Noted and corrected guys...my bad...please accept my apologies.
L24

> I agree with you PoorBoy, I couldn't have said it better myself. I think
> Lobstah24 is overly excited that credit is now being awarded (sigh, not for me
> yet), but this is exactly my point: people who are frustrated don't need
> immature taunting from these message boards; it just may drive them away.
>
>

Credits have been "awarded", "granted", "Sloshed out" in the past, so I wouldn't get too excited.

As I stated in another thread, I will be very interested in how many credit's I gain after the 3 days.. Not for credit sake, but it will give me a gauge as to how much work (science) I've done in the last month or so, that was used..... So far today, I've gotten a couple chrunching days worth of "delayed credit". If after 3 days of validation, it apprears to me that there was a "substantial loss", I'll let people know the numbers...... I started today with 5197 credits, so we will see.....

With the new SNAP box, shouldn't we be able to see pending credits in the near future? It was stopped as the DB was bogged down, after a few days of "live". The SNAP box, as I've read from SetiBoinc, is 10 times faster than the old DB, and should be able to handle the load! Should be no problem, as Predictor has never disabled this functionality, and they have 1/5 the crunchers on old HW!

Woody,

I am so glad Predictor never removed the pending units capability, but they also wisely stopped accepting new members when they started having server and connectivity problems. SAH has chosen NOT to take this route, and so people are STILL migrating over, further cluttering up the project.

Becasue of this difference in registration, Predictor has been able to much better control the rate of their expansion, and that may be one of the reasons they still have pending up.


Feel free to take a look inside my brain

After the last crash, the database was moved back to the original raid array after it had been reformatted and the cluster size increased which gave us some modest gains in performance.

The SNAP device is still being used for the upload and download folders though, since access to those directories are primarily sequential read/writes in nature.

We’ll give the SNAP device another try after a few issues are resolved.

----- Rom
BOINC Development Team, U.C. Berkeley
[url=http://www.boinc.dk/index.php?page=user_statistics&project=sah&userid=85465]<img border="0" height="70">

> We�ll give the SNAP device another try after a few issues are resolved.
>
> ----- Rom
> BOINC Development Team, U.C. Berkeley


What are the issues? When might we expect another multi-day outage to move the database?

> Woody,
>
> I am so glad Predictor never removed the pending units capability, but they
> also wisely stopped accepting new members when they started having server and
> connectivity problems. SAH has chosen NOT to take this route, and so people
> are STILL migrating over, further cluttering up the project.
>
> Becasue of this difference in registration, Predictor has been able to much
> better control the rate of their expansion, and that may be one of the reasons
> they still have pending up.


Given that the "backbone" is the same software, Predictor had consistanly shown they know how to manage a project like this! The posts there with "don't do it like Seti", says alot!

> What are the issues? When might we expect another multi-day outage to move
> the database?

At this point I don't know, we are waiting on an OS update for the SNAP device last I heard.

----- Rom
BOINC Development Team, U.C. Berkeley

reading this thread, I'm wondering how many are serious about the science...

if the science of this project is important to you then I have a hard time finding a reasonable reason to leave the project. I'd think it would make you stay through the good AND the bad times, right to the end.

Tom Koenig,

This post is not to 'flame' you or anyone else. Just to clairify my position on a few things.

> So, MajorKong, you think I was calling you mindnumbingly stupid? I reserve
> that label for others in this thread, but if you took offense . . .
>

Well, you did list my 'flippant' response in your post...

> Look, I am aware of the hardware and software problems (just barely based on
> the news column, but more than the news in classic SETI), but are you so blind
> in your support of this project that you don't hear the frustration voiced by
> Battlemate? Do you have such near-sightedness that you can't understand his
> rant--offensive as it was--as a warning of a possible contraction or even
> failure of the project? If you and the rest of the 'Magnificent 7' are all
> that's left, will this project survive?

Sir, prior to the credits starting to come in today, I had just as much frustration as anyone, if not more. The last time (before today) that I had recieved any credit for BOINC/S@H was on July 25th. Rather than litter this board with angry rants about things and insults to the BOINC and S@H developers and staff, I just chose to channel my frustration into more productive areas. Two of them being BOINC/Predictor@Home and BOINC/CPDN Beta. My computers rarely were idle, though that did happen a time or two. Yes, it sucks when there is a loss of 'credits & results' due to hardware malfunction, but you know... It has happened in every D.C. Project I have ever been in, at one time or another. As has extended down time. It just isn't worth getting mad about it.

>
> Maybe it won't come to that, but his warning is valid.
>

I agree. His warning is valid, even if he was quite 'beyond the pale' in the way he expressed it.


> And I do take this project seriously. I wonder about some others who say,
> "WOOHOO!!! More work for the rest of us!" That flies in the face of what
> distributed computing is trying to do.
>

As do I. I first started participating in the S@H project in late May, 1999 (on my first account... Lost the e-mail address, and didn't have the password, so I started a new one on Jan. 3, 2000 -- and wised up about the p/w...). I became a S@H Beta tester when they were working to go from v2.x to v3.x, and I am a BOINC Beta tester as well. I enjoy trying new things, and wish to do my part to help improve the project.

Now, there are very valid reasons why myself and many others make occasional flippant comments about 'more work for us', and it ties directly into both the design philosophy of BOINC and the finite supply of work units.

Read this page:
http://setiweb.ssl.berkeley.edu/sah_about.php

And especially this paragraph:
------------
Finite work supply

Each work unit is now processed a limited number of times (typically three). When we have no work for your computer, you'll get a 'no work available' message. We encourage you to participate in other BOINC-based projects; then, when SETI@home has no work, your computer can stay busy doing other scientific research.
------------

There is only so much data to be had from Arecibo. While it is due to increase quite a bit shortly, in the past S@H balanced the participants desire for work with the amount of work available by sending out the same work many, many times (7x, 10x, even more in some cases). In other words, 'busywork'. This will no longer be done in BOINC/S@H. Each work unit will be processed only the minimum number of times necesssary for verification. They ARE going to run out of S@H work from time to time. So, those that wish to ONLY process S@H work need to get used to the idea of 'idle time'.

Have a good day, Mr. Koenig.
------------
KWSN-MajorKong
KWSN Forum Admin (retired)
http://www.kwsnforum.com

BOINC Beta tester
Member of the 'Magnificent 7'

> reading this thread, I'm wondering how many are serious about the science...
>
> if the science of this project is important to you then I have a hard time
> finding a reasonable reason to leave the project. I'd think it would make you
> stay through the good AND the bad times, right to the end.
>
>


I am 'in it' for about 90% science, 5% good times with friends with similar interests, and 5% 'the credits'. I am not going anywhere. :)

Ok,

After taking the time to read the original thread and then all of the replies to it I have completely stopped being angry about BOINC and just about the thread in general
Ok, my spelling won't be perfect nor my grammar but this is what I have to say.
The point of this project is to unite many for a common good. While there may be ups and downs in systems, servers and clients the project goal remains the same.
We stand united ladies and gentlemen in our goal to find other life out there. Now whether we use the BOINC client or the Classic, our goal is one in the same. So what, there are some issues with the new client, it happens. If you think writing code and maintaining servers can be done perfectly the first time, try doing it sometime. I tried to code once and realized I could never do it because of the errors.
My final point is this. If we are all united in our goal, then why fuss over credits and such. Sure friendly competition is good and all but, not when it takes away from our true goal. If we cannot be united in the fact that this isn't just a competition for credits but a united scientific endeavour, then we have no point being here in the first place. I may be no expert but, I say from the looks of this thread, many of us have forgotten our goal and instead have focused more on the temporary client/server/credit issues. If you want to go back to classic until BOINC has everything fixed fine, if you want to stick with BOINC also fine. Just remember this isn't just a game or competition. It is a search for something bigger than us. A search for someone else out there with an intelligence enough to say "hi" over the interstellar airwaves. Please don't forget that.
I just started using BOINC a few days ago and have attained no credit whatsoever. I don't really care though. It's all in the name of science anyway.

Chris

>
> I am 'in it' for about 90% science, 5% good times with friends with similar
> interests, and 5% 'the credits'. I am not going anywhere. :)

happy to hear that :-)

personally, I was in it 100% for the science when I joined SAH I back in '99 - later I discovered the fun in looking at rankings and all that :-) But still, at least 90% was the science - actually I'm tempted to say 100%, as I've always seen the other stuff as a kind of funny sidekick *hmm*

Still, my computer (maybe plural in future) will keep crunching seti-numbers for the science :-) I'm absolutly sure that there's life outthere, and I want to contribute to find the proof. I might not be the one who get "the wu", but that's not the important part - the cpu ime I contribute, however petit it might be, will help speed up the process. Maybe nothing will be found with the project, but still it won't have been a waste of time, as we've then learned that this just wasn't the right way to find the proof, but maybe the project would have helped learning other ways to find the proof.

applaus :-)

(btw my spelling and grammar isn't perfect either - but then again, english is not my native language ;-) )

I didn't expect you to flame me, you seem like someone who puts some thought into your posts. They may be a little wordy, but so are mine, and I'm about to get wordy on you now.

I'm afraid I haven't communicated very well the reason for my anger at the initial responses to Battlemate: many of the folks posting here just don't see that his outrage, serious enough to cause him to quit (or claim to), is not cause for glee, it's cause for concern. How does distributed computing work? It works by distributing the work over a large number of individuals' computers. Generally, the more participants there are, the more the chance of success in doing the science. And the opposite is true.

I suppose, to be more polically correct, I should have said that your response to Battlemate was mindnumbingly stupid, rather than you personally. I suspect that clarification won't make you feel much better. I cited those responses to point out what offended me, and if you (all of you) didn't like my words, then how do you justify your insensitive remarks to Battlemate in the first place? By the way, Battlemate did not direct his remarks to you, but to the project, you on the other hand directed your remarks to him.

I know you are a BOINC beta tester, and you should be proud of that, but don't let your pride blind you to the legitimate (though sometimes obscene) rantings of people like Battlemate. There's a message there for you, the developers and all of us.

OK. There's a finite work supply. But "classic" has WU's to send out, and if SETI II can't keep it's hoppers full, it wasn't ready for release--and that's part of Battlemate's frustration. If you've always had work you are lucky; but if you've always been busy and working at capacity, then how much more can you do now that Battlemate is gone and no longer doing any work?

I hate to nitpick, but it seems to me that there are very few problems with the actual software - all of the problems encountered over the last little while seem to have been server side.

> I hate to nitpick, but it seems to me that there are very few problems with
> the actual software - all of the problems encountered over the last little
> while seem to have been server side.
>
>
My premise stands. No matter the reason, the project is driving people away.

> > Woody,
> >
> > I am so glad Predictor never removed the pending units capability, but
> they
> > also wisely stopped accepting new members when they started having server
> and
> > connectivity problems. SAH has chosen NOT to take this route, and so
> people
> > are STILL migrating over, further cluttering up the project.
> >
> > Becasue of this difference in registration, Predictor has been able to
> much
> > better control the rate of their expansion, and that may be one of the
> reasons
> > they still have pending up.
>
>
> Given that the "backbone" is the same software, Predictor had consistanly
> shown they know how to manage a project like this! The posts there with
> "don't do it like Seti", says alot!
>
>


Give it a rest woodrow please.........

This 'SPACE' for Rent.

M7

> I didn't expect you to flame me, you seem like someone who puts some thought
> into your posts. They may be a little wordy, but so are mine, and I'm about
> to get wordy on you now.

Fine by me! :)

> I'm afraid I haven't communicated very well the reason for my anger at the
> initial responses to Battlemate: many of the folks posting here just don't see
> that his outrage, serious enough to cause him to quit (or claim to), is not
> cause for glee, it's cause for concern. How does distributed computing work?
> It works by distributing the work over a large number of individuals'
> computers. Generally, the more participants there are, the more the chance of
> success in doing the science. And the opposite is true.

Generally yes... Up to a point. But what about the case when there isn't enough data to supply all the participants? THIS is the situtation S@H is in, and one of the primary driving forces behind BOINC development. As Dr. Anderson stated in his presentation at the CPDN 'open day' at Oxford, S@H-Classic solved the 'lack of enough data' problem by processing each work unit many more times than is necessary, effectively making most of the 'work' on S@H-Classic useless busywork. Last time I estimated the redundancy at S@H-Classic, I came up with figures that said each work unit was processed between 7 and 10 times. This is WAY too many times. The BOINC/S@H developers (including Dr. Anderson) have decided that 3 times is enough to satisfy the scientific requirements. This means that S@H-Classic is doing, conservatively, 2-1/3 times the amount of work that is necessary. Dr. Anderson has stated that this is a 'waste of energy', and I happen to agree with him.

> I suppose, to be more polically correct, I should have said that your response
> to Battlemate was mindnumbingly stupid, rather than you personally. I suspect
> that clairfication won't make you feel much better. I cited those responses
> to point out what offended me, and if you (all of you) didn't like my words,
> then how do you justify your insensitive remarks to Battlemate in the first
> place? By the way, Battlemate did not direct his remarks to you, but to the
> project, you on the other hand directed your remarks to him.

I was not offended. I was just pointing it out. I have a THICK skin and am not easily angered. No problems.

> I know you are a BOINC beta tester, and you should be proud of that, but don't
> let your pride blind you to the legitimate (though sometimes obscene) rantings
> of people like Battlemate. There's a message there for you, the developers
> and all of us.

Yes, I tend to be 'full of myself' at times... :P But in this case, I think there is an issue that maybe you are not giving sufficient consideration to.

> OK. There's a finite work supply. But "classic" has WU's to send out, and if
> SETI II can't keep it's hoppers full, it wasn't ready for release--and that's
> part of Battlemate's frustration. If you've always had work you are lucky;
> but if you've always been busy and working at capacity, then how much more can
> you do now that Battlemate is gone and no longer doing any work?
>

OK. Now for the crux of the issue. Yes, S@H-Classic has WUs to send out. But BOINC/S@H (aka S@H II) will NOT likely keep its hoppers full. EVER. The demand will be TOO great. Consider this. There are about 500,000 active participants in S@H-Classic. I am not sure of the exact number of computers that these 500,000 people are running on the project. Lets assume, for sake of argument, that the average number of computers per participant is 2 (the exact number really doesn't matter in this computation, since it will be assumed the same under both S@H-Classic and BOINC/S@H). Ok, thats 1,000,000 computers running S@H-Classic. When S@H-Classic ends, let us further assume that all 500,000 active participants migrate over to BOINC/S@H with ALL their computers, giving an average number, again, of 2. Let us then consider that BOINC/S@H has the exact same data feed as S@H-Classic (as it will in the near term, at any rate).

Since BOINC/S@H will only produce 3 work units for every 7 that S@H-Classic produced to keep those 1,000,000 computers busy 24/7, each one of these computers on BOINC/S@H will have work only 42.9% of the time. Or, one can consider that only 42.9% of those computers will have work available. This translates to only enough work for 429,000 computers (out of the 1,000,000), or only 214,500 participants. Now then, just what are the other 285,500 participants gonna do?

This is where the multi-project capacity of BOINC is VITAL! This will allow everyone to get their share of S@H work, and still keep their computers busy on other projects while staying under a unified credits scheme. If people are going to refuse to run ANY other project than S@H, and they are going to get mad because their computers are idle much of the time, then perhaps now is the best time for them to find something else to do. Again, this point is vital. If participants refuse to run anything but S@H, and they are unwilling to allow their computers to be idle much of the time, they are GOING to be unhappy.

Yes, Berkeley made a mistake with S@H. Their mistake was in allowing the project to expand to the point where they needed such obscene amounts of busywork to keep the participants happy. Now, they are stuck trying to wean participants off of the near-constant S@H data stream they are used to. I don't envy Dr. Anderson and the rest of the staff their task in the least. Frankly, I would rather have to give a bath to a feral cat without benefit of body armor, than to be in their shoes.
------------
KWSN-MajorKong
KWSN Forum Admin (retired)
http://www.kwsnforum.com

BOINC Beta tester
Member of the 'Magnificent 7'

> But BOINC/S@H (aka S@H II) will NOT likely keep its hoppers full. EVER. The
> demand will be TOO great.

> Now then, just what are the other 285,500
> participants gonna do?

> This is where the multi-project capacity of BOINC is VITAL!

Wow, not many people understand that there won't be enough work untill seti starts up astropulse and the second telescope to search for wide signals and signals from the southern hemisphere.

I'm happy to see that at least some see this point. (Not flaming anybody here, haven't even read the entire thread...)

Things will get a bit better when climateprediction.net gets online in a few days. Then you can do resource sharing between seti and cpdn.

MajorKong,

As I understand it, you only get credit when three people return the same result, then all three get credit; if only two agree, then no one gets credit. This is supposed to make cheating harder. Why then shouldn't the BOINC platform send out four or more copies and credit those three (or more)who match? Eventually, a WU (for lack of a better term) that didn't match the first time has to be sent out again to be processed, but if the first three times it's processed do not match, the original recipients just wasted their time even if one or two of them did it right. The 7 - 10 times classic sent out WU's may make more sense.

Also, there are fewer than 30K at BOINC now, yet many have not been able to regularly download work (even though the same data is being processed under both versions). I have gone back to classic (to try for a personal goal) and I read that many others have done the same. When you say that there will not EVER be enough work, then you are suggesting that BOINC has all the participants it needs, and I may as well not be here. Is this the message that Berkeley wants to send? That there is no need for (more) participants? That Battlemate can go, cause they don't need him? Is that the message you want to send?

Anyway, how the work is doled out is still off point: even if Battlemate's leaving gives you more work, you said yourself that other projects under BOINC should have work available to process, but if too many people leave, none of the many projects will be done efficiently. And SETI II will suffer if people shift over to the other projects because it's so messed up.