Earliest Deadline First

Message boards : Number crunching : Earliest Deadline First
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · Next

AuthorMessage
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 18996
Credit: 40,757,560
RAC: 67
United Kingdom
Message 225480 - Posted: 4 Jan 2006, 2:43:31 UTC - in response to Message 225466.  

Andy when did you switch to "crunch3rs'" optimized apps/cc? were you using a V4 client immediately prior to that or nearly so?


I only use standard BOINC clients been using 5.2.x since 20th Oct,
And switchedto Crunch3r's app on 14 Dec.
ID: 225480 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 18996
Credit: 40,757,560
RAC: 67
United Kingdom
Message 225485 - Posted: 4 Jan 2006, 2:49:02 UTC

You're right that BOINC/Seti doesn't like Pent M's due to benchmark/large L2, Probably done over 500 units using Crunch3r's app. And DCF is 0.132685.
ID: 225485 · Report as offensive
John McLeod VII
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 15 Jul 99
Posts: 24806
Credit: 790,712
RAC: 0
United States
Message 225522 - Posted: 4 Jan 2006, 3:29:16 UTC

OK, one more time.

The HIGHEST priority is getting work reported before deadline.

The LOWEST priority is keeping the CPU scheduler running round robin.

In between these two is honoring resource shares over the long term.

Basically if the work is getting reported before deadline and the resource shares are honored over the long term (max (computation time / resource fraction) for all projects), I don't believe that anyone really cares that a host stays in EDF for most of that time.


BOINC WIKI
ID: 225522 · Report as offensive
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225529 - Posted: 4 Jan 2006, 3:38:34 UTC - in response to Message 225522.  

OK, one more time.

The HIGHEST priority is getting work reported before deadline.

The LOWEST priority is keeping the CPU scheduler running round robin.

In between these two is honoring resource shares over the long term.

Basically if the work is getting reported before deadline and the resource shares are honored over the long term (max (computation time / resource fraction) for all projects), I don't believe that anyone really cares that a host stays in EDF for most of that time.


So let's just pick any old algorithm and who cares if what it says is correct or not. It is bad style to not be able to work in a standard operating mode. Everything that Boinc is supposed to aid is broken because everyone has to be held to a little rule you made up for dial up. There should be a cache setting. There should be a connect rate. WHEN MY COMPUTER IS NOT OVERCOMMITTED IT SHOULDN'T SAY SO. I WANT THREE DAYS OR WHATEVER I SET WORTH OF WORK ON COMPUTER ANd ONLY THAT AMOUNT. I DON'T WANT A RANDOM SET OF WORK UNITS AT ANY GIVEN TIME. I SHOULD KNOW OR PREDICT WHAT THE SCHEDULER IS GOING TO DO FOR ANY DATE IN FUTURE.

ID: 225529 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 225534 - Posted: 4 Jan 2006, 3:43:53 UTC

So let's just pick any old algorithm and who cares if what it says is correct or not. It is bad style to not be able to work in a standard operating mode. Everything that Boinc is supposed to aid is broken because everyone has to be held to a little rule you made up for dial up. There should be a cache setting. There should be a connect rate. WHEN MY COMPUTER IS NOT OVERCOMMITTED IT SHOULDN'T SAY SO. I WANT THREE DAYS OR WHATEVER I SET WORTH OF WORK ON COMPUTER ANd ONLY THAT AMOUNT. I DON'T WANT A RANDOM SET OF WORK UNITS AT ANY GIVEN TIME. I SHOULD KNOW OR PREDICT WHAT THE SCHEDULER IS GOING TO DO FOR ANY DATE IN FUTURE.


Do you have anything helpful to tell John or is ranting and raving all you know?



Boinc....Boinc....Boinc....Boinc....
ID: 225534 · Report as offensive
KB7RZF
Volunteer tester
Avatar

Send message
Joined: 15 Aug 99
Posts: 9549
Credit: 3,308,926
RAC: 2
United States
Message 225538 - Posted: 4 Jan 2006, 3:48:32 UTC - in response to Message 225534.  


Do you have anything helpful to tell John or is ranting and raving all you know?


Ranting, raving, bitching, whining, moaning, complaining, not listening. He does it all. Better to just ignore him. You can say the same thing 500 times, never makes sense to him, won't change his mind, BOINC sucks in his eyes, everyone else is wrong.

:-)

Jeremy
ID: 225538 · Report as offensive
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225542 - Posted: 4 Jan 2006, 3:51:15 UTC - in response to Message 225534.  
Last modified: 4 Jan 2006, 3:58:42 UTC

So let's just pick any old algorithm and who cares if what it says is correct or not. It is bad style to not be able to work in a standard operating mode. Everything that Boinc is supposed to aid is broken because everyone has to be held to a little rule you made up for dial up. There should be a cache setting. There should be a connect rate. WHEN MY COMPUTER IS NOT OVERCOMMITTED IT SHOULDN'T SAY SO. I WANT THREE DAYS OR WHATEVER I SET WORTH OF WORK ON COMPUTER ANd ONLY THAT AMOUNT. I DON'T WANT A RANDOM SET OF WORK UNITS AT ANY GIVEN TIME. I SHOULD KNOW OR PREDICT WHAT THE SCHEDULER IS GOING TO DO FOR ANY DATE IN FUTURE.


Do you have anything helpful to tell John or is ranting and raving all you know?



Well he said why do we need to worry about anything but Highest priority and lowest priority.

The highest priority and lowest priority are derived from rules that give wrong answers.

I have a defined set of things I would like. They are based on the fact that the scheduler says things that are wrong. The answer given indicates it's too hard to do what is right. John can admit that the computer is overcommitted isn't correct most of the time. The biggest issue is the 2 times correct rate rule. You probably never had heard from me if that one rule wasn't instituted. It's annoying that I can't controll what goes into cache. I wanted to volunteer time to a few projects. I can't stand something is accepted as normal that states something that is obviously wrong.
ID: 225542 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 225565 - Posted: 4 Jan 2006, 4:12:38 UTC

Ranting, raving, bitching, whining, moaning, complaining, not listening. He does it all. Better to just ignore him. You can say the same thing 500 times, never makes sense to him, won't change his mind, BOINC sucks in his eyes, everyone else is wrong.

:-)

Jeremy


Thanks Jeremy......my ignore list has been updated! (not you)


Boinc....Boinc....Boinc....Boinc....
ID: 225565 · Report as offensive
Profile Tern
Volunteer tester
Avatar

Send message
Joined: 4 Dec 03
Posts: 1122
Credit: 13,376,822
RAC: 44
United States
Message 225573 - Posted: 4 Jan 2006, 4:23:41 UTC - in response to Message 225542.  

I have a defined set of things I would like. They are based on the fact that the scheduler says things that are wrong. The answer given indicates it's too hard to do what is right. John can admit that the computer is overcommitted isn't correct most of the time. The biggest issue is the 2 times correct rate rule. You probably never had heard from me if that one rule wasn't instituted. It's annoying that I can't controll what goes into cache. I wanted to volunteer time to a few projects. I can't stand something is accepted as normal that states something that is obviously wrong.


The defined set of things that you would like is worthless crap. The scheduler is correct but you are too stupid to understand it, or too pig-headed to listen to what 100 people have been telling you, for weeks. Nobody cares if you can or can't control what goes into cache, or if you want to volunteer or if you just go away (actually, most of us would prefer you just go away). The things that are accepted as normal are only "obviously wrong" to your deranged, mentally ill little mind. Everyone else here understands this but YOU! If you are the ONLY person on the planet who believes something, then here's a hint - YOU'RE PROBABLY WRONG! Your constant verbal flatulance on these boards is accomplishing nothing.

The scheduler DOES have problems; they are minor ones; they are being worked on; they are NOT the ones you are inventing out of thin air. EDF is not an "error", and "overcommitted" isn't either. BOINC does not do "DOS attacks" on your private network with the broken router. But maybe that will be the factor that finally gets you to uninstall it and go back to running something you can actually understand, like Solitaire.

Have a nice day.
ID: 225573 · Report as offensive
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225579 - Posted: 4 Jan 2006, 4:31:37 UTC - in response to Message 225573.  

I have a defined set of things I would like. They are based on the fact that the scheduler says things that are wrong. The answer given indicates it's too hard to do what is right. John can admit that the computer is overcommitted isn't correct most of the time. The biggest issue is the 2 times correct rate rule. You probably never had heard from me if that one rule wasn't instituted. It's annoying that I can't controll what goes into cache. I wanted to volunteer time to a few projects. I can't stand something is accepted as normal that states something that is obviously wrong.


The defined set of things that you would like is worthless crap. The scheduler is correct but you are too stupid to understand it, or too pig-headed to listen to what 100 people have been telling you, for weeks. Nobody cares if you can or can't control what goes into cache, or if you want to volunteer or if you just go away (actually, most of us would prefer you just go away). The things that are accepted as normal are only "obviously wrong" to your deranged, mentally ill little mind. Everyone else here understands this but YOU! If you are the ONLY person on the planet who believes something, then here's a hint - YOU'RE PROBABLY WRONG! Your constant verbal flatulance on these boards is accomplishing nothing.

The scheduler DOES have problems; they are minor ones; they are being worked on; they are NOT the ones you are inventing out of thin air. EDF is not an "error", and "overcommitted" isn't either. BOINC does not do "DOS attacks" on your private network with the broken router. But maybe that will be the factor that finally gets you to uninstall it and go back to running something you can actually understand, like Solitaire.

Have a nice day.



EDF is not the error it is the error mode it goes into when it complains that the computer is overcommitted. Overcommitted is what my computer has not been. Like I said In the short term the easiest thing would put a switch to turn off the 2 times connect rule. I just allowed BOINC to work and my network stopped responding. It was not acting like it should of. The only messages I get are scheduler messages of http 500 server error. Your ignorance is showing.
ID: 225579 · Report as offensive
KB7RZF
Volunteer tester
Avatar

Send message
Joined: 15 Aug 99
Posts: 9549
Credit: 3,308,926
RAC: 2
United States
Message 225581 - Posted: 4 Jan 2006, 4:36:24 UTC - in response to Message 225579.  
Last modified: 4 Jan 2006, 4:37:32 UTC

I just allowed BOINC to work and my network stopped responding. It was not acting like it should of. The only messages I get are scheduler messages of http 500 server error. Your ignorance is showing.


Ya know, why is it that there are people, who have a whole hell of a lot more computers attached, running through a network, crunching happily, reporting work, not bitching about this??? You say u have 5 computers, yet you have hardly over 12k credit, and BOINC stopped your network access? I'm no network person, nor do I care to be, but when I can see several others running BOINC just fine on 10 computers or more, on a network, with no problems, I think your full of crap. Your best solution, as many have pointed out, but your too dense to listen, is to get a new router, and fix the problem on YOUR end.

Reading your posts gives me more than a headache, I need a beer now.

Jeremy
ID: 225581 · Report as offensive
Profile Tern
Volunteer tester
Avatar

Send message
Joined: 4 Dec 03
Posts: 1122
Credit: 13,376,822
RAC: 44
United States
Message 225588 - Posted: 4 Jan 2006, 4:44:23 UTC - in response to Message 225581.  
Last modified: 4 Jan 2006, 4:48:39 UTC

Reading your posts gives me more than a headache, I need a beer now.


The ones from a couple of weeks ago were worth a beer. At this point, there's some bourbon in the cabinet...

The sad part is, that I think somewhere back there, before the troll hijacked the thread, Winterknight may have had a valid problem. I got lost when there were (I think?) two different computers being discussed, but then there were so many irrelevant messages, I finally had to give up trying to figure it out. Winterknight, if you still have the problem or want to discuss it, my email address has been posted a few times, or you can more easily find me (and other EDF help) on the BOINC boards, or Einstein, or Rosetta, or anywhere that's less zoo-like.
ID: 225588 · Report as offensive
Profile ksnash

Send message
Joined: 28 Nov 99
Posts: 402
Credit: 528,725
RAC: 0
United States
Message 225594 - Posted: 4 Jan 2006, 4:50:21 UTC - in response to Message 225581.  

I just allowed BOINC to work and my network stopped responding. It was not acting like it should of. The only messages I get are scheduler messages of http 500 server error. Your ignorance is showing.


Ya know, why is it that there are people, who have a whole hell of a lot more computers attached, running through a network, crunching happily, reporting work, not bitching about this??? You say u have 5 computers, yet you have hardly over 12k credit, and BOINC stopped your network access? I'm no network person, nor do I care to be, but when I can see several others running BOINC just fine on 10 computers or more, on a network, with no problems, I think your full of crap. Your best solution, as many have pointed out, but your too dense to listen, is to get a new router, and fix the problem on YOUR end.

Reading your posts gives me more than a headache, I need a beer now.

Jeremy

CPDN is working fine. Setiathome is only thing that is uploading but not finishing. Why do people not want it to work better. They are not going to give any cache control. If the 2 times connect rate rule were not forced. I'd have one CPDN 15 Seti@home and 10 Einstein. The computer can finish it in about a week. At least I would know what would be there. I get widely different workunits at any time.

ID: 225594 · Report as offensive
KB7RZF
Volunteer tester
Avatar

Send message
Joined: 15 Aug 99
Posts: 9549
Credit: 3,308,926
RAC: 2
United States
Message 225598 - Posted: 4 Jan 2006, 4:55:13 UTC - in response to Message 225594.  


CPDN is working fine. Setiathome is only thing that is uploading but not finishing. Why do people not want it to work better. They are not going to give any cache control. If the 2 times connect rate rule were not forced. I'd have one CPDN 15 Seti@home and 10 Einstein. The computer can finish it in about a week. At least I would know what would be there. I get widely different workunits at any time.


CPDN works fine for me, SETI works fine for me, Einstein works fine for me, SIMAP works fine for me, and Rosetta works fine for me. Everything uploads just fine for me, unless theres a "server problem" on the projects side. I have uploaded, downloaded, uploaded some more, downloaded some more, all with no problems. As it has been suggested numerous times, LOOK AT YOUR NETWORK PROBLEMS and fix it. How many threads does there have to be explaining the same damn thing over and over to you before you understand that its not no one elses problem but YOURS

Wheres the Jack Daniels at???
ID: 225598 · Report as offensive
John McLeod VII
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 15 Jul 99
Posts: 24806
Credit: 790,712
RAC: 0
United States
Message 225600 - Posted: 4 Jan 2006, 4:58:19 UTC - in response to Message 225594.  

CPDN is working fine. Setiathome is only thing that is uploading but not finishing. Why do people not want it to work better. They are not going to give any cache control. If the 2 times connect rate rule were not forced. I'd have one CPDN 15 Seti@home and 10 Einstein. The computer can finish it in about a week. At least I would know what would be there. I get widely different workunits at any time.

So, why does a different mix of work at different times bother you? I have one computer attached to more than 20 projects (not all of which are giving out work yet) and the mix of results on the host vary from day to day a great deal.

I have one host that is in EDF with a 6 month deadline Sulphur result that will take -- 5 and a half or so months to complete - guaranteed EDF, the other projects will make up the time later.

The work gets done on time (except for the time that a project handed out work that had estimates that were a factor of 800 low - got too much to complete by deadline that time but BOINC did as well as possible with the garbage data that it got). My resource shares are honored over a few months (but not every day).

BTW, detecting a normally disconnected host is a bit harder than it seems, and having a setting for this with only 3 venues available is also not as helpful as it seems.


BOINC WIKI
ID: 225600 · Report as offensive
Profile kev1701e
Avatar

Send message
Joined: 28 Dec 99
Posts: 138
Credit: 10,216,553
RAC: 0
United States
Message 225632 - Posted: 4 Jan 2006, 5:33:41 UTC - in response to Message 225579.  

I have a defined set of things I would like. They are based on the fact that the scheduler says things that are wrong. The answer given indicates it's too hard to do what is right. John can admit that the computer is overcommitted isn't correct most of the time. The biggest issue is the 2 times correct rate rule. You probably never had heard from me if that one rule wasn't instituted. It's annoying that I can't controll what goes into cache. I wanted to volunteer time to a few projects. I can't stand something is accepted as normal that states something that is obviously wrong.


The defined set of things that you would like is worthless crap. The scheduler is correct but you are too stupid to understand it, or too pig-headed to listen to what 100 people have been telling you, for weeks. Nobody cares if you can or can't control what goes into cache, or if you want to volunteer or if you just go away (actually, most of us would prefer you just go away). The things that are accepted as normal are only "obviously wrong" to your deranged, mentally ill little mind. Everyone else here understands this but YOU! If you are the ONLY person on the planet who believes something, then here's a hint - YOU'RE PROBABLY WRONG! Your constant verbal flatulance on these boards is accomplishing nothing.

The scheduler DOES have problems; they are minor ones; they are being worked on; they are NOT the ones you are inventing out of thin air. EDF is not an "error", and "overcommitted" isn't either. BOINC does not do "DOS attacks" on your private network with the broken router. But maybe that will be the factor that finally gets you to uninstall it and go back to running something you can actually understand, like Solitaire.

Have a nice day.



EDF is not the error it is the error mode it goes into when it complains that the computer is overcommitted. Overcommitted is what my computer has not been. Like I said In the short term the easiest thing would put a switch to turn off the 2 times connect rule. I just allowed BOINC to work and my network stopped responding. It was not acting like it should of. The only messages I get are scheduler messages of http 500 server error. Your ignorance is showing.


ksnash, is this computer of yours doing CPDN and Einstein as well as SETI? If so, then it is overcommitted. In just SETI terms, running 24/7, it has ~13 days worth of work on it (86 units in process * 13000 seconds/unit). I would be surprised if BOINC wasn't complaining of being overcommitted.

kev



kev

X2 4400+,4200+ @2.75GHz, XP1800+ @1.65GHz, P4 @1.6GHz
ID: 225632 · Report as offensive
Profile Paul D. Buck
Volunteer tester

Send message
Joined: 19 Jul 00
Posts: 3898
Credit: 1,158,042
RAC: 0
United States
Message 225731 - Posted: 4 Jan 2006, 11:36:12 UTC

What I liked especially much is a person telling John that he does not understand how the CPU Scheduler works. That is priceless.

Oh, and the red light that tells me that the cruise control is on is signaling an error condition and I need to have my car checked. Just to make a better example than the Oil light ...

Oh, and we have forgotten that the point is to download, process, and return work to the project. And, to have as much of this work returned within the deadlines as possible. Not to have large queues on the participant's computers.

@WinterKnight,

One of your last posts finally made me see, I think, what JKeck has been trying to get through my thick skull ... have you opened another thread on it as Bill asked? If so, where?

Well, I need to think on this a little bit, but, I think in the essentials the point is that the current EDF is too simplistic, or, at least, the way we are handling it is too simplistic ...
ID: 225731 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 18996
Credit: 40,757,560
RAC: 67
United Kingdom
Message 225734 - Posted: 4 Jan 2006, 11:47:49 UTC - in response to Message 225731.  
Last modified: 4 Jan 2006, 11:48:07 UTC

What I liked especially much is a person telling John that he does not understand how the CPU Scheduler works. That is priceless.

Oh, and the red light that tells me that the cruise control is on is signaling an error condition and I need to have my car checked. Just to make a better example than the Oil light ...

Oh, and we have forgotten that the point is to download, process, and return work to the project. And, to have as much of this work returned within the deadlines as possible. Not to have large queues on the participant's computers.

@WinterKnight,

One of your last posts finally made me see, I think, what JKeck has been trying to get through my thick skull ... have you opened another thread on it as Bill asked? If so, where?

Well, I need to think on this a little bit, but, I think in the essentials the point is that the current EDF is too simplistic, or, at least, the way we are handling it is too simplistic ...


I'm still here, just, as Bill said this is making my 'ead 'urt. I haven't as yet takien this to another place, but will soon. I've got things to do at moment.

And, which post I made that many, I think I'm going to have to cut and paste bits into one document so I can see, the arguments and counter-arguments more clearly, from those that have genuine input.

ID: 225734 · Report as offensive
Astro
Volunteer tester
Avatar

Send message
Joined: 16 Apr 02
Posts: 8026
Credit: 600,015
RAC: 0
Message 225748 - Posted: 4 Jan 2006, 13:04:08 UTC

Andy, Did you recently UP your cache, or just add PPAH? If so, that's why you're now experiencing these things
ID: 225748 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 18996
Credit: 40,757,560
RAC: 67
United Kingdom
Message 225766 - Posted: 4 Jan 2006, 13:41:22 UTC - in response to Message 225748.  

Andy, Did you recently UP your cache, or just add PPAH? If so, that's why you're now experiencing these things

No Cache is as it was and not joined any thing new on this computer since SetiB where I am 351.
ID: 225766 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : Number crunching : Earliest Deadline First


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.