Just an Idea.

hih_tv-Greg
hih_tv-Greg
Joined: 11 Feb 05
Posts: 94
Credit: 31,815
RAC: 0
Topic 189597

Hi Everyone,
I took some time to go through the message boards of the different BOINC projects. With a few exceptions most of the project seems to be running pretty well which started me thinking.

Distributed computing is still relatively new, the work that we all are apart of could at best be called "long shot" projects. The science is sound…but the probabilities are low.

Perhaps it is for this reason that the researcher needed to turn to BOINC as a means of refining their search parameters. As stated in one of the postings " the cost return" of this method of computation can't be beat.

As I have stated on E@H, we the common folk get to take part in projects that could rival the "Manhattan Project" of the Second World War. I don't know what the numbers are but
I would venture to say that we could give the "BEAST" at the NRO a real run for its money when it comes to brute force computing, never mind all of the "Super and Hyper Computing Centers" around the world.

What this system lacks is redundancy. We proved that distributed computing is do-able, what the developers need to do is think about distributed severs. Secondary severs disperse around the globe in the event the "main" server goes down or is in need of repair. Depending on where you are when you read this, you are familiar with your local weather and geological conditions.

At the moment we in California are in the tail end of a heat wave (one of many I expect for this summer). Yesterday Southern Cal was place on a stage 2 power alert, at stage 3 we start rolling black outs two of the project are located in California as far as I know. We are also prone earthquakes.

SETI at the University of California at Berkeley sits right on top the Hayward Fault As such an event could knock the system offline. The last big event in the "Bay Area" back in the 90's took the area down for a week.

How difficult would it be to include a "mini service" in the BOINC download that allows the user to not only participate in science, but allow that person should they choose too help insure the reliability of the system. The Master System could "poll" all the subsystems depositing bit of information in all of them. Should the Master System go down all the subsystems would "poll" each other and recreate the primary functions of the Master System.

This might be a stupid idea from a user that has way too much time on his hands…if it is forgive me. But with all the talent and skills I've seen here I know a solution can be found.

Greg

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

Just an Idea.

This concept has been discussed in the past. Sad to say, it is going to be years before it becomes practical. Heck, we are still trying to get multi-project DC going...

As to the other point. You are correct, when you look at the computing power we rival almost all of the supercomputers out there. I am not sure we are way over the top of them yet, though we do have a huge resivouir of power.

Heffed
Heffed
Joined: 18 Jan 05
Posts: 257
Credit: 12,368
RAC: 0

Partial functionality of that

Partial functionality of that is already built into BOINC.

One of the original intentions of BOINC was also to allow users to donate disk storage space. Projects could store data on your PC. Additional credit would be given for the amount of data stored.

Ocean Archer
Ocean Archer
Joined: 18 Jan 05
Posts: 92
Credit: 368,644
RAC: 0

Heffed -- Not a bad idea

Heffed --

Not a bad idea if you're connected 24/7/365, but if you're a dial-up, or you replace a hard drive, what then??


If I've lived this long - I gotta be that old!

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

RE: Heffed -- Not a bad

Message 14495 in response to message 14494

Quote:

Heffed --

Not a bad idea if you're connected 24/7/365, but if you're a dial-up, or you replace a hard drive, what then??


It would be stored on more than one machine. Say we issue 10 copies, you lose one ... we still have 9 copies. The files can be encrypted to prevent access and corruption can be done using MD5 and returning a list of signtures. You lose too many, your computer becomes de-certified to hold data...

Simple really ... just no one is using it directly. CPDN uses something like this INDIRECTLY by asking participants to hold on to the model data if they can ... but there is no reward if you do, no penalty if you don't ... I don't mind, so I have saved most of the sucess ones that I have done.

John McLeod VII
John McLeod VII
Moderator
Joined: 10 Nov 04
Posts: 547
Credit: 632,255
RAC: 0

The project would also phone

The project would also phone home occasionally to see if there was any need for some of the data on the HD. Massive uploads would be a problem for some users (slow connections and hundreds of MBs of data to transfer.

The Pirate
The Pirate
Joined: 11 Nov 04
Posts: 57
Credit: 23,332,769
RAC: 0

Bit-Torrent already does this

Bit-Torrent already does this with software. Software or fragments of software are stored on more than one computer. And all help with the download.


Tom Awtry
Tom Awtry
Joined: 18 Jul 05
Posts: 100
Credit: 520,861
RAC: 0

I have several old Windows

I have several old Windows “98” machines, which we use here in the office for training and emergency over scheduled clerical work. My question is; has anyone ever attempt to implement an E@H application using this old of OS?

Understandability, computation time will be comparable to performing E@H calculations with a slide rule, but I would like to having these machines doing something in their idle time.

Thanks,
Tom

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,870
Credit: 115,941,270,578
RAC: 35,454,715

RE: I have several old

Message 14499 in response to message 14498

Quote:

I have several old Windows “98” machines, which we use here in the office for training and emergency over scheduled clerical work. My question is; has anyone ever attempt to implement an E@H application using this old of OS?

Yes, you can get perfectly acceptable performance under Win98 or WinMe. I have four such machines sharing two projects, EAH and SAH, 70/30. Here are some approximate WU crunch times (the SAH client is optimised, which helps):-

Machine/OSVer . . Project . . . . Crunch times
=======================================
Cel466 /Win98 . . EAH/SAH . . . 34hrs/10.0hrs
P2-350 /Win2K . . EAH/SAH . . . 46hrs/14.5hrs
P3-450 /WinMe . . EAH/SAH . . . 35hrs/ 9.5hrs
P3-600 /WinMe . . EAH/SAH . . . 28hrs/ 8.0hrs

The celeron 466 is actually overclocked to 525 which is why it seems to be on a par with the P3 450. All machines run 24/7 and all except the P2 350 submit one result for each project in less than 2 days. The P2 takes around 2.5 days. No problems, even with the short EAH 7 day deadline.

Cheers,
Gary.

Tom Awtry
Tom Awtry
Joined: 18 Jul 05
Posts: 100
Credit: 520,861
RAC: 0

RE: I have several old

Message 14500 in response to message 14499

Quote:
I have several old Windows “98” machines, which we use here in the office for training and emergency over scheduled clerical work. My question is; has anyone ever attempt to implement an E@H application using this old of OS?

Quote:
Yes, you can get perfectly acceptable performance under Win98 or WinMe. I have four such machines sharing two projects, EAH and SAH, 70/30. Here are some approximate WU crunch times (the SAH client is optimised, which helps)

Gary – Thanks for informative and quick response concerning the 98’s, the only problem now is me getting up the energy this afternoon to put everything together. Sorry, for the poor English, in the last posting, no excuse, just thinking ahead of my fingers.

Take Care

berashith
berashith
Joined: 2 Jun 05
Posts: 2
Credit: 676
RAC: 0

PLEASE PARDON MY IGNORANCE

Message 14501 in response to message 14493

PLEASE PARDON MY IGNORANCE BUT I AM NEW TO "BOINC". what can we do with the credit we earn? it is enough for me to know that my donation of system resources helps to facilitate worthy research but if credits are issued, to what end?

thank you...
Berashith Niemand

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.