What you propose (1000 tasks per host) is example of wrongly applied rescheduling and I would warn against such actions until your host can complete all those 1000 tasks in few days (look current average turnaround for particular estimate how much days it will be).

And more common note - "if all" never will be implemented as I stated before. If all would care about theyr hosts right behavior there will be no limits at all. The same is true regarding rescheduling. Never "if all" will be reached for absolutely different reasons starting from "cause it's not fair" feeling of some to "I just doesn't care" feelings of others.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

I was sleeping and there are so many post to read/comment.
I dont´like to use names in my post to avoyd conflicts and missunderstandigs, but this time i´m forced to use.

@Raistmer you are part of the "good reshedulers" group, those i mention in my initial post as 2 group, the small circle who realy knows how to rescheduler without impacting the project, so i see no problem with your practice, on the contrary, what you make is try to optimize you resources and do more science, so it´s good for the project. And if IIRC you stay withint the limits, so emotions aside you are right from my POV. But the problem is the circle of those who realy knows is very small, very few realy knows how to do that.

In my opinion, call the use of optimized apps as "cheating" is wrong, why, the project allow their use, they use are within the roules, so you can´t "cheat" if you are within aproved path. Insist on this point is futile.

@Tbar There is nothing wrong if you have 400 ready crunched WU in your cache, if you have thats simple shows you do a lot of science and probabily you have an internet access problem. What is wrong is if you "cheat" to optain a 400WU cache of "pendings" because this are above the project 200WU limit Period. The same aplies to add mores hosts to your farm, add more hosts is aproved by the project, so nobody could said that is "cheating" too.

@Wiggo we agree in a lot of points, the only point i can´t agree with you is when you insist in point someone directly, that only could flame the thread, and we need to keep the thread alive. Help us, let´s focus on the real big problem, those who "cheat" and build larges caches just for credit (my called 3 group).

To all, i belive the problem now is allready exposed, what we need now is a way to fix it, but that is with the project dev´s and admin, we could make little from our side beside ask all the comunity to avoid the bad rescheduler practice and keep the good rescheduling between the aproved limit.

Just for curiosity... after 5 days without new AP WU avaiable form the project and our caches where empty of AP WU for days... lets see what the big 3 anonymous "cheaters" have in their caches...

http://setiathome.berkeley.edu/results.php?hostid=6637324 pendings 214

http://setiathome.berkeley.edu/results.php?hostid=6016862 pendings 727

http://setiathome.berkeley.edu/results.php?hostid=6794209 pendings 210

Now compare that number with the 82,613 Results out in the field (by last Server Status Page). That give us 1.4% of the actual total AP WU avaiable, in just 3 hosts who produces less than 0,1% of the project daily output, so anyone could see that´s wrong. And the number is a lot worst, why because the results on the filed figure count CPU+GPU WU so if you made the same math just on the GPU WU i´m sure you will get something very impresive (if anyone knows who to do that i´m curious to see the real number)

Anything else must be added?

And please take no ofenses from my posts, sometimes the translations are not good but please be sure that´s clearely not my intention.

<edit> Still don´t understand why anonymous hosts are allowed on SETI or any other project. Hide the list of your host theoricaly is to keep your privacy and that is questionable, anonymous hosts are not. I know it´s hard to separate both. The use of anonymous hosts leave allways to situations like this one, use of the anonymous courtain to break the roules. But that is for another thread.

---

Apologies, I was being sarcastic, which doesn't always translate well.
I have NO intention of rescheduling anything.
I was just trying to point out what would happen if many people did it and my take on it from a competitive point of view.


Reg

@Raistmer you are part of the "good reshedulers" group, those i mention in my initial post as 2 group, the small circle who realy knows how to rescheduler without impacting the project, so i see no problem with your practice, on the contrary, what you make is try to optimize you resources and do more science, so it´s good for the project. And if IIRC you stay withint the limits, so emotions aside you are right from my POV. But the problem is the circle of those who realy knows is very small, very few realy knows how to do that.

Almost but not quite. I'm above limits (well, already under but it's just becasue usual AP tasks shortage). We on common ground treating hosts whit too many tasks to process in time as misbehaving ones but differ in approach to limit per se. I don't consider it as "God's revelation" and "unbreakable".
Hence there are some differencies in my and yours definitions of "good rescheduling group". My little (little? perhaps yes indeed cause number who does rescheduling correctly small) wider.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

http://setiathome.berkeley.edu/results.php?hostid=6016862 pendings 727

http://setiathome.berkeley.edu/results.php?hostid=6794209 pendings 210

Only last one has 5 errors but 5 still not too big number, hardly it got those erros because of rescheduling.

EDIT:
http://setiathome.berkeley.edu/results.php?hostid=6637324 pendings 214

this one shows 22 errors. Something to check on this host indeed.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

You are right, but we are not talking here about error, i never say they are wrong because the error they produce, heavely error producing hosts if for a compleately diferent thread. I inclusive clearely said on one post, they inclusive return their crunched WU in time... but that not makes them right.

I´m sure if their pendings WU where distributed alond the community all this WU (besides timeout or error of course) will allready crunched and reported days ago, and those WU where allready cleared from the DB.

Lets focus on the 727 pending host. IIRC it has about 1100 WU pendings 5 days ago, so it crunches about 72 WU/day... Their cache still holds for 10 days more... so if my calculations are correct it work with 15 days of cache... Sorry that can´t be right.

The old individual vs community discusion... he don´t have more right than any other one.

<edit> LOL - If you are a little above the 200 WU limit, don´t worry, you and the Lunnatics team have a lot of credit with the community... we simple look on the other direction... but please not extrapolate...

---

@Tbar There is nothing wrong if you have 400 ready crunched WU in your cache, if you have thats simple shows you do a lot of science and probabily you have an internet access problem. What is wrong is if you "cheat" to optain a 400WU cache of "pendings" because this are above the project 200WU limit Period. The same aplies to add mores hosts to your farm, add more hosts is aproved by the project, so nobody could said that is "cheating" too...

Right, gotcha. It's fine if my AP only Mac works through 400 APs in 4 days as long as it's in groups of 200. If I run them all on the same host number though it's cheating. Right...that makes sense...not.

Nothing wrong with my FIOS internet, I can download 8 APs in a few seconds.
As of this moment, my two Mac hosts have around 140 APs left, that's enough for about a day and a half. Then the machine goes cold.

If I run them all on the same host number though it's cheating. Right...that makes sense...not.

Yes it not makes sense, but if you have 2 or + host numbers even on the same computer, nobody could blame you for nothing, it´s between the roules, so is right. I could agree the rules are foulty, but until someone change the roule, you are right. BTW There are few who actualy do that...
About the internet, i´m sure you understand what i try to said, 400 ready crunched WU without reporting indicates your host are not connecting with the project servers for a long time, normaly that happening only when you have some connection problem (other thread).

---

@ Juan regarding anon hosts

That's off-topic and I am sure somebody has a better explanation for that design decision, but the reason for allowing hosts to be anonymous is privacy. You lose users if they have to put their cards on the table.
Reasons for hiding your hosts are as diverse as the people.
You should respect the need for privacy.
Of course some people hide their hosts because they are up to no good and don't want their name associated with their wrongdoing.
There's always bad nuts.
And staff can always look it up anyway. It's just the wider public that can't satisfy their curiosity/voyeurism. And why should they? You don't invite the world into your living room either!

People always have reasons. They might be good or bad but there are always reasons. It's unfair to have a go at the law abiding majority because of a few crooks.

You'll be aware I must have reasons, too. Sometimes anonymity and unreal identities are tantamount to being able to do what you do - it's not just the bad guys that hide behind a mask.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

@william

I Agree with the privacy and have nothing against, i inclusive said hide the user list of computers is easy to understand.

But use anonymate to do "bad" things is wrong, i don´t said we all need to know who is the owner of the anonymous hosts, that´s sure only will flame the threads, what i try so said is, somebody (maybe something like you the mod´s do in the forums) need to keep the anonymous hosts "on track" so they not extrapolate their limits. They could remain anonymous without any problem, but they need to respect the roules as anyother member of the community, just that. Maybe something like a PM telling... hey what you are doing is not good... please avoid that practice... who knows?

LOL - I knows what you means... Did you ever see the movie... I know what you did the last summer...

---

People always have reasons. They might be good or bad but there are always reasons.

there, i think, is another state. reasons not be divided in good nr bad, but wrong and logically.
there always is peoples, who think logically, rational,a nd lot of that, who think alogically. and nobody can prove them, for example, a 2 x 2 not the same as 5 x 2, because is large category of that, who be sure, a 2 x 2 = many, and 10x10 also is many, as so, all that is equivalent.
as so, if someone not understand things, when it is clearly writed, further discussion is waste of time, as i understand from my experience.
as says old russian proverb - "если человек идиот - то это надолго".
i publish that only therefore, i be sure, some people there, maybe because a small living experience, try to prove some opponents again and again - looks like he thinks, he can prove peoples, who cannot point some logical structures and so on. as always, nothing personally, but that is reality, in what we lives....

---

...i can´t agree with you... you insist in point someone directly, that only could flame the thread.

I was about to say the same. Although maybe not nice enough to pass the "annoy or antagonize other people" & "messages that are deliberately hostile or insulting" rules of moderation :P

On a side note, it has made sense for years to build many & mediocre crunchers for Seti. I've said it before & I'll say it again: my netbook has never starved for work in over 3 years, except for when the Lab shut down for a week or more. A smart person would build as many single-card (let's say) GT 650 machines as they could afford if they really wanted to avoid the stress & babysitting.

I think David Anderson called it "the hot-rod mentality" in an Nvidia keynote, many years ago. Now I'm all for "hot-rods"... I really am.. they just haven't made sense for years, as far as Seti is concerned.

On another side note, these threads are just a prelude. I imagine they will get worse before they get better. It appears Raistmer has kicked a hornets nest and that is almost definitely a good thing. What he is doing manually now will likely be done automatically (based on his experience & findings) one day by Boinc. Sure it'll take like 5 years but whatever:)

On another-other side note, going up against a volunteer guru has the painfully obvious effect of biting the hand that feeds you. And it's OK to bite the hand that feeds you but you better have a faaaaar better argument than dogma.

Edit: And this from someone running stock apps and AP de-selected:)

You are right, but we are not talking here about error, i never say they are wrong because the error they produce, heavely error producing hosts if for a compleately diferent thread. I inclusive clearely said on one post, they inclusive return their crunched WU in time... but that not makes them right.

I´m sure if their pendings WU where distributed alond the community all this WU (besides timeout or error of course) will allready crunched and reported days ago, and those WU where allready cleared from the DB.

Lets focus on the 727 pending host. IIRC it has about 1100 WU pendings 5 days ago, so it crunches about 72 WU/day... Their cache still holds for 10 days more... so if my calculations are correct it work with 15 days of cache... Sorry that can´t be right.

The old individual vs community discusion... he don´t have more right than any other one.

<edit> LOL - If you are a little above the 200 WU limit, don´t worry, you and the Lunnatics team have a lot of credit with the community... we simple look on the other direction... but please not extrapolate...

But that's very my point, it's not because some have "lot of credit with community" (the single result of this "credit" would be to more deeply analyze what such peoples saying, it's not make them right just because of that but "average Joe" opinion and specialist opinion just can't have similar weight), but for completely other reasons. And I disagree with "if they would distributed among...". If those tasks were distrbuted to stock app they would be completed much later even than quite excessive hoarder with fast GPU can provide.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

If I run them all on the same host number though it's cheating. Right...that makes sense...not.

Yes it not makes sense, but if you have 2 or + host numbers even on the same computer, nobody could blame you for nothing, it´s between the roules, so is right. I could agree the rules are foulty, but until someone change the roule, you are right. BTW There are few who actualy do that...

And this is again I can't agree with. It's not "rule", it's technical limit (!).
It's not some law, rule and whatever. Some peoples on this board trying to make such rule from this limitation but it's not.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

You'll be aware I must have reasons, too. Sometimes anonymity and unreal identities are tantamount to being able to do what you do - it's not just the bad guys that hide behind a mask.

+1

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

as says old russian proverb - "если человек идиот - то это надолго".
i publish that only therefore, i be sure, some people there, maybe because a small living experience, try to prove some opponents again and again - looks like he thinks, he can prove peoples, who cannot point some logical structures and so on. as always, nothing personally, but that is reality, in what we lives....

LoL, perhaps so :) But it's hard to tolerate if you see witchhunting and wrong accusation on public, it's purely illogical, indeed, but sense of justice disobeys.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

but you better have a faaaaar better argument than dogma.

LoL, exactly :) I can be wrong as anyone can, but not because of dogma, +10.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Apologies, I was being sarcastic, which doesn't always translate well.
I have NO intention of rescheduling anything.
I was just trying to point out what would happen if many people did it and my take on it from a competitive point of view.


Reg

I allready use you and I as an example, just don´t remember the post, with your farm you could easely acumulate 5% of the entire avaiable AP WU´s if you decide to do competitive rescheduling and don´t care about timeouts. Now imagine Mark or MJS...(10% at least), in other words, we 4 could easely acumulate >1/4 of the total project avaiable AP WU, resulting in a compleate starvation to the rest of the community... and a lot of enemies of course...
But guy´s don´t panic, i´m sure no one of us will ever try to do that. So Reg go back to your evil plywood construction plans. My new 690 is allready in transit. :)

But it's hard to tolerate if you see witchhunting and wrong accusation on public, it's purely illogical, indeed, but sense of justice disobeys.

That´s i totaly agree. No witchhunting that only flames the threads.

---

it's technical limit (!).

I´m sure you knows this is not the real true...

Friends/Enemies. To avoid emotions, flaming, etc. I will do partialy what was sugested by Chris, I will left this thread not SETI... at least for now, and back to my normal life. Left to each one concience what is right or wrong to do with rescheduling, hope a lot choose the right path.

---

And staff can always look it up anyway.

Really, is that so? Staff as in Mods or Admins?

Staff, as in Berkeley payroll. Mods certainly can't.