Uploading

Message boards : Number crunching : Uploading
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 · Next

AuthorMessage
edtx

Send message
Joined: 25 Nov 03
Posts: 9
Credit: 43,058
RAC: 0
United States
Message 137675 - Posted: 17 Jul 2005, 9:53:05 UTC

Surely one of the ways to cut some of the communications problems is to stop sending units to those still using Boinc CC Versions before 4.19. They are causing download errors, which will upset people who crunch the unit sucessfully because there isn't enough good returns to form a quorum. I've already aborted 3 units for this reason, I check because I've crunched these before (on old computer 14hrs/unit) only for the system to say too many errors, unit abandoned.

And maybe they should rethink the upload sequence. Rather that blindly sending the results to an overloaded/faulty server, a short message be sent from the host requesting bandwidth to upload x number of units with a deadline of MM/DD/YY HH:mm
and then if the server was busy it would give preference to those near the deadline until the problem at the server is resolved.

I've got one unit that finished crunching at about 14:00 GMT on the day of the breaker replacement it's still sitting here trying to upload, under the transfers/time col. it says 02:34. I'm glad I'm not on a modem with auto-connect. But other units seem to finish and report within minutes.

Andy






Now, this suggestion makes a lot of sense to me.


I have Wu's on boroadband that are still trying to upload and I have Wu's on modem that are still trying t load, but when you look at the the server status page it says all servrs are up and running, and there is no backlog. Are Seti and I even on the same page?
ID: 137675 · Report as offensive
Diablos
Volunteer tester

Send message
Joined: 7 Jun 05
Posts: 2
Credit: 10,509
RAC: 0
United Kingdom
Message 137681 - Posted: 17 Jul 2005, 10:25:24 UTC - in response to Message 137672.  

The point is we donate our CPU time voluntarily to do the work that Seti asks of us volunteers, yet Seti is always down, and I mean down when compared to the other Boinc applications. They may have problems once in a while, but not constantly as Seti does. I am about to drop Seti and just do the work of the other Boinc projects


To the person, who pretty much posted that that the world does not end if they cannot upload a few Seti Wu's, I would like to say to you that I hope we don't make contact with alien life anytime soon as it would probaly bore you.


You too, Good Morning!



I wouldnt be here if i thought it would bore me , but maybe i shouldn`t read forums when i just woke up lol (^_^)
ID: 137681 · Report as offensive
Profile Warren

Send message
Joined: 13 Dec 00
Posts: 16
Credit: 57,768
RAC: 0
New Zealand
Message 137690 - Posted: 17 Jul 2005, 11:18:23 UTC

I for one will wait a few more days but have changed the amount of time seti gets to give the other projects more time. It is the weekend so lets see what happens in the next few days. If nothing is resolved then I have lost nothing. I give my computer time free to the team, so what the team does with it I am not overly concerned.

I am sure that in time all the bugs in the system will be solved and then we will have to find something else to bitch about.

Have a good night all.
A simple country boy waiting for...
A sign of any kind.
ID: 137690 · Report as offensive
Brian Silvers

Send message
Joined: 11 Jun 99
Posts: 1681
Credit: 492,052
RAC: 0
United States
Message 137691 - Posted: 17 Jul 2005, 11:20:05 UTC - in response to Message 137642.  



Take a look around any games/programs forums and see how many people neglect to read the readme text where things are explained. before you bring up that splendid idea.



I think what is being asked for is either context-sensitive help or bubble help. As a developer for a Point-of-Sale software system for a US Fortune 500 company, my team has to take into account the user-friendliness aspect...or else we will end up getting tickets being called in from the stores due to not understanding how to do things. We have had such issues in the past...because WE as developers understood that you should go from Point A to Point B to Point C, however sometimes it is not so clear to the end user, so they start at A, go to B, then get frustrated because they went to D, then W, then H...and couldn't figure out how to get back to C, so they started over... then complain about how "complex" or "slow" the system is... The difference in my environment is the store has an external customer standing there, so they will keep going until they figure it out or hand-write a sales ticket. People here if they don't understand it or think that it is not working right may end up just giving up and going elsewhere.

Respectfully...

Brian
ID: 137691 · Report as offensive
Profile ML1
Volunteer moderator
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 21253
Credit: 7,508,002
RAC: 20
United Kingdom
Message 137699 - Posted: 17 Jul 2005, 12:29:35 UTC - in response to Message 137630.  
Last modified: 17 Jul 2005, 12:32:21 UTC

...But no, most will not even try to figure out how things work, but they will manage to come on here and say that the old was Better!

Windows 95 was better than 3.11, Windows 98 was better than 95. XP is apparently better than 2000. I'll always ask them why they want to upgrade their whole computer, hardware and software, but refuse to go with a better program to find Seti. If someone can answer that to me, I'll stop asking.

Plain. Simple.

Notice how in the MS world, each newer OS version introduced ever more MS-Virus vulnerabilities? (I finally gave up on the fiasco at WinXP. Too much user time wasted!)

And then the hardware REQUIRED upgrading to cope with the MS software bloat. However, this push has also been very good for s@h for the ever increasing excess of idle CPU cycles :)

I agree, that Boinc is the good way forward and is a good improvement on functionality over classic. Unfortunately, it is also vulnerable to problems with those new features until the system settles.

Meanwhile, the MS OSes seem never to settle. To my mind they remain a perpetual ever more expensive frustrating mess.

(OK, so I use Linux :) )

Good luck,
Martin

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 137699 · Report as offensive
Miklos M.

Send message
Joined: 5 May 99
Posts: 955
Credit: 136,115,648
RAC: 73
Hungary
Message 137705 - Posted: 17 Jul 2005, 13:02:41 UTC

"July 15, 2005
The data server is continuing to drop connections, which is very unusual this long after an outage. We are working to figure out why."
It is the 17th and I am losing wu's due to the delays and drops. It had been this way most of the week. It is a shame! What is the point of getting new units if the done ones cannot be returned for credit? Busy work? Waste of electricity?

Nick
ID: 137705 · Report as offensive
Profile Tigher
Volunteer tester

Send message
Joined: 18 Mar 04
Posts: 1547
Credit: 760,577
RAC: 0
United Kingdom
Message 137715 - Posted: 17 Jul 2005, 13:35:10 UTC

ML1 you are right for sure....I still use MS OSs and Linux...they get me where I need to be but it can be hard work for sure and $$$ with MS.

I also agree BOINC is the right approach. I still question though (yes....endlessly) -

UCB's means of communicating problems to us and how they will be fixed. I feel we are left hanging in mid air for days on end.

I still cannot understand why it has taken so long to move classic off and re-deploy the hardware for the premier system - boinc. This would be so useful right now.

I still don't understand why there is not a seperate UL/DL server given the work it has to get through - especially after down time periods.

There needs to be better documentation for folk to refer to on how it works. Wiki is a start so well done Paul.

I still think a model of the whole system would be appropriate to highlight weaknesses and to better identify the efftects of points of failure; thus improving the ruggedness of the system.

I still believe the overall availability of the entire system is lower than it should be hence reliability can be considered poor.

I understand the design goals of doing other projects - some people only want seti.

I understand the grant given to this work was commended for its interesting ways of achieving good user engagement - I read the report - but I feel these (which included this forum, credits, competition and teams) are being neglected somewhat and users are becoming disenchanted. Disengagement threatens!

I think hardware and software testing (including integration testing) is the poor relation. This forum is littered with examples of new core clients etc not working. I have to ask how is the testing conducted? Do they need a bigger audience? More time invested here is always worth while in my experience.

I believe the impact of classic coming across is unknown. I have asked them to estimate it but I doubt this has happened. Right now I think the few that have migrated add dirctly to long recovery times. It WILL get worse. There must be a plan including some excellent PR.


To address the High Priest / Priestesses & Gurus (aka Cheer Leaders) criticisms of the above or anything else I say:

- This is everyone's forum.....you do not own it.
- All are free to express their views and should not be the subject of petty harrassment and rubbishing.
- Your view is a view....make it...its no better than anyone else's and no worse........don't deprecate other peoples' views.





ID: 137715 · 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 137734 - Posted: 17 Jul 2005, 14:25:38 UTC - in response to Message 137715.  
Last modified: 17 Jul 2005, 14:26:41 UTC

{edit} though posted in response to one other post, this is not directed at any one person ...{/edit}

There needs to be better documentation for folk to refer to on how it works. Wiki is a start so well done Paul.

Thank you ... I lobby for more clarity here, We have a model where statistics are done external to the projects, and I would argue that we should either clearly make that the choice, or the other way where the documentation is clearly hosted by the BOINC Home site under direction of the BOINC Development team. To this point in time I have been, um, less than successful.

I still think a model of the whole system would be appropriate to highlight weaknesses and to better identify the affects of points of failure; thus improving the ruggedness of the system.

I ask again, since you think this is needed, can you, or you and I, or you, I, and the rest of the world, develop said model? Seriously, if it is something that you feel is needed, well, build it ... I think Documentation is important and so, I write it ... health permitting, believe it or not, I spend around 12 hours a day working on it. Want to know how hard it is ... ask any of the other editors ...

(aside: Azwoody, we are at about 977 pages and growing.)

With BOINC being Open Source comes a certain amount of freedom, coupled with a certain amount of responsibility. I build documentation because that is something I know how to do, I have the time, and I have the tools. You think we need a model, fine, build one. The developers have a lot on their plate and if this is a needed tool, lets build them one ...

I understand the design goals of doing other projects - some people only want SETI.

Agreed, the only issue I have with those participants electing the choice is that they are upset of the limitation that each and every BOINC project states is a given if you elect to one, and only one, project. You elect to only run one project, you are going to run out of work, you are going to have idle computers, that is the given for the choice you elect. Make that choice, LIVE with the consequences ...

To address the High Priest / Priestesses & Gurus (aka Cheer Leaders) criticisms of the above or anything else I say:

- This is everyone's forum.....you do not own it.
- All are free to express their views and should not be the subject of petty harassment and rubbishing.
- Your view is a view....make it...its no better than anyone else's and no worse........don't deprecate other peoples' views.

Agreed. And I struggle very hard not to ... so, why aren't you living up to your own standard and deprecate my views... because it does not agree with yours?

I don't call you names, why do you, and some others call me, and other people names? In context, the word "cheerleader" is neutral. Within the message boards it is an appellation that has connotations of ignorance, bias, loss of contact with reality, and intolerance ...

So, when will I be free of harassment? When I state that, yes this is bad, but this is also expected, it is what we signed up for, we should expect this...

Will you accept this as an honest opinion and appraisal? Or are you going to call me a cheerleader?

I will tell you honestly, debates like this that degenerate into name calling make me ashamed to be associated with this group.

I agree with you that there is a problem, in almost all cases I acknowledge the issues that are raised and that yes, they do exist. Where I differ is in my conclusions.

* I do NOT conclude that the developers are incompetent.
* I do not invent cute names to call you by.
* I do not tell you that you don't know what you are talking about.
* I do not slander people I have never met.
* I do not impugn their motives.

I do think we are better off now than we have ever been historically, LHC@Home is officially off the Beta roster. We now have 5 projects fully into production mode with nearly twwice as many in open testing modes ... and we are doing work ...

Anyway, we should all think about our conduct ... walk in the other person's moccasins ... if, and I speak to ALL who posted here ... if that post you made, was a reply to YOU, what would you feel ... I know that as I write this I am shamed by the very thought that something like this needs to be said. My rule is that if I have nothing good to say, then I should say just that ... nothing ... and I broke that rule just now and I am shamed by it ... maybe, just maybe, some of you likewise should be shamed ...
ID: 137734 · Report as offensive
Profile Prognatus

Send message
Joined: 6 Jul 99
Posts: 1600
Credit: 391,546
RAC: 0
Norway
Message 137744 - Posted: 17 Jul 2005, 15:12:12 UTC
Last modified: 17 Jul 2005, 15:18:36 UTC

I agree with those who say to close SETI Classic ASAP, where there is a reasonable timeframe for the transition and that it's announced properly to all.

But Ageless said something that made me think... maybe it's a good idea, you decide: why not zip up WU's - both uploads and downloads? Of course, the server and client should be responsible for automaticly packing and unpacking the ZIPs.

The main problem is connecting to the UL/DL server, not bandwith - although that is also stressed - but when you have one file to deal with, instead of 15 or 200 or whatever, it should be easier to clear the queue too. Besides, a zipped file is usually also smaller than its individual members combined size. Plus, it would be fewer connection requests when users had only one file to upload.

Maybe another thing to consider also: why not give bandwidth depending on the client's needs? The definition on "need" can be a lot and I don't mean to discriminate or favorize anyone, but one have to realize that some are in greater need than others.

People with big farms, maybe spread wide geographically and 400 uploads or more, can't easily just press the "Retry Now" button on every WU - as those who have one machine and a handful of WU's. (Myself, I have 3 machines and with lots of patience I have managed to UL/DL all units so far.) Also, people who are in risk of not meeting deadline(s) because of server problems, should maybe get a temporary higher bandwith priority, just to make it easier to reach the deadline(s). The client and server could arrange such things between them on the fly - depending on that the client gets connection once in a while.

Please don't jump down my throat for suggesting solutions. We need to discuss these things I beleive, and I want your opinions, but please consentrate on the issues and not the person(s). This thread is heated enough as it is.. And Yes, I know about the arguments about making your cache size to fit your needs and that it's your own fault if you miss deadlines, and that you can always attach to another project, and all that... but we also have to realize that the server has problems here and you cannot blame users for that.
ID: 137744 · 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 137752 - Posted: 17 Jul 2005, 15:31:28 UTC
Last modified: 17 Jul 2005, 15:42:45 UTC

The first question is what is the bottle neck, is the server I/O bound or CPU bound? If I/O bound then you need a faster/wider pipe or compression that will reduce the consumption. If it is Compute bound, ziped files will make that worse.

Also, compression of data works best when there is repitition of data. The actual result data files for SETI@Home are not that big. I don't recall the actual size, but, they are just not that big. So, unlike CPDN where the results in compressed form are 300M, I wonder what the savings might be.

One more note, if the files have certain characteristics "compressing" the file can make it grow ... which is why you can see the message "adding..." on compression as it has been found that an attempt to compress the data is making an larger output file than it was originally ... hence no compression ...

{edit}
The output, Result Data Files I had on my system (I had 4 I tried) are about 13 to 20K in size. Compreession took those down to about 3K (the files are structured XML and so ARE good candidates for compression). Compression ratios were between 3 and 7 to one. So, the average size of the output was down to about 3-4K from 13 to 21K ...

But, we still do not know what the actual cause of the problem IS, right now. A dropped connection is a dropped connection and a compressed file will be just as out of luck as a non-compressed file... But, it is a thought with merit if one of the primary causes of the drops is a server that is I/O Bound.

My memory says that Matt had said that the server was running at max CPU though ...
ID: 137752 · Report as offensive
Profile Tigher
Volunteer tester

Send message
Joined: 18 Mar 04
Posts: 1547
Credit: 760,577
RAC: 0
United Kingdom
Message 137753 - Posted: 17 Jul 2005, 15:31:33 UTC

Paul
Thank you for your reply.

I am saddened to think that you feel in any way the need to respond to this. You are, quite candidly, the last person on these forums at whom I would ever frown.

I both respect and admire your almost single handed efforts to improve this project with the limitations you have often referred to (again here) and the constraints you have when it comes to the development team and their willingness or otherwise to walk a wise path.

I stand by everything I said and I thank you for your comment. You make your points and you back them up with fact or considered opinion which is always qualified as such when needed. I have no problem with this at all. Others do not!


To your responses sir:

In respect of the model my maths is not good enough. Straight and simple and already said. My suggestion was that a UCB undergrad be asigned. Ideal work for them. So I would if I could but I cannot. I can contribute to ideas around it of course and have made many such suggestions here already in other threads.


In respect of those that want only seti. Yes you are right but equally I think you may be wrong. They have no choice but to move from classic to boinc. They believe this is "being done to them" not for them. They did not want anything else and now they are told they have to or they will have spare CPU cycles. Wear their hat for a minute and take their view. So.....it falls to the project to make things reliable so they have no more cause to complain than they might have with classic and it falls to the project to educate them about dc and how they might provide benefits across many science areas. That being said....and if agreed.....then the project must react.


High Priests

What I am frustrated with are people who just snipe and critcise without ever offering a decent idea themselves. Very quick to call the police but will never go to court and testify. They are destructive in their MO and add no value. I personally feel I do make points worthy of discussion and debate rather than straight rubbishing. Sometimes the points are not well thought through I admit but I would like to be put down nicely and construcively rather than hammered! If this forum cannot support such decent debate then it does not deserve to carry on.

I have not called you a cheer leader and I ask that you let me know if you think I have. I am so sad you put yourself in this class and hopefully this is a two sides of the pond mis interpretation of cheer leader. The Titanic was unsinkable....the company said so...the architect said so.....so did the cheer leaders who were the last seen on deck rearranging deck chairs. A cheer leader is never neutral....never......and is there to get support for one team or one argument against another. So.....that's this side of the pond. What's your side say?


I will go no further until you let me know as the second part of your response is based purely on how you have interpretted a word pair and then applied it, I think unnecessarily, to yourself.



Regards






ID: 137753 · Report as offensive
Profile Tigher
Volunteer tester

Send message
Joined: 18 Mar 04
Posts: 1547
Credit: 760,577
RAC: 0
United Kingdom
Message 137758 - Posted: 17 Jul 2005, 15:39:25 UTC - in response to Message 137744.  

I agree with those who say to close SETI Classic ASAP, where there is a reasonable timeframe for the transition and that it's announced properly to all.

But Ageless said something that made me think... maybe it's a good idea, you decide: why not zip up WU's - both uploads and downloads? Of course, the server and client should be responsible for automaticly packing and unpacking the ZIPs.

The main problem is connecting to the UL/DL server, not bandwith - although that is also stressed - but when you have one file to deal with, instead of 15 or 200 or whatever, it should be easier to clear the queue too. Besides, a zipped file is usually also smaller than its individual members combined size. Plus, it would be fewer connection requests when users had only one file to upload.

Maybe another thing to consider also: why not give bandwidth depending on the client's needs? The definition on "need" can be a lot and I don't mean to discriminate or favorize anyone, but one have to realize that some are in greater need than others.

People with big farms, maybe spread wide geographically and 400 uploads or more, can't easily just press the "Retry Now" button on every WU - as those who have one machine and a handful of WU's. (Myself, I have 3 machines and with lots of patience I have managed to UL/DL all units so far.) Also, people who are in risk of not meeting deadline(s) because of server problems, should maybe get a temporary higher bandwith priority, just to make it easier to reach the deadline(s). The client and server could arrange such things between them on the fly - depending on that the client gets connection once in a while.

Please don't jump down my throat for suggesting solutions. We need to discuss these things I beleive, and I want your opinions, but please consentrate on the issues and not the person(s). This thread is heated enough as it is.. And Yes, I know about the arguments about making your cache size to fit your needs and that it's your own fault if you miss deadlines, and that you can always attach to another project, and all that... but we also have to realize that the server has problems here and you cannot blame users for that.


HI
I agree that Wus should be compressed if there is a saving in size. Its trivial cpu time to compress and decompress but a significant chunk of bandwidth across all users. I did see one of the delevopers warm to the idea a few weeks back. Perhaps they have already picked it up but this is a useful reminder. Thanks


ID: 137758 · Report as offensive
Profile Prognatus

Send message
Joined: 6 Jul 99
Posts: 1600
Credit: 391,546
RAC: 0
Norway
Message 137771 - Posted: 17 Jul 2005, 16:17:47 UTC
Last modified: 17 Jul 2005, 16:23:16 UTC

Tigher; I didn't know that the developers was already thinking about this, but that only shows that ideas are seldom unique. :) It's great if they find any improvements in this area.

Paul; you're right about that we don't know yet exactly what the problem about dropped packets is. And I have also heard several times (I don't remember from where) who said the UL/DL server is CPU bound. So we'll see which solutions (if any) are feasable in the end. I have absolute confidence that the developers are on top of the situation. In the mean time, we mortals can fabulate around the topic some more. ;)

Here's another idea (probably not unique either): you know those phone queues, where an answering machine tells you what your place in the queue is? Maybe the UL/DL server could do something similar... The client could display its place number in the queue, so that the user may relax a bit - knowing his/hers turn is coming and (approximately) when.

Then, when the connection is made eventually, and it's your turn to talk to the server, just maybe the client/server chat should not end with that the server hangs up abruptedly and go the next client before the packet has been re-tried a couple of times.

The anologue to this situation is when you wait a long time in a phone queue, and when you finally reach a person then he/she just hangs up because he/she can't hear you properly within 1,5 seconds or so... and you have to start all over and call back... I'm sure most of us know how frustrating that can be! :)

ID: 137771 · Report as offensive
Profile Tigher
Volunteer tester

Send message
Joined: 18 Mar 04
Posts: 1547
Credit: 760,577
RAC: 0
United Kingdom
Message 137779 - Posted: 17 Jul 2005, 16:34:23 UTC - in response to Message 137771.  

Tigher; I didn't know that the developers was already thinking about this, but that only shows that ideas are seldom unique. :) It's great if they find any improvements in this area.

Paul; you're right about that we don't know yet exactly what the problem about dropped packets is. And I have also heard several times (I don't remember from where) who said the UL/DL server is CPU bound. So we'll see which solutions (if any) are feasable in the end. I have absolute confidence that the developers are on top of the situation. In the mean time, we mortals can fabulate around the topic some more. ;)

Here's another idea (probably not unique either): you know those phone queues, where an answering machine tells you what your place in the queue is? Maybe the UL/DL server could do something similar... The client could display its place number in the queue, so that the user may relax a bit - knowing his/hers turn is coming and (approximately) when.

Then, when the connection is made eventually, and it's your turn to talk to the server, just maybe the client/server chat should not end with that the server hangs up abruptedly and go the next client before the packet has been re-tried a couple of times.

The anologue to this situation is when you wait a long time in a phone queue, and when you finally reach a person then he/she just hangs up because he/she can't hear you properly within 1,5 seconds or so... and you have to start all over and call back... I'm sure most of us know how frustrating that can be! :)



Well that's an interesting idea but I'm not sure how well it would work. Clients make a connection request to the server. That connection would need to held open for a long time perhaps until it is the turn of that client. Using that connection you could transfer the place in the queue. But there will be a restriction on the number of open connections allowed for sure....TCP/IP commonly does this and for good resource reasons. So at some point before the client is serviced the connection may be dropped. So the server would have to connect to the client to conclude the transaction. I think with most peoples' firewall arrangements this would not work as inbound connections will certainly be closed off. This is especially the case for corporate sites for sure. Which means the client would have to try again and no doubt be at the back of any queue. Tell me if you think my analysis is wrong?


ID: 137779 · 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 137785 - Posted: 17 Jul 2005, 16:45:24 UTC - in response to Message 137753.  

Paul
Thank you for your reply.
you are welcome ... :)


I am saddened to think that you feel in any way the need to respond to this. You are, quite candidly, the last person on these forums at whom I would ever frown.

I knew I did not separate the parts well enough and I beg pardon. No, you have not, others have. So, I was responding in part directly to your post and ALSO to the general tone. So, my fault for not being precise enough or clear enough. I, like you say in the post I am now quoting from, also like the debate and the exploration of the ideas. Case in point is the one on data compression which I did a little thought expiriment and then 4 small test cases ... and the basic conclusion is still *MAYBE*... All we can do now is to maybe recast this as a note to the developers mailing list to make the suggestion.

The capability to zip/unzip is in BOINC, I just don't know what the issues are or now hard it would be to make the change. But further thought does seem to make this an idea that has much more merit than I first thought ... especially as the file does NOT have to be decompressed on the Data Server fi I understand things correctly ... so, Maybe I will make the mailing after I wrap this up.

I both respect and admire your almost single handed efforts to improve this project with the limitations you have often referred to (again here) and the constraints you have when it comes to the development team and their willingness or otherwise to walk a wise path.

I stand by everything I said and I thank you for your comment. You make your points and you back them up with fact or considered opinion which is always qualified as such when needed. I have no problem with this at all. Others do not!

Thank you.

In respect of the model my maths is not good enough. Straight and simple and already said. My suggestion was that a UCB undergrad be asigned. Ideal work for them. So I would if I could but I cannot. I can contribute to ideas around it of course and have made many such suggestions here already in other threads.

Well, you are ahead of me. I would not know where to start (and would not spend the time to research a starting point I am too far behind on my corner), but, if you have? Then, lets start there ... even if we cannot finish ... maybe we can start? They may not have a grad student handy ...

In respect of those that want only seti. Yes you are right but equally I think you may be wrong. They have no choice but to move from classic to boinc. They believe this is "being done to them" not for them. They did not want anything else and now they are told they have to or they will have spare CPU cycles.

I think this is where you and I part is only on the conclusion. Yes, it *IS* being done *TO* them. But, the choice is up to them. Volunteer or not volunteer. UCB does not *OWE* anything to the prior participants beyond a thank you. IF, a participant does not like BOINC. Fine. Don't volunteer. Do something else. SETI@Home is moving on. They can, or they can say no. Again, this is the point where we differ. I am a volunteer, I know it and love it. I *KNOW* the project people appreciate us. But, do we really need them to be telling us that they appreciate us, or, doing their part of the project?

My considered choice is that I would rather they spent more time on the project. Not holding my hand. I write documentation ... Tony, bless his heart, spends his time patrolling the forums and directing traffic. Sends me notes or directs me to a post that he thinks could use my help ... others, to numerous to mention do other things ... but the point is ... I volunteer my time, computers, efforts and dreams to the project. But, they owe me NOTHING. Not the time of day, not a message, not even a thank you ... but burried in that part of your post is the implication that UCB and the SETI@Home team *OWES* us something ... they *DON'T* ... sorry, I am getting worked up and shouldn't ...

I have not called you a cheer leader and I ask that you let me know if you think I have. I am so sad you put yourself in this class and hopefully this is a two sides of the pond mis interpretation of cheer leader. The Titanic was unsinkable....the company said so...the architect said so.....so did the cheer leaders who were the last seen on deck rearranging deck chairs. A cheer leader is never neutral....never......and is there to get support for one team or one argument against another. So.....that's this side of the pond. What's your side say?

You did not. But others have. Again, I was not clear, and hungry so, I posted with a simple edit rather than clarity. So, breakfast was good and I apologize for the inadvertant implication that you had.

I will go no further until you let me know as the second part of your response is based purely on how you have interpretted a word pair and then applied it, I think unnecessarily, to yourself.

Well, it has been applied to me and in one sense it is 100% correct because I am passionate about BOINC, I think it is one of the most significant advances in Distributed Computing. I can now participate in 5-10-20-100 projects easily without having to limit myself to one DC project to one computer. And running the risk of running out of work ... so, I accept the term on that basis. I am just disappointed that so many cannot address the issues directly but so quickly degenerate into name calling as a last resort.

And, so, I will likely not continue here ... I see little likelyhood that this is going to turn into a thread with merit ...
ID: 137785 · Report as offensive
Profile Prognatus

Send message
Joined: 6 Jul 99
Posts: 1600
Credit: 391,546
RAC: 0
Norway
Message 137787 - Posted: 17 Jul 2005, 16:51:47 UTC - in response to Message 137779.  

Tell me if you think my analysis is wrong?

It sounds right. I agree with you about the technical implications involved here. Although I still think the model, as a fair principle, is valid enough to think about further. Maybe it can be implemented somehow... You never know. :) Then again, there may be additional concerns that tip over the whole thing.

ID: 137787 · 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 137788 - Posted: 17 Jul 2005, 16:52:00 UTC

Just when I thought I could get away.

The basic idea used in this is collision detection and back off like what is used in Ethernet. The trouble with this is that saturation starts to take place when the network starts to get above 60% of capacity... (others say it is efficient into the 90's but I think they are dreaming).

Token ring, used the sort of mechanism Prognatus suggests. If you don't have the token, you can't talk. This scheme can go almost to 100% loading with little loss of capacity. Response times are also deterministic ... but, then you have to pass the token around.

The undirected nature of BOINC is still probably a better fit with the ethernet model much as it is causing us problems RIGHT NOW ...

Worse, to implement it would require a change in both the clinets and the servers and this would have to be done across all projects at the same time ... :(

I don't see how to make this work ... :(
ID: 137788 · Report as offensive
Pascal, K G
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 2343
Credit: 150,491
RAC: 0
United States
Message 137792 - Posted: 17 Jul 2005, 17:00:30 UTC - in response to Message 137550.  

?

Who would've thought that my little comment 3 days ago would have sparked a 58 page (on my printer) conversation/debate/arguement/discussion/insertyourwordhere about the situation.[/quote]
But interesting reading, isn't it? It runs the gamut . . .[/quote]
On his printer? Read from the monitor and save some trees![/quote]

This reminds me of years ago when I was a electrician. One of my least favorite jobs was pulling wire, especially the big wire pulls. 3 or 4 men would set it up and then need 5 to 8 more guys to help. Without fail one or more of the *helpers* would try to take over and reorganize the setup. Once when this happened I told the wannabe alpha male that we didn't need any help replanning, that we had it covered. When he wouldn't give up telling us what to do with our job I asked him what his foreman had told him to do. He looked surprised and said "To come over and help us with the pull." I then said that we wanted him to go over by a paticular reel of wire and make sure it didn't tangle up as it was coming off. And if he couldn't do that to go back to his foreman and tell him that he couldn't help us. That shut him up and a couple of hours later the wire was in the conduit.

I was asked to help, not take over the job. I've been doing that for a while, and there have been some problems, but I think an amazing amount of work has been done and I am glad I helped. I will continue to help as long as it is wanted.

SETIATHOME was only supposed to last a couple of years but it has evolved into BOINC and is a substanstial achievement in advancing scientific knowledge.

Congrats to SETI,[/quote]


Bump for the nay sayers and wantabe Alpha males..........You know who you are rah rah sis boom bah BOINC BOINC BOINC.........;O)
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.
ID: 137792 · Report as offensive
Profile Tigher
Volunteer tester

Send message
Joined: 18 Mar 04
Posts: 1547
Credit: 760,577
RAC: 0
United Kingdom
Message 137793 - Posted: 17 Jul 2005, 17:05:37 UTC - in response to Message 137785.  

Paul
Thank you for your reply.
you are welcome ... :)


I am saddened to think that you feel in any way the need to respond to this. You are, quite candidly, the last person on these forums at whom I would ever frown.

I knew I did not separate the parts well enough and I beg pardon. No, you have not, others have. So, I was responding in part directly to your post and ALSO to the general tone. So, my fault for not being precise enough or clear enough. I, like you say in the post I am now quoting from, also like the debate and the exploration of the ideas. Case in point is the one on data compression which I did a little thought expiriment and then 4 small test cases ... and the basic conclusion is still *MAYBE*... All we can do now is to maybe recast this as a note to the developers mailing list to make the suggestion.

The capability to zip/unzip is in BOINC, I just don't know what the issues are or now hard it would be to make the change. But further thought does seem to make this an idea that has much more merit than I first thought ... especially as the file does NOT have to be decompressed on the Data Server fi I understand things correctly ... so, Maybe I will make the mailing after I wrap this up.

I both respect and admire your almost single handed efforts to improve this project with the limitations you have often referred to (again here) and the constraints you have when it comes to the development team and their willingness or otherwise to walk a wise path.

I stand by everything I said and I thank you for your comment. You make your points and you back them up with fact or considered opinion which is always qualified as such when needed. I have no problem with this at all. Others do not!

Thank you.

In respect of the model my maths is not good enough. Straight and simple and already said. My suggestion was that a UCB undergrad be asigned. Ideal work for them. So I would if I could but I cannot. I can contribute to ideas around it of course and have made many such suggestions here already in other threads.

Well, you are ahead of me. I would not know where to start (and would not spend the time to research a starting point I am too far behind on my corner), but, if you have? Then, lets start there ... even if we cannot finish ... maybe we can start? They may not have a grad student handy ...

In respect of those that want only seti. Yes you are right but equally I think you may be wrong. They have no choice but to move from classic to boinc. They believe this is "being done to them" not for them. They did not want anything else and now they are told they have to or they will have spare CPU cycles.

I think this is where you and I part is only on the conclusion. Yes, it *IS* being done *TO* them. But, the choice is up to them. Volunteer or not volunteer. UCB does not *OWE* anything to the prior participants beyond a thank you. IF, a participant does not like BOINC. Fine. Don't volunteer. Do something else. SETI@Home is moving on. They can, or they can say no. Again, this is the point where we differ. I am a volunteer, I know it and love it. I *KNOW* the project people appreciate us. But, do we really need them to be telling us that they appreciate us, or, doing their part of the project?

My considered choice is that I would rather they spent more time on the project. Not holding my hand. I write documentation ... Tony, bless his heart, spends his time patrolling the forums and directing traffic. Sends me notes or directs me to a post that he thinks could use my help ... others, to numerous to mention do other things ... but the point is ... I volunteer my time, computers, efforts and dreams to the project. But, they owe me NOTHING. Not the time of day, not a message, not even a thank you ... but burried in that part of your post is the implication that UCB and the SETI@Home team *OWES* us something ... they *DON'T* ... sorry, I am getting worked up and shouldn't ...

I have not called you a cheer leader and I ask that you let me know if you think I have. I am so sad you put yourself in this class and hopefully this is a two sides of the pond mis interpretation of cheer leader. The Titanic was unsinkable....the company said so...the architect said so.....so did the cheer leaders who were the last seen on deck rearranging deck chairs. A cheer leader is never neutral....never......and is there to get support for one team or one argument against another. So.....that's this side of the pond. What's your side say?

You did not. But others have. Again, I was not clear, and hungry so, I posted with a simple edit rather than clarity. So, breakfast was good and I apologize for the inadvertant implication that you had.

I will go no further until you let me know as the second part of your response is based purely on how you have interpretted a word pair and then applied it, I think unnecessarily, to yourself.

Well, it has been applied to me and in one sense it is 100% correct because I am passionate about BOINC, I think it is one of the most significant advances in Distributed Computing. I can now participate in 5-10-20-100 projects easily without having to limit myself to one DC project to one computer. And running the risk of running out of work ... so, I accept the term on that basis. I am just disappointed that so many cannot address the issues directly but so quickly degenerate into name calling as a last resort.

And, so, I will likely not continue here ... I see little likelyhood that this is going to turn into a thread with merit ...



Well Paul again you argue your case well , you support it. For my part I think we are very much in agreement; with the exception of classic perhaps. I will think about the modelling but time is against me....but I will think about it.

If cheer leader translates into passion then fine.......we have clarified what it means in both our dictionaries. There is nothing wrong with passion and my remarks come out of similar , but with somewhat less dedication, sentiments.

As for this thread having no merit......well may I disagree.

I hope all who read here see that a marker has been set down...a line in the sand.....that will encourage people to speak up with ideas and views and know they will not be slammed. A marker that will encourage people to articulate what they feel is wrong and how that might corrected AND what is right and give due credit. I think it serves a good purpose to get this in the open and serve notice that debate is important and valuable. So......glad the breakfast was good!

ID: 137793 · Report as offensive
Profile Fuzzy Hollynoodles
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 9659
Credit: 251,998
RAC: 0
Message 137799 - Posted: 17 Jul 2005, 17:13:57 UTC - in response to Message 137584.  


/me sends Nurse Fuzzy that way. Way better than Nurse Rita. ;)


Ageless, Nurse Rita is waiting impatiently for you for your appointment!!!


"I'm trying to maintain a shred of dignity in this world." - Me

ID: 137799 · Report as offensive
Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 · Next

Message boards : Number crunching : Uploading


 
©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.