anyone running the old boinc v4?

[NeoGen]

We could use a hand or two at Tanpaku. It seems a japanese team is coming up strong behind us.
http://issofty17.is.noda.tus.ac.jp/top_teams.php?sort_by=total_credit

Our #1 spot will be theirs soon if we don't do something.

I'm running it but somehow, don't know why, boinc can't queue more than 1 or 2 workunits and then only contacts the server in 1 hour. I finish the 2 workunits in less than half hour, so I'm time-sharing it with HashClash. It was either that or manually clicking the "Update" button every half hour to get more workunits. :?

[vaughan]

I found the same thing. I was using a P4 2.53 that normally runs SOB 24/7. I put Boinc V4 on it and Tanpaku grabbed 2 or 3 tasks, crunched them in less than 15 minutes each then sat there idling for ages. If I started up SOB it took the CPU 100% even at idle setting. Tanpaku's workunit expiry time is on;y an hour or so after you receive the tasks so you cannot cache very many if you try.

[Brucifer]

yeah, I quit running it. I'm on linux, don't have an optimized client for it, and the time limit stuff, etc. I'm still very unhappy with the boinc world on linux anyway. :(

[moving fusion]

I don't know if anyone saw (a while ago now) but i did have a problem running BOINC behind a proxy sever requiring authentication. No matter what I did I couldn’t get it to work.

What is the project you are running?

Sam

[NeoGen]

Proxy servers and stuff like that is totally off my field of knowledge. I only know boinc has a few settings to configure about it in the options.

The project we're talking here is Tanpaku, a japanese bio-medical project, similar to predictor but with even shorter workunits. (Always around 10 minutes, maybe less on more powerful computers)
Here's the site:
http://issofty17.is.noda.tus.ac.jp/index_E.php

The thing is that it only works with Boinc v4. If you try to attach to it using boinc v5 it won't allow. That's why very few people are running it.


EDIT: When attaching to the project on boinc, you have to use the original japanese link:
http://issofty17.is.noda.tus.ac.jp/
The site I gave you before was of the English translated pages. :roll:

[Brucifer]

If you are going to run *any* boinc project you positively need to get an optimized boinc client so that you can obtain as much credit as possible. This is another project that is pretty stingy on granting credit unfotunately.

However, that doesn't answer the original issue which was that more help was needed. I have a system that is seg faulting on sieving, so I'll move that over onto Tanpaku.

edit: Actually you sounded so totally *pitiful* Neogen that I put four systems on it. Linux of course. I guess it's time to get serious and try and get an optimized client for this.

:)

Edit II: heh... linux boinc 4.43 is the correct version for this. So if anyone knows where to find the source for it I would appreciate the help.

[Brucifer]

NeoGen, you are talking about the team that is #1 now right? And I think that is the group that is doing the project, or rather a group of the students there. So they may have some assets that will drown us, but one never knows until they give it the old *college* try, huh? Har Har whatta pun, college for a university etc., I know, I know, rather lame..

[NeoGen]

I don't know if the "Protein structural analysis room Japan" team is the one behind the project, I have seen it on one or another projects too already. (Now that I check for it... they're on Predictor, Rosetta, WCG, SIMAP, XtremLab, and more...)

The boinc binaries are easy to get, just go over to one of the boinc mirrors, like this one for example:
http://einstein.aset.psu.edu/download/boinc/dl/

And did you know that Crunch3r still has a couple of old optimized boinc v4 clients for linux? :)
http://www.guntec.de/Crunch3r/boincx86.html

[Brucifer]

That is for stock oem boinc clients, no source code. Need the source to re-compile with the cpu switches set.

As for the crunch3r client, I'll give the 4.7x one a try and see how that goes.

Thanks for the info. :)

[Brucifer]

I'm not going to keep crunching with tanpaku. Credit is just too low and I'd have to throw too many assets on it to get anywhere which would pull me off other projects. Sorry. :(

[NeoGen]

That's allright Brucifer. We'll catch up with the Japanese boys some other time. I too don't understand why the project grants so little credits, but it is remarkable that the same thing happens with the other Japanese project Cell Computing...

Two projects... both japanese... both use boinc v4... both award very little credits for workunits... is there a pattern here? :mad: