another person involved with the site and I will follow up on that, I\'m just getting it down here as it occurs to me.

Thanks for listening, and tell me what you think.

Quentin

---