LHC@home

This is a response to Discover magazine's October issue.
============================================================

The October edition did an Ok job covering many things. But, ... (you knew the "but" was coming, didn't you?)

I was a little surprised that the issue did not mention one of the most exciting frontiers for the average person. That is the ability of an ordinary person to be a part of the excitement of science through the use of distributed computing. The Berkeley Open Infrastructure for Network Computing (BOINC) allows the average "Joe" or "Jane" to connect to one of several projects in production work and thus become a part of the scientific discovery process.

The BOINC System allows the participant to join one or more projects and harness the "idle" power of their computer and to donate that power to what is probably the worlds largest supercomputer. I note that there are about 3,000 computers added to the processing pool *DAILY* (see http://www.boincstats.com/index.php). This number is larger than many supercomputers' total number of nodes. And we add numbers like these regularly, bringing the total nodes to 742,411 (though I grant that not all nodes stay connected forever). But, even if you only count a quarter we still beat the current leader, IBM's 65,536 node cluster machine (BlueGene/L), which is currently rated the #1 supercomputer (http://www.top500.org/lists/plists.php?Y=2005&M=06).

The current production project list and *MY* classification includes:

SETI@Home, exploratory (UC Berkeley originator/developer of BOINC)
ClimatePrediction.net, weather research
Predictor@Home, Biology research
Einstein@Home, Physics research
LHC@Home, Engineering models for the building of the LHC (covered on the last page of the issue).

One of the reasons I bring this up, other than it is my passion, (and in the spirit of full disclosure I have to admit to a small site documenting BOINC to which I am the primary, but not the sole, contributor http://boinc-doc.net/boinc-wiki/index.php?title=Main_Page) is that in the early stages of the LHC@Home project there was considerable surprise at the speed in which test series were completed. If I recall correctly, we were doing about the same amount of work that their supercomputer was with equivalent numerical accuracy. In prosaic terms, we would "run 'em dry without any sweat". To quote an administrator of the BOINC Servers "... we underestimated computing power of "mere" 5000 users."

I believe that there was a short article about BOINC some time ago, but, if distributed scientific computing isn't a frontier I don't know what one is. Especially as with the currently active projects, and others that are in alpha/beta testing, touches on several of the frontiers that you did mention. Oh, and with BOINC being open source, any scientist with a problem that needs supercomputer time can tap into this pool with relative ease. I say with relative ease as many of the projects only have one or two permanently assigned staff. Super-computing for the masses.

Paul D. Buck

---

Well, you can perhaps add SZTAKI ?
This is for now the sixth project and it's not alpha or beta, so ...

---

Do you want to get banned for 31 years, your account and credits deleted at a Boinc project ? Predictor@home is your best choice.

Too late now. Already sent.

And I had not heard that Sztaki was out of beta testing. I know a couple of the other editors are running that and would probably have made the change already. But, I will look at it.

Nope, can't see anything that indicates that they are out of Beta testing.

One of the "problems" with BOINC is that there is not a really well defined transition from one state to the next. In general, the project in alpha testing is very restricted. But, this can also be the case for small projects or projects with limited resources.

Beta tests tend to be more open to participants. But, then, again, they may be still "testing" but are confident enough with the application to regard the successful results to be scientifically valid. SETI@Home certainly was this way. Though I know many would contend that the current application and BOINC are still only Alpha test quality.

Oh, and in looking I saw a new project has been added called Rosetta@home ...

I tried to look around in the site for both of these projects and still do not have a clear idea which state they are in. John Keck added Rosetta@Home in the beta phase ... so I will have to go do some more studies...

---

Rosetta@Home -- sounds like it is very similar to Folding and Protein --

http://boinc.bakerlab.org/rosetta/

---

[b]Ok Paul.......when I came here about an hour ago I was going to ask you if you had sent that in yet.

I guess you could read my processor's mind



I should have known you would get that taken care of!



Good work as always..........

---

Volunteer Mad Scientist For Life

Actually, I got "the word" because I pay attention to the "Recent Changes" page in the Wiki, a practice I recommend to everyone. John Keck discovered it and made the initial page entry.

I will be adding a/the developer of Rosetta@Home to the editors list tomorrow so that he/they can start to add material as they wish. To be honest, they have a dynomite site ...

Not sure about the work yet, I have one work unit for my PowerMac and it is listed as being 200+ hours, still, and 0% progress. I been watching football and reading all day, so I am not sure if this is a "stuck" work unit or not. Monday when I will be working (fate willing) I will do a "force" and see what is going on if it is still stuck at 0% done.

---

<blockquote>Not sure about the work yet, I have one work unit for my PowerMac and it is listed as being 200+ hours, still, and 0% progress. I been watching football and reading all day, so I am not sure if this is a "stuck" work unit or not. Monday when I will be working (fate willing) I will do a "force" and see what is going on if it is still stuck at 0% done.</blockquote>
This is one bug I have already seen mentioned there. The initial estimates are really bad, and the % complete jumps around. The workunits are similar in processing time to E@H. It should suddenly get to 6.67% and a reasonable estimate sometime on clients earlier than 4.70. Later clients, with the durration correction factor, will have to complete a few workunits to see reasonable estimates.

---

BOINC WIKI

BOINCing since 2002/12/8

Well, the first rosetta@Home completed for me. Looks like a couple hours (9,000 seconds) for the one. The new one I DL is still a 10 days to complete. So, I will do a few to see if it gets more reasonable. I wonder if Willy is set up to get XML from them or not ...

---

And I had not heard that Sztaki was out of beta testing. I know a couple of the other editors are running that and would probably have made the change already. But, I will look at it.

Nope, can't see anything that indicates that they are out of Beta testing.
And where did you read (or see) that they are in Beta testing ?

Then I quote "you" here :
Hungarian project. Details hard to come by for an English speaker.
This is difficult for All people who don't speak Hungarian and not only for an English speaker.

Some update : they are translating the web site in english
I quote them here :
2005. august 30.
Good news for our international users that the translation of the webpages has begun. The pages will transform dinamically from now on.

---

Do you want to get banned for 31 years, your account and credits deleted at a Boinc project ? Predictor@home is your best choice.

<blockquote>[/i]Then I quote "you" here :
Hungarian project. Details hard to come by for an English speaker.
This is difficult for All people who don't speak Hungarian and not only for an English speaker.
[/i]
</blockquote>

I made a minor change to this entry. Tell me if it is more accurate now.
"Details hard to come by for the non-Hungarian reader."

Wiki Main Page

---

Well, I looked in the Wiki first ... :)

Seriously, it is more of a case where we can see a positive declaration that the project is in production. Since I cannot read the pages I cannot see if they are in one state or another. So, I asked ... now I have to wait for an answer.

---

<blockquote><blockquote>[/i]Then I quote "you" here :
Hungarian project. Details hard to come by for an English speaker.
This is difficult for All people who don't speak Hungarian and not only for an English speaker.
[/i]
</blockquote>

I made a minor change to this entry. Tell me if it is more accurate now.
"Details hard to come by for the non-Hungarian reader."

Wiki Main Page</blockquote>

better indeed :)

---

Do you want to get banned for 31 years, your account and credits deleted at a Boinc project ? Predictor@home is your best choice.