Distributed Computing Platform (EU@Home)

[ab5tract]

Hey all, big admirer of Thunderbolts and of EU in general. I was wondering if there might be any usefulness in developing some EU@Home like distributed computation software. I don't know, it just seemed like it might strike a good balance of utility and evangelism. Plus I would love to have plasmas rendering while a screensaver plays already-rendered data.

Is there a use case for this? Would it be useful to develop a distributed supercomputer to throw experiments at?

[davesmith_au]

Giday ab5tract, and welcome. Are you just thinking out loud or do you yourself have skills in the area of software development which could come in handy? We do have one colleague who is currently developing a simple "EU quiz" which should be an interesting prospect when it comes to life. If you've got skills to offer, just let us know.

Would it be useful to develop a distributed supercomputer to throw experiments at?

What do you mean by "distributed" here? Also, how much money would be needed? It doesn't grow on trees, after all...

Perhaps a little expansion on what you have in mind is in order.

Cheers, Dave.

[DustyDevil]

This is an exceptionally good idea, especially now that both the software and the hardware are becoming more accessible to the non-funded scientist.

Berkeley’s Open Infrastructure for Network Computing (BOINC), is one of the leading distributed computing platforms in the world, and it is freely available at the following web site:

http://boinc.berkeley.edu/

BOINC is currently used in several distributed computing projects such as SETI@home; here's a link to the page with other BOINC projects listed:

http://boinc.berkeley.edu/projects.php

The BOINC source code can be downloaded and adapted to new projects, so a lot of the necessary software development has already been done.

On the hardware front, BOINC has been adapted to run fastest on graphics processors such as NVIDIA or Playstation 3. And the Playstation 3 game consoles have been pressed into service as super computers! Check these links out:

http://www.physorg.com/news148749271.html
http://www.ps3cluster.org/

I doubt there is a cheaper way to build a supercomputer, at least, at this time.

Here's another link for NVIDIA and BOINC:

http://www.nvidia.com/object/io_1229516081227.html


For the time being, that is until someone with a lot of time and money chooses to design an EU distributed computing project and/or supercomputer, the most effective use of BOINC from an EU perspective would be for this forum and its users to support a pre-existing project such as Einstein@Home:

http://www.einsteinathome.org/about/index.html

Einstein@Home is processing the data from LIGO:

http://www.ligo.caltech.edu/

Einstein@Home is searching for gravitational waves from rotating neutron stars, black holes, and supernovae; see the following link:

http://www.einsteinathome.org/gwaves/sources/index.html

IF LIGO fails to detect gravitational waves from these sources, then this would be an enormous, and impossible to ignore, failure of General Relativity. IF black holes exist, the ONLY direct evidence for their existence will be the gravitational waves they produce. For black holes to exist, gravitational waves MUST exist, and LIGO is sensitive enough to detect gravitational waves. So the sooner all that LIGO data gets analyzed, the sooner we'll know if black holes and neutron stars really do exist. And if they don't, then something other than gravity alone must be responsible for the structure of the universe.

LIGO seems to be made to order for undermining current cosmological theories.

Shouldn't we be helping?

[Steve Smith]

I used Seti@Home for several years analyzing the data from Arecibo.

What data stream do you have in mind for the potential EU@Home software to analyze? We aren't operating a satellite network or collecting observations from space that need to be upconverted from their bitbursts.

[ab5tract]

Sorry for the lateness of my reply. Thank you DustyDevil for clearing up my intentions and providing an excellent resource. I will be creating a team as soon as I get a chance to set up an account (It will be sad to replace my http://electricsheep.org but clearing up the whole black hole travesty comes first

@davesmith_au

I do have a reasonable amount of programming experience. I've never attempted anything as large as this would be, but there's always a first time for everything. What I do not have is any formal physics beyond high school. In other words, I can implement whatever math required into algorithms but I don't know enough about plasma physics to come up with said math.

And the best thing about distributed computing is its price! Assuming the software is developed without a fee (which it would be), the entire project is completely free (minus bandwidth costs associated with the main server that ties it all together).

@Steve Smith

Whatever we want! What kind of investigations are going on right now? What kind of computational experiments are we already conducting? I understand that plasma physics involves a lot of computer modeling--this could be farmed out to the distributed computing platform (which was my original thought).

Part of the goal of this project would be to expose the public to the science behind our "wild" claims, and that's starting with me

[ab5tract]

Alright I've started a team for Einstein@Home here.

[Steve Smith]

While I am a supporter of all input that is meant to increase exposure to, or analysis of Electric Universe concepts, I'm still at a loss as to what data you expect the system to use?

We are not generating large interactive databases; we aren't doing any data mining -- there isn't really a mountain of information that requires us to search within it; we aren't working with probes that require us to review volumes of telemetry; there's no backlog of information waiting for the appropriate software to crack it open; and the "plasma research" that is going on is mostly in distributed garages. No mathematicians are conducting prime number explorations or particle physics research or formulating equations that need to go beyond complex algebra.

You see my point?

If someone told me about the best hammer on Earth and that I just had to use it, if I didn't have a nail to drive it would do me no good.

What nail did you have in mind?

[bboyer]

What nail did you have in mind?

Help Team Thunderbolt Prove A Point

[ab5tract]

What nail did you have in mind?

Plasma simulations. I figured if researchers are using computers in their labs to render plasma simulations, then we could build a supercomputer that those renderings could farm out to.

Second to that, WHATEVER the community might NEED it for. If it is unnecessary, than by all means, we won't waste our time. However, as I tried to explain before, part of the reason for my post was to determine its utility and figure out what kind of experiments are being done in EU right now. As we all know, there is no theory without experimental evidence. Sorry if I sound ignorant or anything, I just have a large number of interests and not enough time.

I mean, let's model a nebula or something, I don't care. I want to do experiments, and if plasma experiments are done on computers at all, I would love to help out and this seemed like a good way.

[Steve Smith]

OK.

"...the reason for my post was to determine its utility and figure out what kind of experiments are being done in EU right now."

Distributed garages is the methodology currently in use.

And the reason for my post was to understand what it is you're trying to accomplish. As I mentioned, there are no "plasma simulation" models in the works. Do you realize the software work that would be necessary for supercomputer modeling?

"...let's model a nebula or something, I don't care. I want to do experiments, and if plasma experiments are done on computers at all, I would love to help out and this seemed like a good way."

How does one "model a nebula?"

The "plasma experiments" currently underway involve highly complex procedures. A tesla coil electrode is pointed at a chunk of material and the power is turned on for varying time periods. Pictures are taken and a video might be made.

Distributed computing is used for massive projects. As I wrote, there are no databases of information, there are no experimental research files waiting to be analyzed, there is no equipment that can be used to create the realities upon which simulations could be modeled. We have no monitoring devices, no high-speed cameras, no high-power transformers, capacitor banks, or laboratory facilities, in general. There are no reams of data being created.

I applaud your enthusiasm and hope your ideas will bear some fruit for you. What you're suggesting is probably something for five years from now.

[DustyDevil]

What nail did you have in mind?

Plasma simulations. I figured if researchers are using computers in their labs to render plasma simulations, then we could build a supercomputer that those renderings could farm out to.

Second to that, WHATEVER the community might NEED it for. If it is unnecessary, than by all means, we won't waste our time. However, as I tried to explain before, part of the reason for my post was to determine its utility and figure out what kind of experiments are being done in EU right now. As we all know, there is no theory without experimental evidence. Sorry if I sound ignorant or anything, I just have a large number of interests and not enough time.

I mean, let's model a nebula or something, I don't care. I want to do experiments, and if plasma experiments are done on computers at all, I would love to help out and this seemed like a good way.

ab5tract, I have a couple of suggestions you might consider investigating. As far as I know, they have not been set up for distributed computing; but if they were, they have the potential to move EU research forward. At least, I think they could.

In the 1980's, Dr. Anthony Peratt used a supercomputer to model the formation of spiral galaxies from interacting Birkeland filaments. You can read about his simulation at the following link:

http://www.plasma-universe.com/index.ph ... _formation

Peratt's simulation could be updated to increase its level of detail and complexity. Since the simulation is basically computational, it would probably not require a large database. Here's a link with current information about Anthony Peratt, including contact info:

http://www.24-7pressrelease.com/press-r ... -51954.php

Dr. Peratt may be able to get you started with this, or maybe another plasma simulation.

In November 2007, Dr. Gerrit Verschuur issued a press release claiming that he had found correlations between the signals from neutral hydrogen within our galaxy and the structure of the Cosmic Microwave Background (CMB) shown in the WMAP data. If Dr Verschuur is correct, that the neutral hydrogen signals are correlated with the CMB, then the CMB cannot be caused by the Big Bang. Dr. Verschuur's study has the potential to overthrow the Big Bang theory. You can read about Dr. Verschuur's work at the following link:

http://seti.sentry.net/archive/bioastro ... /0061.html

A statistical analysis of the neutral hydrogen data and the WMAP data could be conducted using distributed computing, and it could be a way to help Dr. Verschuur build the case for his analysis. Such a statistical analysis could be done following the example of the distributed computing project called MilkyWay@home which can be found at this link:

http://milkyway.cs.rpi.edu/milkyway/

Contacting Dr. Peratt, Dr. Verschuur, and/or the people on the MilkyWay@home project could help you get started on the experiments you'd like to conduct.

BTW, I have Einstein@home running on 2 computers at my own home, but I haven't yet joined your Thunderbolts group.

[ab5tract]

Whoa, thanks DustyDevil, it's going to take me some time to get up to speed with those resources, but thank you very much for providing them. Dr. Peratt's simulation was the thing that got me thinking about the distributed computing platform in the first place, but I didn't have a link for it so thanks for that.

A friend of mine has pointed out that Finite Element Analysis may provide a fruitful avenue for investigation, so you may hear soon of some developments we achieve in trying to re run Peratt's simulation or possibly others. He also mentioned that the calculations may not be very distributable, so it may be that a distributed platform is not very useful, as Steve Smith predicted.

BTW, I have Einstein@home running on 2 computers at my own home, but I haven't yet joined your Thunderbolts group.

You should join! Also, I would hope that the team could be seen as ours, and not "mine", as I never intended to aggrandize myself in the measure.

[Tzunamii]

What you're suggesting is probably something for five years from now.

What can be done now to achieve that point 5 years from now?
Are the foundations present now, data wise, to build upon to reach that point 5 years from now (or whatever time it would take)?
If not, where would one start?
If super computing is bringing a Howitzer to a knife fight, then cheaper alternatives must be available to at least begin something.
Its encouraging to see people wanting to take an active part
Thanks for the info and Links DD! Ill be busy reading this for a while.

[Steve Smith]

Send money.

[biknewb]

Brilliant plan. Just joined Team Thunderbolt.

Quote of the day: "It is only a matter of time until the gravity-only paradigm collapses under its own weight."
Feel free to copy.

Always willing to help that happen!