LHC@home

If you need the results fast, just shorten the deadline, that's what it's for, and set IR at a decent level.

As I keep mentioning, the best solution to keep the project's decision of the IR set to where it's set and to keep you folks who are up in arms over this happy is to just implement the server-side aborts. No, it won't 100% eliminate "redundant results", but it will cut out a sizeable portion of them...

all these arguments, day after day started because ONE person cant get everyone else to think like him,.. what a shame!!!
@ Keith t,....I think I'll use that link now!

---

A clear conscience is usually the sign of a bad memory

all these arguments, day after day started because ONE person cant get everyone else to think like him,.. what a shame!!!
@ Keith t,....I think I'll use that link now!

Thats right, seems like he conducts. I looked around and discovered that Dagorath is arguing (quite unkind) with other users in several other project forums (i.e. ABC@home, BOINC,...). And that he´s unhappy with the way of some projects are managed (i.e. QMC@home), so why you don't get your own project instead?

Seems like squealing, but i had to quote this nice post of Dagorath from the ABC@Home forum:

Agreed! Now it's time to battle waste and retarded thinking at LHC where Neasan has sunk his own ship as I suspected he eventually would.

P.S.: Why you think i am one of your students. You aren't that popular, are you?

---

"Thats right, seems like he conducts. I looked around and discovered that Dagorath is arguing (quite unkind) with other users in several other project forums (i.e. ABC@home, BOINC,...). And that he´s unhappy with the way of some projects are managed (i.e. QMC@home), so why you don't get your own project instead?"


Seems like squealing, but i had to quote this nice post of Dagorath from the ABC@Home forum:

Agreed! Now it's time to battle waste and retarded thinking at LHC where Neasan has sunk his own ship as I suspected he eventually would.

It all adds up to, you can't/won't answer the question that started this thread so you attack the man/men who asked it and try to destroy his/their credibility.

Dagorath may or may not be guilty of what you accuse him of, however, it's irrelevant.

Answer the question and don't attack the people who asked it!

I hope you can communicate with your fellow scientists better than Neasan and yourself have communicated with Dagorath and I or the possibility of a disaster on the Swiss/French border in 2008 becomes a lot more likely.

---

http://lhcathome.cern.ch/lhcathome/workunit.php?wuid=1726535

Even 5 computers are not grouping 3 same answers...

May you consider that this project could need this IR = 5 ?

An other question : why have the scientists to wait that a WU has been deleted to take information of it ?

If the quorum is okay, why have they to wait that the 2 others answers. The result remain in database to collect the delayed returned results, but in fact, scientists could use the results at the moment the quorum is okay...

If I am right, how a IR=5 can delay the obtention of results ?
It's only consume space to keep the result in database until everyone has responsed, or passed the deadline .

---

------
Thrr-Gilag Kee'rr

L'Alliance Francophone

No, its not quite true, Dagorath only argues if there is some thing blatantly stupid that even a blind person can see :) the question that should be asked is, why do Alex or Neasan not come up with a plausible counter argument as to why they need an IR of 5 :( Alex and Neasan are the ones in fact prolonging the argument by not doing so. :) if seen in perspective the WU's where the protons crash on start up :) IR of 5 is silly. so why not further process the worth while WU's if required and the duds can be put in the dust bin on a IR of 2? as a start may save a few CPU cycles.
Dagorath keep the tone ;) down but fight you battles ;)
Regards
Masud.

"Thats right, seems like he conducts. I looked around and discovered that Dagorath is arguing (quite unkind) with other users in several other project forums (i.e. ABC@home, BOINC,...). And that he´s unhappy with the way of some projects are managed (i.e. QMC@home), so why you don't get your own project instead?"


Seems like squealing, but i had to quote this nice post of Dagorath from the ABC@Home forum:

Agreed! Now it's time to battle waste and retarded thinking at LHC where Neasan has sunk his own ship as I suspected he eventually would.

It all adds up to, you can't/won't answer the question that started this thread so you attack the man/men who asked it and try to destroy his/their credibility.

Dagorath may or may not be guilty of what you accuse him of, however, it's irrelevant.

Answer the question and don't attack the people who asked it!

I hope you can communicate with your fellow scientists better than Neasan and yourself have communicated with Dagorath and I or the possibility of a disaster on the Swiss/French border in 2008 becomes a lot more likely.

---

No, its not quite true, Dagorath only argues if there is some thing blatantly stupid that even a blind person can see :) the question that should be asked is, why do Alex or Neasan not come up with a plausible counter argument as to why they need an IR of 5 :( Alex and Neasan are the ones in fact prolonging the argument by not doing so. :) if seen in perspective the WU's where the protons crash on start up :) IR of 5 is silly. so why not further process the worth while WU's if required and the duds can be put in the dust bin on a IR of 2? as a start may save a few CPU cycles.
Dagorath keep the tone ;) down but fight you battles ;)
Regards
Masud.

Well Neason or Alex???

---

No, its not quite true, Dagorath only argues if there is some thing blatantly stupid that even a blind person can see :) the question that should be asked is, why do Alex or Neasan not come up with a plausible counter argument as to why they need an IR of 5 :( Alex and Neasan are the ones in fact prolonging the argument by not doing so. :) if seen in perspective the WU's where the protons crash on start up :) IR of 5 is silly. so why not further process the worth while WU's if required and the duds can be put in the dust bin on a IR of 2? as a start may save a few CPU cycles.
Dagorath keep the tone ;) down but fight you battles ;)
Regards
Masud.

Well Neason or Alex???

Maybe I remember incorrectly,but they already answered this in this thread.

Well Neason or Alex???

In this very thread there has been given two very plausible answers to that question. From what I can see, it boils down to the following:

This project has a very high host / wu ratio. This means that the extra replications (>3) can be computed essentially "for free" (as seen by the server). Imagine running one WU on an 8-way SMP box... Assuming a CPU bound program, using an IR of 3, 5 or even 7 wouldn't make much of a difference in terms of completion time of the WU. So the extra replications can be done "for free".

If the error rate was zero, a high IR would lead to a shorter time to completion, because there's a higher chance that three of those replications would get crunched by fast computers.

Since the error rate is not zero, using a high IR still leads to a shorter completion time, since there's a higher chance that at least 3 of the 5 results agree. This is especially true, considering what Alex mentioned regarding Intel vs AMD results.

As mentioned in this post, it would seem that the above theory has indeed been confirmed during testing.

The "problem" with Dagorath's analysis, as far as I can see, is that it assumes a lower host / wu ratio, and so it does not apply to the current situation (I believe he said so himself). If the extra replications cannot be computed "for free", then his analysis holds.

Personally, I'm all for completing the LHC as soon as possible, regardless of how many "wasted" CPU cycles my computer must perform. If you'd rather spend your CPU cycles on something else, then do it :)

The "problem" with Dagorath's analysis, as far as I can see, is that it assumes a lower host / wu ratio, and so it does not apply to the current situation (I believe he said so himself). If the extra replications cannot be computed "for free", then his analysis holds.

Personally, I'm all for completing the LHC as soon as possible, regardless of how many "wasted" CPU cycles my computer must perform. If you'd rather spend your CPU cycles on something else, then do it :)

The "problem" with the "for free" argument is that whilst the extra WUs are free as far as LHC is concerned, they are not free as far as other BOINC projects are concerned - every LHC WU crunched after Q is achieved is wasting crunching time that could have been donated to another project. Fast turnarounds can be achieved by IR > Q, but they can also be achieved (without wasting cycles) by short deadlines. It is a pity that LHC can't (or won't?) use that method instead.

However, since the admins have said they will look at this issue again when other applications are added to the project, I'm happy to keep contributing cycles for the being. :-)

---

Fast turnarounds can be achieved by IR > Q, but they can also be achieved (without wasting cycles) by short deadlines. It is a pity that LHC can't (or won't?) use that method instead.

OTOH, short deadlines lead to Earliest Deadline First / High Priority, which irritates the snot out of a large percentage of users who refuse to understand that their resource allocations are honored over the long term, just not on a second-by-second basis in lockstep with the switch interval. Those users end up grumbling about how the project is being "rude" and how they "want to have control of their computer instead of the project having control", etc, etc, etc...

Frankly, the project has had their say in regards to this. They said they would revisit the topic when the BOINC version on the server has been updated. The BOINC version has not yet been updated, thus this topic has not been revisited. If you want to push this issue, then what you should ask for are details about the progress of the BOINC server upgrade.

Beyond that, individuals who find this single subject such a catastrophic issue should probably heed the advice that was given and contribute their processing time to another project. Of all the things to get worked into a lather about, this is not one that I would choose... If you feel other projects are being "shorted", then vote with your clock cycles

---

OTOH, short deadlines lead to Earliest Deadline First / High Priority, which irritates the snot out of a large percentage of users who refuse to understand that their resource allocations are honored over the long term, just not on a second-by-second basis in lockstep with the switch interval. Those users end up grumbling about how the project is being "rude" and how they "want to have control of their computer instead of the project having control", etc, etc, etc...

Would LHC would miss those users any more than it misses those who refuse to crunch while IR > Q? After all, even with the latest avalanche of WUs, there is still more processing power attached to this project than it needs - witness the arguments over the maximum daily WU quota. :-)

Those who are upset with LHC have probably already left / suspended the project. Hopefully they will come back and join the rest of us at some point in the future (when more power is required).

---

Fast turnarounds can be achieved by IR > Q, but they can also be achieved (without wasting cycles) by short deadlines.

Short deadlines won't fix the AMD / Intel issue though... But I agree that the deadlines are very "lax", and could be tightened a lot.

Would LHC would miss those users any more than it misses those who refuse to crunch while IR > Q?

You fail to appreciate the fervor of those who demand that their resource allocations be honored even when viewed by very short timeframes. They are just as, if not more, ridiculously indignant as the people bellyaching about IR > Q. More than that, I am relatively certain that the people who demand resource allocation share honoring greatly outnumber those who are irate about IR > Q. The topic comes up frequently on SETI and Einstein, so much so that I was able to get the ball rolling on suggesting to the Einstein staff that they extend deadlines out some to relieve the pressure on older / slower / less dedicated hosts.

If one *must* tinker with the deadlines, then I'd suggest trimming no more than 1 day off of it.

Also, as an "unintended side-effect", since resource allocations are probably low for this project, you'll start seeing an increase of reporting that work was available, but BOINC determined that the host couldn't complete in time, which would again cause bellyaching about "fair distribution" because they were not informed that there was a change *AND* they don't see why they should have to make a change because it is *THEIR COMPUTER*, dammit!

Trust me, the deadlines are fine. People need to just quit their bitching about this and lobby to get the BOINC server-side components upgraded.

---