> My name is Pascal and this message has no meaning, but still has my approval...
>

Well, I can see why...:)

you can sign up for it? cool! where do i do that?

---

I do not have a superman complex; for I am God, not superman!

> > Predictor - don't forget predictor!
> >
> > Might not have been the "most reliable", but holds high honors for
> "Keeping
> > users informed if there was a problem!"
> >
>
> Then perhaps you should switch to Predictor...since they keep you well
> informed :)
>

Bosh.. Realize you're lost - step 1!

Predictor finished their stuff about a month back.. Now waiting for HW from Dell and SW updates to go to phase 2.....

Bosh... Spend your time reading history and not posting without info!

---

>
> Bosh.. Realize you're lost - step 1!
>
> Predictor finished their stuff about a month back.. Now waiting for HW from
> Dell and SW updates to go to phase 2.....
>
> Bosh... Spend your time reading history and not posting without info!
>
Perhaps you "noobie" do not understand sarcasm...:)

P.S. Read some more...:)

That's why I detached from Pirate too.
===========

Pirates@Home is the only Project I'm running right now, I find it to be less stressful on me & my Computers ... :)

---

One of the points missed with regard to cp.net is that the science is different enough that you cannot always use it as a yardstick to measure other projects.

The cp.net science does not need, or cannot afford, redundent processing of the models. So, they don't send out the WU/Model more than once. So, credit can be instantly posted and awarded.

This also limits the server loads because there is no need to keep work sitting around until the quorum of results is created. The other three projects; SETI@Home, Predictor@Home, LHC@Home all are using the redundent processing of work. This means that they are subject to the problems that arise from that need.

---

Damn Paul...you beat me to posting almost the same thing! But two points...

> One of the points missed with regard to cp.net is that the science is
> different enough that you cannot always use it as a yardstick to measure other
> projects.
>
> The cp.net science does not need, or cannot afford, redundent processing of
> the models. So, they don't send out the WU/Model more than once. So, credit
> can be instantly posted and awarded.

You might want to check the threads over at CPDN as it appears that some units are being sent out to more than one host (though I don't think that such units are being cross-validated as with other BOINC projects).

> This also limits the server loads because there is no need to keep work
> sitting around until the quorum of results is created. The other three
> projects; SETI@Home, Predictor@Home, LHC@Home all are using the redundent
> processing of work. This means that they are subject to the problems that
> arise from that need.

This is true in a double-sense. 1) as you have noted to me and others elsewhere, it makes the DB table much smaller; and 2) no cross-validation of work means fewer scans of that table.

> Damn Paul...you beat me to posting almost the same thing! But two points...

Sorry! :)

> You might want to check the threads over at CPDN as it appears that some units
> are being sent out to more than one host (though I don't think that such units
> are being cross-validated as with other BOINC projects).

Ah, yes ... I forgot about that aspect. Once a model has been run and the results are "interesting" they will re-run that model to see if it is a one-time anomaly or not ... but they do not do the check me twice (or thrice, or fourthsie, or ...) validation that the other projects use. As was stated in other threads ...

> This is true in a double-sense. 1) as you have noted to me and others
> elsewhere, it makes the DB table much smaller; and 2) no cross-validation of
> work means fewer scans of that table.

Yes ... :)

---

Where do you people find all these projects? The only four that I know about are listed on the BOINC home page. :(

Jim

> > switch?
> >
> > Are we all running all 4 projects at once? (only reason why I aint
> runnig
> > pirates is because they are only playing around with a screensaver and
> not
> > doing any actual real work, in their own words)
> >
>
> That's why I detached from Pirate too. I'm waiting for Einstein to start up
> next year, already signed up for it.
>
> L8R....
>
> ---
>
> src="http://boinc.mundayweb.com/pirates/stats.php?userID=78&trans=off">
>
>
>

---

---

> Where do you people find all these projects? The only four that I know about
> are listed on the BOINC home page. :(
>
Pirates@Home
There is a link there to Einstein@Home.

---

When you have a project where the WU's take nearly a week to crunch, it isnt tough to have enough for everyone in time.

---

> > Where do you people find all these projects? The only four that I know
> about
> > are listed on the BOINC home page. :(
> >

Contact,

We are still in early days, but, current projects are:

SETI@Home
Predictor@Home (Off-line upgrading to v 4.x and new servers)
LHC@Home (limited to 5,000 users, currently upgrading servers and may open up again)
cp.net
Pirates/EInstein@Home (currently pre-Alpha and have no operational WUs).

Your best bet for "optimum" results is to keep SETI@Home and join cp.net ... with their long work units you will never run out of work if SETI@Home goes off the air for awhile.

LHC@Home is stop and go as they get results back, then they have to do some back-end magic to send out new work.

In my personal opinion, your mileage may vary, you should not sign up for Pirates. When I was attached to them I had a lot of odd problems. And since they are not issuing real work, there is not a lot of sense for you to attach to them now.

And yes, I did detach all my computers from Pirates@Home ...

---

The formatting in this thread is odd if you happen to scroll down, perhaps another software-glitch...
Some will say -always complaining -
But, for heavens sake WHEN can they even get a small change right the first time?
Too much looks trial and error in my humble opinion.

No offence,
Greetz ;-))

---

> The formatting in this thread is odd
>
Yes. Just as odd: when viewed in 'Reply to Post' format is OK.

The last days everything is / or went odd.

Greetz ;-))

Put your money where it will do the most good!!!
Team Members (not very many)
Team members photos (Two years old)

---

Semper Eadem
So long Paul, it has been a hell of a ride.

Park your ego's, fire up the computers, Science YES, Credits No.

I would like to point out that if seti sent more work units, and allowed people more time before they expired, then it would allow for far less downtime. (which will make it not noticeable)

Also, the whole thrice validation, well while CP just seems to trust its users more, even with thrice validating each result you can be effecient about it.

Make a ram HDD, set the DB on the RAM HDD, set it to only scan it for comparison once a 100 or so results have been turned in (so it will do the comparison more efficeintly, thus reducing system strain even further), and have it save it from ram to HDD every 10 seconds or so.

RAM operates at speeds millions of times faster the HDDs. It makes SENSE to use a ram drive for intense server actions.

---

I do not have a superman complex; for I am God, not superman!

How much RAM would you need and what would be the bottle neck for this fix?????


Perspective
What the Berkeley crew has to do:
1. Seti@h Classic
1a. Maintain website
2. BOINC Core Client Development
2a. Maintain alpha, beta sites
2b. Answer other projects inquires about BOINC
2c. Maintain website
3. Seti@h 2 Application Development
3a. Maintain website
4. Astropulse Application Development
4a. Maintain alpha, beta sites
4c. Maintain website?


This is my guess as to the work required of the CREW, if is incorrect or incomplete please feel free to add or correct, I await your input….

Other projects:
1. Application development
2. Maintain website


Poll; who has the heavier workload…………..

---

Semper Eadem
So long Paul, it has been a hell of a ride.

Park your ego's, fire up the computers, Science YES, Credits No.

the "bottleneck" with that approach would be the amount of ram. You can only make a DB thats so big... That means that the database would have to be split into two databases that reference to each other (which I assumed has already been done, since its the only way to go). There is always a bottleneck, its just its size and location that changes, it will make for a less significant bottleneck, by removing it from the HDD, and placing it on the BUS, ram, or CPU. (all better choice then HDD since they have no moving parts)

I would like to point out that you put maintain sites about 5 times, yet mainting the sites is minimal work, its site CREATION that is hard work, all they have to do to maintain it is plug the occasional news report (which they dont even have to do that often). And, rarely, fix a problem.

Also, they dont need to "do"? seti classic, its already done.

So the only things of significance that needs doing from your list are:
2. BOINC Core Client Development
2b. Answer other projects inquires about
3. Seti@h 2 Application DevelopmentBOINC
4. Astropulse Application Development (which, if I am not mistaken, was halted)

Which are a chore by themselves, but there is no need to unnaturally inflate the list.

---

I do not have a superman complex; for I am God, not superman!