stephenbrooks.orgForumMuon1GeneralBOINC version of Muon1
Username: Password:
Search site:
Subscribe to thread via RSS
Stephen Brooks
2008-04-01 11:00:48
Now that I've checked out that it's working properly, Yoyo from Rechenkraft.net's Yoyo@home project has managed to use his BOINC wrapper to allow BOINC users to do Muon1 simulations as work units.  This required some careful work to make sure the samplefiles and lattices got there correctly (I don't think BOINC natively supports intercommunication like this) and that the results history was preseved between work units.  But now it works, and since I've seen some of the "(Boinc Wrapper)" users in the top simulations by muon % score it's clear that they're making a valuable contribution to the project.  So many thanks to Yoyo for his work and hopefully this gives people another way to "enjoy" Muon1.
yoyo
2008-04-01 19:40:56
This answers also my questions whether the results are valuable or not.
yoyo
caferace
2008-04-02 07:19:06
It would be handy (as in nice, thank you please) if those of us who are also longtime DPAD contributors could "somehow" apply our credits earned through the wrapper to our respective teams.  I like running yoyo and am doing well for TA, but am focusing on OGR right now and continuing with DPAD on external machines with Stephens original coding.

File under: Nice to have. 

-jim
[TA]caferace
yoyo
2008-04-02 08:52:25
Hello,
there are sides like DC-Vault which compares distributed computing teams of boinc and none boinc projects.  If credits counted in yoyo@home and muon to your team your computation is cranted twice with credits.  Therefore I would not go this way.  The goal of yoyo@home is to bring the power of the boinc community to the none boinc projects.  So for the common boinc user the project should look like a boinc project, with the known page layout, credit system, boinc functionality.  Therefore the credits should be in the boinc part.  In the none boinc part the credits should go to a dummy team like now.
yoyo
yoyo
2008-04-02 08:53:51
If you want to have the credits in your muon team you can still use the original muon1.exe.
Stephen Brooks
2008-04-02 15:30:27
--[It would be handy (as in nice, thank you please) if those of us who are also longtime DPAD contributors could "somehow" apply our credits earned through the wrapper to our respective teams.]--

If you make your BOINC username "[TA]caferace" you'll appear as something like "[TA]caferace (Boinc Wrapper) 1234". Then when you list [TA] under Muon1's team stats, you'll see yourself twice, once as "[TA]caferace" and once as the longer name.

Can't offhand think of a way to make it any more integrated than that.  If you change your BOINC name like this I'll merge the ones without the [TA] to the ones with.
shauge
2008-04-22 17:36:20
What is happening here?
The prefix suggestion was working fine, but then it suddenly MY work was no longer credited to my user anymore (and team).  I lost one day of full production here.  I feel ripped off.  Now it seems as the users are slowly being merged into yoyo project user?
yoyo
2008-04-22 18:20:08
It was already discussed here http://www.stephenbrooks.org/forum/?thread=1134 (scroll down).

In short:
Teams in yoyo@home and muon are compared at pages like dc-vault.  If your credits go to your team in yoyo@home and in muon your team would get credits twice.  This is not fair.  Since yoyo@home is mainly intended to be for the Boinc community, for users who only runs Boinc, the credits must be in the Boinc team.
yoyo
[DPC]white_panther
2008-04-22 19:59:42
but why not do it to only the team the user is already in.
muon is already in the dc-vault stats, so i don't get it.

perhaps mine englisch is not that good that i understand it.
yoyo
2008-04-22 21:01:35
E.g. there is a team "TN" in yoyo@home and in muon.  A user which runs yoyo@home with muon should add credits to his team only in one of both projects, because teams are compared across all projects at e.g. dc-vaults.  The user crunches once and the team should get credits only once.  Since yoyo@home is a Boinc project for boinc useres the credits must be granted in the Boinc world.  In muon it should be invisible or granted to one anonymous account. 
User who have individual muon user- and team stats should continue with muon1.exe.
Stephen Brooks
2008-04-23 10:54:02
I also preferred it the way it was before, but I didn't want to start some sort of war with the BOINC/dc-vault stats people... :|
Stephen Brooks
2008-04-23 11:07:56
The main problem is that if yoyo@home was a true "wrapper" then it would not give any points itself, all the points would be scored in the respective projects.  Though then it would not give any "BOINC points" so existing BOINC users might not want to use it.
[DPC]white_panther
2008-04-23 23:27:07
stephen, maybe its an idea to produce a boinc client?
and that way get more users to run dpad
yoyo
2008-04-24 09:01:53
There are 250 user which did run muon in the last 7 days via Boinc:
http://www.rechenkraft.net/yoyo/server_status.php
Most of them were not aware of dpad before and are new to this project.  As usual for boincies they expect and got there credits in the boinc world.
BOHICASETI
2008-04-24 18:43:29
I was sorta curious about the number of results vs. the number of points awarded.  When I compare boinc (Boinc Wrapper) & [DPC] White Panther (they are right next to each other), I notice that boinc has about 1/10th the amount of results turned in but has an almost equal amount of Mpts awarded.  Is there something in the stats that is glitchy or is there an algorithm that I'm not seeing in the scoring of these results?  Not complaining or anything, just curious if I was the only one who noticed or if I'm brain farting something obvious.  Thanks =)
yoyo
2008-04-24 20:57:01
hmm, interesting observation.  I would say White Panther had more short work units and boinc had less work units, but most of them runs longer.  But these are only my very simple view.
Xanathorn
2008-04-24 22:55:55
I would say when the boincers start right away with the current samplefiles they would have mostly the 1000+ mpts results which take a very long time to complete, whereas e.g. WhitePanther and I started from the beginning when the lattice was first put online and the calculations are very short (especially negative scores).
Stephen Brooks
2008-04-25 12:24:11
Yes, the boinc project started when the linac-type lattices (already probably the longest simulations we've done) were already at quite high yields, so they've had a very large number of high-Mpts simulations.  People who have been around when shorter or low-scoring lattices are there have got a lot more simulations.
Hydrox
2008-04-25 20:47:44
Why is the boincteam no longer listed in the stats?
Stephen Brooks
2008-04-26 16:18:43
...because it's not a team any more, it's a user!  Search for "boinc (Boinc Wrapper)" in the user search box.  (Though you don't need to, it's in the top 50 by Mpts overall already).
: contact : - - -
E-mail: sbstrudel characterstephenbrooks.orgTwitter: stephenjbrooksMastodon: strudel charactersjbstrudel charactermstdn.io RSS feed

Site has had 25161229 accesses.