climateprediction.net home page
new to BOINC, and I already think it sucks.

new to BOINC, and I already think it sucks.

Message boards : climateprediction.net Science : new to BOINC, and I already think it sucks.
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user3708

Send message
Joined: 30 Aug 04
Posts: 29
Credit: 35,656
RAC: 0
Message 2262 - Posted: 31 Aug 2004, 3:42:58 UTC

OK, maybe I am being a little dramatic. Still, though...

I just downloaded BOINC this evening and installed it. The major changes I\'ve noted so far is that it\'s opened up whole new worlds of confusion and complication.

For starters, am I supposed to let the old CPDN client keep running while BOINC runs? The old CPDN client isn\'t doing anything anyway, it\'s just stuck with a message saying how there was an error connecting to the registration server, whatever that means. Whenever I click \"OK\" to try and close it it just comes back up again. In other words, it never goes away, unless I close the program.

Meanwhile, I guess BOINC is now doing the model simulation, considering my CPU usage is 100%, although it\'s harder to tell what it\'s doing with the interface that\'s different from the old CPDN client interface.

These things I can figure out and maybe get used to. However, my major complaint is the cumbersome cross-project (or whatever) user ID that I could never hope to remember. It took me a long time to figure out that this is what I was supposed to enter when I attached the project to BOINC. I couldn\'t find anything pertaining to it on the forums.

Another complaint is that for some reason it now says in my information page that I\'ve been a CPDN member since today, even though I\'ve been a member for several months and completed four long runs. All the other relevant information on that page is also zero, despite the fact that (according to my calculations) it should be far from it.

Anyway, I\'m not quite ready to completely give up on BOINC. I\'ll keep giving it a try, and maybe after I get these things figured out and get used to it I may grow to like it. It does have a major redeeming quality: the fact that it can run on Linux. I don\'t know at this point how to set that up, but maybe with some help I can add my Linux box to the project. If it\'s gonna take up space in my room I might as well put it to work ;-)
ID: 2262 · Report as offensive     Reply Quote
Profile astroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 2272 - Posted: 31 Aug 2004, 5:10:10 UTC

Hi, willaby,

Welcome to the wunnerful world of boinc. Admittedly, it can seem more than a bit dysfunctional after the comparatively benign realm of 'Classic' CPDN.

The Classic CPDN Model can be allowed to finish. (It is recommended -- from a scientific viewpoint.)

Your hung Classic Model could be a consequence of the weekend reorganization of the servers. There were several problems. (A retry may be in order -- though when I last checked on the status of a friend's machine in the Top 200, it was still bollixed. Perhaps tomorrow, after the British holiday weekend.)

The cross-project ID is, AFAICT, a means for stats-people to link a user's stats among the sundry BOINC projects. Not something we should concern ourselves with, I think.

This BOINC thing was developed in the worldview of SETI@home (by the same folks) which had an embarrassment of riches in terms of participants -- wherein said participants were regularly fed the same set of work units over and over... to keep 'em busy donchaknow. BOINC gives those users something else to do, if they so choose.

BOINC is sufficiently versatile to permit users to run more than one DC client (and bleed away some of SETI's embarrassment) and not suffer loss of productivity during SETI's notorious outages. However, if one's interest is, say, climate modeling, i.e., science as opposed to merely accumulating credit, BOINC has less appeal.

That said, BOINC permits folks to participate with machines booted in OS other than the offal of Redmond (across the Coast Range & Puget Sound from my home).

The really good part is that BOINC allows P4 CPUs and servers to utilize the full capability of those very capable machines. (It doesn't use the [epithet deleted] M$ Registry.) At least at this point in evolution, P4s are reaping the benefits of the CPDN move to BOINC. (In Alpha/early Beta, M$ version was more effient than Linux but that has been reversed, at least for now.)

There are some Linux versions not supported. (Use the search option if you have a fringe flavor to see if it works here.) Recent major releases are known to work. (I have three boxes running SuSE 9.0 [but not 9.1 {been there, failed that}].) Linux runs "nicely" with CPDN. However, it is only on C/L -- no GUI.

Forget the stats for now. Carl worked the weekend to change servers & databases, and to archive some Classic CPDN stats into the BOINC world, not without difficulties. It will surely be sorted out in the near future.

Meanwhile, if you want to do Linux, jump on in! There is plenty of help here on the Boards if you get into a twist.


Download and install is similar to M$ version -- but check to be sure that the BOINC exe has proper permissions before trying ./xxxxxxx.

HTH.


________________________________________________
Indeed I tremble for my country when I reflect that God is just.
-- Thomas Jefferson
ID: 2272 · Report as offensive     Reply Quote

Message boards : climateprediction.net Science : new to BOINC, and I already think it sucks.

©2024 climateprediction.net