Is our participation useful?

tullio
tullio
Joined: 22 Jan 05
Posts: 2,043
Credit: 39,991,044
RAC: 9,114

Roadrunner has reached 1.026

Roadrunner has reached 1.026 petaflops sustained on the Linpack benchmark. Folding@home is claiming 2.426 petaflops. I suspect we are comparing apples with oranges.
Tullio

Stranger7777
Stranger7777
Joined: 17 Mar 05
Posts: 436
Credit: 366,242,025
RAC: 102,955

RE: Roadrunner has reached

Message 82471 in response to message 82470

Quote:
Roadrunner has reached 1.026 petaflops sustained on the Linpack benchmark. Folding@home is claiming 2.426 petaflops. I suspect we are comparing apples with oranges.
Tullio

I think it's OK to compare us with supercomputers in a raw perfomance.
But as I see in a statistics page we have only ~76000 active users right now.
And I now suspect they are bringing about 160-30=130 TFlops to the project.

So, now I understand that we are small only apart. But together we are the power. Ok, back to work. And let us see, who will find GWs.

JoeB
JoeB
Joined: 24 Feb 05
Posts: 121
Credit: 53,934,792
RAC: 33,175

RE: Roadrunner has reached

Message 82472 in response to message 82470

Quote:
Roadrunner has reached 1.026 petaflops sustained on the Linpack benchmark. Folding@home is claiming 2.426 petaflops. I suspect we are comparing apples with oranges.
Tullio

I just looked up the June list of the Top 500 Supercomputers at http://www.top500.org/list/2008/06/100 . The results seem to be system independent and only measure TFLOPs.

If I read the list right the E@H user applied computing power would place the E@H project in the top 10 based on the E@H server status page. This is remarkable performance by everyone associated with the effort! If other programs are doing as well or better it only proves the power of the distributing computing concept and the people supporting it.

Joe B

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,125
Credit: 127,324,678
RAC: 11,673

RE: I think it's OK to

Message 82473 in response to message 82471

Quote:

I think it's OK to compare us with supercomputers in a raw perfomance.
But as I see in a statistics page we have only ~76000 active users right now.
And I now suspect they are bringing about 160-30=130 TFlops to the project.

So, now I understand that we are small only apart. But together we are the power. Ok, back to work. And let us see, who will find GWs.


You might want to look at this brief glance at the E@H 'demographics' .... power(s) indeed! :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter. Blaise Pascal

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 51,668,899
RAC: 4,113

RE: I've just pulled the

Quote:
I've just pulled the statistics for top 500 participants into Excel and found that 2 participants on the top do more work then 500 others and so on.
I think now that our participation is unuseful, because ATLAS when working on this project when it is idle of other tasks do much more then thousands of us.
Moreover it has more power efficiency per watt that most of us have.
So, I beg Bernd and others to open our eyes on this dark side of computing. Is our power still useful for the project or we can step by step move to the others or even turn off our rigs just for the economy?


Numbers, numbers ...and numbers!
For example: with ATLAS we crunch S5R3 ~ 330 days. Without ATLAS ~ 400 days. ATLAS only ~ 1000 days. Last number may be equivalent of "without any analysis".

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,125
Credit: 127,324,678
RAC: 11,673

After a bit of a search I

After a bit of a search I found this post by Bruce Allen ( circa '05 ) on the cost savings due to us volunteers on power costs.
I guess there's a lot of ways that 'value' can be assessed for one's involvement in the project. We seem to look at this issue recurrently, which is a fine thing. Some questions which probe this:

If, say, the current contribution of was replaced by the equivalent processing contributed by an otherwise unrelated bunch of volunteers - would that change one's view?

If, say, an even greater fraction was done in house by the project ( 99% + ), and only a select few outsiders were invited to assist - would that change one's view if one were selected? If one weren't?

If, say, no large clusters at all were involved, but as a result the project had less total power - how would one feel about that?

I'm not searching for 'right' or 'wrong' here, just trying to define the intent.

I think it seems fair to say that the success of E@H depends upon volunteer efforts - we are the base horsepower.

There are probably many non-targeted and non-technical benefits, mainly of a social nature. For instance :

- the open nature of involvement acts as a social force in that for whichever countries materially contribute formal funds to LIGO and allied GW efforts, then those burdened taxpayers can have a hands on glimpse of the outcome of such spending through involvement. [ Contrast that with the Manhattan Project ... ] For me DownUnda it would be close to a mere few cents per taxpayer directly or otherwise ( AIGO say ).

- stimulation of youth to pursue science. Heaven knows they need a shove in that direction these days. I understand Livingston LIGO has a nifty visitors centre frequented by all ages and Hanford does tours. E@H is an important public web presence. This sounds like a waffly benefit but should not be under-rated. Many a successful scientist will 'date' their personal epiphany in following a science career from very specific exposure in their youth.

- it crosses national, ethnic, religious, gender, and political boundaries, AND is a neutral preserve for those measures [ see moderation policy ]. In a world beset with a plethora of divides this is an incredible blessing. A similiar situation occurred during the Cold War with mathematics.

Cheers, Mike.

( edit ) and of course, it exposes a multitude of Windows users to the benefits Unix type operating systems. :-)

I have made this letter longer than usual because I lack the time to make it shorter. Blaise Pascal

Stranger7777
Stranger7777
Joined: 17 Mar 05
Posts: 436
Credit: 366,242,025
RAC: 102,955

I've answered in your thread

I've answered in your thread and asked there a question. I'll ask here again.
Can you measure the work these first players do against others? We already pay our taxes to the science sphere. And we all (all the people in our country) pay these taxes.
Right here (in this project) only small (better to say - microscopic) part of us pay a valuable price. The others, I suppose, only "participate" (message boards, news reading, screen saver, etc).
So, it is better to buy 1,2 or 3 supercomputes with huge perfomance but from taxes payed by all country, than to pay for the excess elictricity individually.

tullio
tullio
Joined: 22 Jan 05
Posts: 2,043
Credit: 39,991,044
RAC: 9,114

My PC consumes 200 watts.

My PC consumes 200 watts. When I am using it with its LCD monitor I am not watching TV with its big cathodic tube monitor, which eats more electricity, and I keep it off, not in standby. So crunching Einstein,SETI, QMC, climateprediction.net, CPDN Beta and LHC (when this gives me something to do) I am sparing energy. And I can also listen to music via Internet, so my stereo radio is off too. PCs are wonderful machines.
Tullio

RandyC
RandyC
Joined: 18 Jan 05
Posts: 2,911
Credit: 111,139,797
RAC: 131

RE: I've answered in your

Message 82478 in response to message 82476

Quote:
I've answered in your thread and asked there a question. I'll ask here again.
Can you measure the work these first players do against others? We already pay our taxes to the science sphere. And we all (all the people in our country) pay these taxes.

Yes, the work can be measured, but I'll ask a question back: Why does it matter (to you or anyone else)? The work needs to be done and it CANNOT be done by one individual or even one or more super-clusters in an efficient or cost-effective way. Each volunteer contributes according to their resources and personal desires...if not, then they would not be attached here would they?

Quote:

Right here (in this project) only small (better to say - microscopic) part of us pay a valuable price. The others, I suppose, only "participate" (message boards, news reading, screen saver, etc).

I disagree with you on this. The majority of participants here contribute (to use your words...) a microscopic amount to the total. But that small effort outweighs the total amount contributed by the super-clusters and individuals with large computer farms. Those of us who post on these boards contribute an even smaller (though very visible) part.

Quote:

So, it is better to buy 1,2 or 3 supercomputes with huge perfomance but from taxes payed by all country, than to pay for the excess elictricity individually.

No.

An analogy: consider an ant or termite colony. It consists of a queen (in some cases multiple queens); soldiers, and workers (I am simplifying here). Could the colony survive or meet its goals if it consisted of only the queen(s)? Soldiers? Workers? No! It needs at a minimum, the queen(s) and the workers.

A BOINC project needs at a minimum, the project developers (i.e. queens) and the volunteers (workers/crunchers). Adding soldiers (large farms and super-clusters) helps, but is not necessary for the success of the project.

HTH

Seti Classic Final Total: 11446 WU.

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 51,668,899
RAC: 4,113

RE: I've answered in your

Message 82479 in response to message 82476

Quote:
I've answered in your thread and asked there a question. I'll ask here again.
Can you measure the work these first players do against others? We already pay our taxes to the science sphere. And we all (all the people in our country) pay these taxes.
Right here (in this project) only small (better to say - microscopic) part of us pay a valuable price. The others, I suppose, only "participate" (message boards, news reading, screen saver, etc).
So, it is better to buy 1,2 or 3 supercomputes with huge perfomance but from taxes payed by all country, than to pay for the excess elictricity individually.


Hi!

I think that is not right.

First reason:
Huge clusters like Nemo, Merlin and ATLAS will not be upgrade with the same rate, as our computers. Now, Atlas (but only Atlas(!) from {Nemo, Merlin, Atlas} group) have a good performance/watts ratio. But I think that my overclocked Intel Core 2 -> 3.15 more energy efficient than Atlas nodes or as minimum, close to it.

Second reason:
We already have our computers. Of course, after shutdown of BOINC many modern computers decrease energy consumption. But 200 watts with 100% of efficiency better that 50 or 100 watts with 1% of efficiency.

Distributed computations is the best system for computations. (If algorithm allow to use it).

Comment viewing options

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