I have mentioned this before. I have rethought how to express it. While I do not like the method, personal anecdote may be the best method.

I started in May 2000. As an early user and one who did not turn off his computer even back then, I quickly rose to the 98th percentile even though starting a few months late.

For several reasons I dropped out of the project in the Fall of 2007. I came back in November 2009 and found I was in the 91st percentile. Duh?! At the moment I am in the 95th percentile.

Going from memory and which the percentiles seem to support the number of active users levelled around 2004. That is an obvious explanation for dropping from 98th to only 91st with a two year absence.

I have no memory much less records of the rate of increase of users in the early years. Were I to speculate the last significant increase was when S@H was worked into the script of the TV show JAG.

With this clear limit on the number of people who choose this project -- in fact any BOINC project -- given the number of users have been more or less level for five or so years who long will it take to double again?

That depends what you count. Counting by users it will remain level. Counting by CPU hosts it should double in three years or less.

This is simply because users will replace their old machines. Today one has to be on a very restricted budget to settle for a computer with a single core. If one spends the same amount as on the last computer it will have at least two cores. That will mean a doubling of processing equivalent to doubling the people participating.

Those who find themselves on restricted budgets should be balanced out by those who buy four core computers. So I will stick with doubling the number hosts REAL SOON!

I have no idea if 8, 16 and more core CPUs are in the pipeline.

So I ask if S@H ready to produce double the number of WUs? If not is software to do finer analysis of previous apps in development?

=====

If you step back this is in the category of awesome in the finest five year old tradition of misusing the word awesome.

The computing capacity available is going to increase for no other reason than multi-core will replace single core for no other reason than failure and/or replacement of existing computers.

If nothing better than four cores appears in the next ten years then the potential world distributed computing capacity will increase four-fold in the next ten year even if another 300 million Chinese do not start using computers.
[/b]

---

Unvarnished
Haaretz
Jerusalem Post
The origin of the Yahweh Cult

I have mentioned this before. I have rethought how to express it. While I do not like the method, personal anecdote may be the best method.

I started in May 2000. As an early user and one who did not turn off his computer even back then, I quickly rose to the 98th percentile even though starting a few months late.

For several reasons I dropped out of the project in the Fall of 2007. I came back in November 2009 and found I was in the 91st percentile. Duh?! At the moment I am in the 95th percentile.

Going from memory and which the percentiles seem to support the number of active users levelled around 2004. That is an obvious explanation for dropping from 98th to only 91st with a two year absence.

I have no memory much less records of the rate of increase of users in the early years. Were I to speculate the last significant increase was when S@H was worked into the script of the TV show JAG.

With this clear limit on the number of people who choose this project -- in fact any BOINC project -- given the number of users have been more or less level for five or so years who long will it take to double again?

That depends what you count. Counting by users it will remain level. Counting by CPU hosts it should double in three years or less.

This is simply because users will replace their old machines. Today one has to be on a very restricted budget to settle for a computer with a single core. If one spends the same amount as on the last computer it will have at least two cores. That will mean a doubling of processing equivalent to doubling the people participating.

Those who find themselves on restricted budgets should be balanced out by those who buy four core computers. So I will stick with doubling the number hosts REAL SOON!

I have no idea if 8, 16 and more core CPUs are in the pipeline.

So I ask if S@H ready to produce double the number of WUs? If not is software to do finer analysis of previous apps in development?

=====

If you step back this is in the category of awesome in the finest five year old tradition of misusing the word awesome.

The computing capacity available is going to increase for no other reason than multi-core will replace single core for no other reason than failure and/or replacement of existing computers.

If nothing better than four cores appears in the next ten years then the potential world distributed computing capacity will increase four-fold in the next ten year even if another 300 million Chinese do not start using computers.
[/b]

How can you talk increase without mentioning Cuda? It is way easier and cheaper to add a Cuda card and those will most likely be included in the machines you are talking about upgrading, I see it tripling within a couple years and NO Boinc is not ready for this kind of action.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

I'm not sure why you say BOINC isn't ready for more crunching power. Most of the projects I'm working in regularly run out of work, or have server or other project specific issues that slow distribution of work, but I don't see BOINC holding things up. It may be different for super-crunchers (anybody with a RAC bigger than mine) I guess.

I agree it will be interesting if the projects, SETI included, take advantage of this increase in crunching power, by adding new projects or making their projects more crunch-intensive (is that a real word?).

---

Maybe I am missing the point here, but if you can go from 1 to 4 cores so can S@H. I am also sure that these fine people who look after the hardware will be on top of this and already have their xmas shopping list ready. It just needs a few more people to donate $US10 each.

I'm not sure why you say BOINC isn't ready for more crunching power. Most of the projects I'm working in regularly run out of work, or have server or other project specific issues that slow distribution of work, but I don't see BOINC holding things up. It may be different for super-crunchers (anybody with a RAC bigger than mine) I guess.

I agree it will be interesting if the projects, SETI included, take advantage of this increase in crunching power, by adding new projects or making their projects more crunch-intensive (is that a real word?).

Because when my machine gets confused it isn't Seti but the Manager(Boinc) that is getting confused. Although you may consider me a Super Cruncher, I have always found that my "Fast Machine" is usually only Par in a year or two.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

Moore's Law works for & from every angle until some law of practicality hits (ex: desktop keyboards generally don't shrink for a reason). S@H benefits from piggybacking as other initiatives pay the dough to move the football forward.

even if another 300 million Chinese do not start using computers.

Far & Away the best thing that can happen to the science is China. They're our probable S@H project future --> Here.

Moore's Law works for & from every angle until some law of practicality hits (ex: desktop keyboards generally don't shrink for a reason). S@H benefits from piggybacking as other initiatives pay the dough to move the football forward.

even if another 300 million Chinese do not start using computers.

Far & Away the best thing that can happen to the science is China. They're our probable S@H project future --> Here.

I read and saw it at the Intel site some months ago, that Intel is working on an 80 core processor, that was supposed to be out sometime in 2011. I think they were having trouble getting the memory to sync up correctly, as some of the other I/O. Can you imagine 80 cores overclocked to the max, crunching away? Wow!

---

Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website

There is only one telescope recording data for SETI@Home.

During the past decade, the receiver has been upgraded to record more data (the ALFA multibeam receiver) once. It took a consortium of users to fund the new receiver.

There has been talk of a southern SERENDIP for years, but so far it is only talk.

The receiver and recorder has a "floor" -- a point where it can't distinguish very weak signals, and a resolution -- a point where subtle differences are simply indistinguishable.

The receiver has a fairly fixed bandwidth (frequency range) and increasing bandwidth means redesigning or replacing the receiver and/or recorder.

In other words, there is a finite, and fairly consistent flow of data.

It might be possible to squeeze a little bit more out of old recordings, but there is a finite number of old recordings.

It all boils down to a promise that SETI@Home has always made: there will be times with no work.

Unless they resort to "make-work" techniques (trying to search below the noise floor, for example) there will be times that there is no work.

I don't see an obligation for SETI@Home to produce more work unless they can justify it scientifically.

---

There is only one telescope recording data for SETI@Home.

During the past decade, the receiver has been upgraded to record more data (the ALFA multibeam receiver) once. It took a consortium of users to fund the new receiver.

There has been talk of a southern SERENDIP for years, but so far it is only talk.

The receiver and recorder has a "floor" -- a point where it can't distinguish very weak signals, and a resolution -- a point where subtle differences are simply indistinguishable.

The receiver has a fairly fixed bandwidth (frequency range) and increasing bandwidth means redesigning or replacing the receiver and/or recorder.

In other words, there is a finite, and fairly consistent flow of data.

It might be possible to squeeze a little bit more out of old recordings, but there is a finite number of old recordings.

It all boils down to a promise that SETI@Home has always made: there will be times with no work.

Unless they resort to "make-work" techniques (trying to search below the noise floor, for example) there will be times that there is no work.

I don't see an obligation for SETI@Home to produce more work unless they can justify it scientifically.

Myself, I just hope this is merely the beginning, and that as time goes on the scientists will be figuring out other means ET may use to communicate and that we will even look thru the same data again for something different. Time will tell.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

It all boils down to a promise that SETI@Home has always made: there will be times with no work.

Unless they resort to "make-work" techniques (trying to search below the noise floor, for example) there will be times that there is no work.

I don't see an obligation for SETI@Home to produce more work unless they can justify it scientifically.

Myself, I just hope this is merely the beginning, and that as time goes on the scientists will be figuring out other means ET may use to communicate and that we will even look thru the same data again for something different. Time will tell.

Me too.

Just making the point that the proverbial tail does not wag the dog.

The main goal is to try to find a persistent signal, not of natural origin, in among a whole great big ruddy stack of noise.

If they find another telescope (or several), if for example we start getting recordings from the Allen array (hopeful, but that's a political question) or from Parkes, or who knows what other sources might be possible, that'd be great.

We also need to be prepared (especially as we look to funding) for a time when most work requests get "no work available."

Edit: this is also why a few hours of outage does not really matter. As I understand it, the backlog of "good" recorded data is small and most of the tapes in the archive will produce a lot of -9's.

So, we can be out of work from an outage, or we can be out of work from being out of work.

---

If we do find THE signal will we be called upon to help with the translations? That should keep us busy for some time as no one could know how they communicate.

---

PROUD MEMBER OF Team Starfire World BOINC

If we do find THE signal will we be called upon to help with the translations? That should keep us busy for some time as no one could know how they communicate.

SETI@Home isn't looking for a signal that can be demodulated or decoded, just for some pattern that is sufficiently non-random to bear further investigation.

To put it another way, SETI@Home is more about filtering out the "uninteresting" data -- because the vast majority of the recorded data is stunningly boring.

... and there is so much of it that it'd be easy to miss the one "good" recording in all of that trash.

---

If we do find THE signal will we be called upon to help with the translations? That should keep us busy for some time as no one could know how they communicate.

I think that would be the time Uncle Sam steps in to take over.
Not much chance of an alien signal being left in the hands of those who've simply discovered it.

TOP SECRET-----NATIONAL SECURITY----NEED TO KNOW BASIS-----TOP SECRET

---

I do not fight fascists because I think I can win.
I fight them because they are fascists.
Chris Hedges

A riot is the language of the unheard. -Martin Luther King, Jr.

I agree with Ned, it may be that many projects just reply "no work available" in a short while, as the number and strength of hosts grows.

Taking a broad view, I hope that other scientific enquiries, SETI related or otherwise, can come up with clever ways to use all this growing computing power.

Taking a narrow view, where will all the credit hounds go when SETI and Milkyway can't provide full time work? I wonder if somebody is working on a new project to take advantage of the newer CPUs/GPUs?

---

In Dan Wertheimer's request for donations last December, one of the goals for 2010 is stated as:

Expand the frequency coverage of our search beyond the current 2.5MHz band.

The ALFA receiver has about 300 MHz bandwidth...

                                                                Joe

I agree with Ned, it may be that many projects just reply "no work available" in a short while, as the number and strength of hosts grows.

Taking a broad view, I hope that other scientific enquiries, SETI related or otherwise, can come up with clever ways to use all this growing computing power.

Taking a narrow view, where will all the credit hounds go when SETI and Milkyway can't provide full time work? I wonder if somebody is working on a new project to take advantage of the newer CPUs/GPUs?

I am Seti only, those are the only credits my hounds go after. It would all depend on what is available and if I am motivated. Although admitting to being a credit hound, Seti means a lot to me and that is the reason I crunch. If there are no projects looking for strangers out there, I would probably retire my machines.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

From the Useless Knowledge & Speculation Department, the Total vs. Active User and Host Graphs now go back two years.
I don't know what, if any, conclusions can be drawn from them, but feel free to gaze upon them.

it looks like overall more people start using BOINC to run seti then some migrate to other projects. there doesnt appear to be any major spikes up or down for users other than may 2009. IIRC we had an outage then

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

it looks like overall more people start using BOINC to run seti then some migrate to other projects. there doesnt appear to be any major spikes up or down for users other than may 2009. IIRC we had an outage then

Yeah... there was some sort weirdness that occurred back then... too many light beers in the distance for me to remember the details. But I think you're right, that big whooptie bump was some of the fallout of a larger than average outage.

I think, based on CPU cores, there will be an increase in both users (slow gain) and hosts. The real gain will come from the number of hosts with increasing numbers of cores.

But, as has been suggested earlier, the crunching game is moving away from CPUs to GPUs - CUDA on nVidia GPUs (rather slow) and anonymous platforms on ATI GPUs (the fast crunchers ATM).

Upgrading is turning towards adding graphics cards and not possibly CPU cores.

---

It's good to be back amongst friends and colleagues