BOINC needs a overhaul

Message boards : Number crunching : BOINC needs a overhaul
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · 4 . . . 5 · Next

AuthorMessage
Luke
Volunteer developer
Avatar

Send message
Joined: 31 Dec 06
Posts: 2546
Credit: 817,560
RAC: 0
New Zealand
Message 974642 - Posted: 28 Feb 2010, 2:52:16 UTC
Last modified: 28 Feb 2010, 4:56:00 UTC

I'm going to get utterly scathed for saying this... BUT, before you do rip me apart, LISTEN to what I have to say.

Perhaps a complete code overhaul or ground up design is something to prioritize for a future BOINC 7 release???

It seems to me and many other users around the boards, that every version and release of BOINC seems to be a step backward and in the wrong direction. More and more bugs seem to be introduced every time, with little being fixed. And quite apparent that most ideas and bug reports aren't getting through, or they are, except the devs don't seem to acknowledge them. Comparing it to a commercial business, isn't it good customer service to reply or answer your calls and emails so the customer knows they have been acknowledged?

It also seems that there is a positive, ready, community willing to contribute and discuss many ideas for releases, except it isn't harnessed by the developers:
Reporting about it on these forums does not make any sense as the developers HARDLY read these forums."
and in my humble view that is terribly inefficient.

And yes, some of you will argue that it is better for them to spend time developing and writing the actual code than listening to the communities ideas and issues that arise with each and every release. But you all know what can go wrong if the dev team shuts out a community buzzing with great ideas.

If the view that BOINC is bloated and buggy spreads beyond a select few, as a whole that could turn potential users away from the idea of (which, if correctly implemented, is a great humanitarian and scientific tool) distributed computing. Current users may be put off updating to new releases, or others may simply refuse to (like me, I'm not upgrading until I see that the majority of the bugs are fixed). Some, dare I say it may even leave.

My message to the BOINC dev team is to (politely) halt producing 'stop-gap' releases (6.10.32, 6.10.34, 6.10.35, 6.10.xx) strip the program right down to it's core code, and work upwards from there. It may take a lot of time and effort, but it'll be worth it in the end.

- Luke.
- Luke.
ID: 974642 · Report as offensive
Profile RFP
Avatar

Send message
Joined: 21 Jan 10
Posts: 44
Credit: 29,197
RAC: 0
United States
Message 974651 - Posted: 28 Feb 2010, 3:17:29 UTC

I am in total aggrement with you, but it seems to be following the standard development model ( started my MS)

1. Do a little up-front designe
2. No formal documentation
3. No pere reviews
4. Do testing ( find the 90% of the bugs self evedent)
5. Let users do real time testing to find the remainder ( 10% ) of the bugs.
6. Patch and loop back to step 5.

It seems the SEI (softwarew Eng Inst) wasted a lot of time and money trying to establish a good development mentodology.
ID: 974651 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 974670 - Posted: 28 Feb 2010, 4:29:29 UTC

Go for it Luke! I agree totally.
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 974670 · Report as offensive
Luke
Volunteer developer
Avatar

Send message
Joined: 31 Dec 06
Posts: 2546
Credit: 817,560
RAC: 0
New Zealand
Message 974678 - Posted: 28 Feb 2010, 4:54:34 UTC

Thanks RFP & hiamps!!!
- Luke.
ID: 974678 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 19012
Credit: 40,757,560
RAC: 67
United Kingdom
Message 974679 - Posted: 28 Feb 2010, 4:58:38 UTC
Last modified: 28 Feb 2010, 5:12:20 UTC

No disagreement here.

It needs to be redesigned to reflect current hardware (multi-cpu's and gpu's) and that projects can have more that one application.

edit] It also needs mechanism to extend deadlines after outage. And ways of giving piority to downloads, to feed hungry processors, also after outage. The rule of stopping d/loads when uploads is 2*cpu's is a disaster.
ID: 974679 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 974690 - Posted: 28 Feb 2010, 6:22:51 UTC - in response to Message 974679.  
Last modified: 28 Feb 2010, 6:23:47 UTC

No disagreement here.

It needs to be redesigned to reflect current hardware (multi-cpu's and gpu's) and that projects can have more that one application.

edit] It also needs mechanism to extend deadlines after outage. And ways of giving piority to downloads, to feed hungry processors, also after outage. The rule of stopping d/loads when uploads is 2*cpu's is a disaster.


Disaster???
The project wide backoff is a disaster.

Server saving? Maybe.
But getting things through to the servers in tough times is just stupidly impossible, when the rig stops trying after a couple of attempts and backs off for hours or days. I had one back off over 23 hours at one point.

The most recent outage was a good example. All of my 'normal' rigs had managed to phone home within a short time after things came back online.
The Cuda rigs, which require the newer Boinc mangler, took days.

Yes, it eventually sorted.

More distressing to me than the rigs, I suppose.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 974690 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13720
Credit: 208,696,464
RAC: 304
Australia
Message 974692 - Posted: 28 Feb 2010, 6:39:28 UTC - in response to Message 974690.  

Disaster???
The project wide backoff is a disaster.

Yep.
Much better to let the recovery take a long as possible, rather than make it as quick as possible.
Grant
Darwin NT
ID: 974692 · Report as offensive
Luke
Volunteer developer
Avatar

Send message
Joined: 31 Dec 06
Posts: 2546
Credit: 817,560
RAC: 0
New Zealand
Message 974718 - Posted: 28 Feb 2010, 9:05:14 UTC

Looks like we have 8 who want change (if you count 2 others I've messaged off-boards). Any more???
- Luke.
ID: 974718 · Report as offensive
Profile Spectrum
Avatar

Send message
Joined: 14 Jun 99
Posts: 468
Credit: 53,129,336
RAC: 0
Australia
Message 974725 - Posted: 28 Feb 2010, 9:34:41 UTC
Last modified: 28 Feb 2010, 9:44:49 UTC

Hi all.

Firstly I have to agree with Luke's opening post something needs to happen to make the system work with greater efficiency, every time the system goes down the forums are flooded with complaints and questions, suggestions, conspiracy theories and raves by button abusers and attention whores, I am patient and understanding and may have a naive overview of the of the workings of the project and the complexity of the programming, hardware requirements, infrastructure and science involved in getting and keeping a project such as seti@home running and I may have posted to a thread asking a question or two or said hello to people who I have met in the forums (that I would never know existed without seti@home) who have helped me, given advice, offered a place in a team and made me feel really welcome, but I cannot recall ever getting so pissed off that I started a thread flaming the staff simply because I know that I don't know enough to post a reasonable argument given the complexity of this global project.

Luke I don't think anyone would or should flame you for your post and RFP you have submitted a succinct outline of a way not to move forward, hiamps I say this with joviality no disrespect intended you are a button abuser and attention whore I enjoy your posts please don't ever change:), Mr Sattler meow meoow mew meooww purrr purrrow keep those kitties crunching, loved the auction and it is a brilliant idea to raise funds for the project.

OK enough of my ranting I have just got home after 13 hours of airports and flight changes and have to do it again when I go back to work next week so I am tired and a little tipsy as the duty free is a bonus when coming home.

Most of us have spent money upgrading our systems to better crunch for the cause not to mention the power bills that we pay (are we crazy or just passionate?) I just hope someone in charge reads this thread and acts on it as I can see some people not being as tolerant or understanding leaving the project.

Only my two cents worth in what I believe is probably the largest post I have ever undertaken and no disrespect only respect and encouragement to all.

Lou AKA Spectrum
ID: 974725 · Report as offensive
Profile Spectrum
Avatar

Send message
Joined: 14 Jun 99
Posts: 468
Credit: 53,129,336
RAC: 0
Australia
Message 974726 - Posted: 28 Feb 2010, 9:36:02 UTC
Last modified: 28 Feb 2010, 9:47:51 UTC

Hi all.
Luke you have my vote.
ID: 974726 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 974742 - Posted: 28 Feb 2010, 10:46:52 UTC - in response to Message 974730.  

And you've got my vote too, I simply don't upgrade, change, cause,
I've seem to have lost track of the endless stream of BOINC versions, since BOINC 5.10.45.

And the implementation of CUDA, doesn't work as smoothly, as it should, IMHO.

ID: 974742 · Report as offensive
Profile gizbar
Avatar

Send message
Joined: 7 Jan 01
Posts: 586
Credit: 21,087,774
RAC: 0
United Kingdom
Message 974748 - Posted: 28 Feb 2010, 12:04:17 UTC

Stick my vote in as well. Boinc seems to be being upgraded for Boinc's sake now, and not for the benefit of the projects it's managing. It seems to me that the features added are now undermining the core operation and stability. I have not upgraded since 6.10.25, and I only upgraded to that because of a discussion over at Lunatics.

I'm glad that people are willing and able to test it so throughly. I don't have the patience to sit there and try the same thing over and over again to replicate a fault. I just want to install it, and let it run. It has to be a pretty big and bleeding obvious bug for me to catch it, so it's usually been dealt with by the time I get to it.

regards, Gizbar.



A proud GPU User Server Donor!
ID: 974748 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 974751 - Posted: 28 Feb 2010, 12:41:55 UTC - in response to Message 974692.  

Quick as possible for WHO? The backoff made mine go into 24 hr wait even after things were going again. Was not good for the project either if machines can't start working again. Plus boinc can't handle large loads which are growing with every Nvidia card added.
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 974751 · Report as offensive
Profile 52 Aces
Avatar

Send message
Joined: 7 Jan 02
Posts: 497
Credit: 14,261,068
RAC: 67
United States
Message 974755 - Posted: 28 Feb 2010, 13:07:19 UTC

Perhaps a complete code overhaul or ground up design

What do you want Boinc 7 to be based on? Cloud computing? Then you'd know what hurt really feels like ;)

You folks have no idea how good you've got it. I run public builds of Boinc (ok, it's dogfood) as soon as possible. When things tank on me, it's 90% the S@H project, 10% my own fault, and 0% boinc.

The rule of stopping d/loads when uploads is 2*cpu's is a disaster.

I pinpointed and posted here the source to revise that. Build it to suit. You don't like the backoff process? Change (or disable) it. Or script it to your liking with boinc.exe. It's not that hard, especially if you're remotely motivated and upset by something. We all pride ourselves in being among forward creative & scientific thinkers for the planet. So then, invest in what irks you, either financially (in Boinc) or sweat equity (recompile and run). Hell, wasn't that the idea behind Lunatics in the first place, folks were upset by performance, and motivated to do something about it. Sure, they could have instead whined in the wrong forum about it for years, but they learned to crawl, walk, run... and made a huge difference in the process for the entire project.

wrong if the dev team shuts out a community buzzing with great ideas

Ya know, 70 out of 100 ideas around here really aren't that great because the evangelist don't understand all the issues (fails to scale, etc). 10 might be great, but they just don't itch the passion of the devs to work on it (it's just reality that if a dev is unexcited about something, it won't get done, or at least not well). The other 20 ideas probably get done and are either unacknowledged by the community, or quickly forgotten within days (that's greatness and gratitude for ya).

But you've already solved the problem in a way that suits the people here. If Early Adoption Livin' In Internet Time doesn't float your boat, then pick a wagon towards the end of the wagon train. Something which best maximizes happiness and stay with it until something supercedes it. Slap it with the 'Luke Seal Of Approval' ;-)

Plus boinc can't handle large loads

That's got far more to do with hardware (of the project) than software provided under the boinc umbrella. The legos are in place to scale out, but the money isn't there for S@H to make it happen -- thought that was well understood.

Cheers :-)
ID: 974755 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 974758 - Posted: 28 Feb 2010, 13:19:54 UTC - in response to Message 974755.  
Last modified: 28 Feb 2010, 13:27:30 UTC

(off topic)

Plus boinc can't handle large loads

That's got far more to do with hardware (of the project) than software provided under the boinc umbrella. The legos are in place to scale out, but the money isn't there for S@H to make it happen -- thought that was well understood.

And most of the large loads are self-inflicted by people who deliberately push the envelope far beyond what BOINC was designed for - and delay the return of scientific results in the process.

I'm still trying to formulate a more substantive reply on the main subject - maybe later, or I may just send a PM to Luke. I'll see how the discussion goes.
ID: 974758 · Report as offensive
Profile Bill Walker
Avatar

Send message
Joined: 4 Sep 99
Posts: 3868
Credit: 2,697,267
RAC: 0
Canada
Message 974764 - Posted: 28 Feb 2010, 14:04:53 UTC

Seems to me that this petition would be valid if you can get the majority of active users (not just posters) to sign up. I count about 12 "in favours" so far, you need another few 100,000.

I share your frustration sometime guys, but you are missing the point completely of BOINC, SETI, etc. It is not about keeping our machines (and egos) fed. It is about doing science: within the budgets, time constraints, and abilities of the scientists we serve. They don't serve us.

If you don't like that, I suggest you start up your own distributed computing system, where the points are more important than the science.

ID: 974764 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 19012
Credit: 40,757,560
RAC: 67
United Kingdom
Message 974771 - Posted: 28 Feb 2010, 14:29:49 UTC - in response to Message 974755.  

The rule of stopping d/loads when uploads is 2*cpu's is a disaster.

I pinpointed and posted here the source to revise that. Build it to suit. You don't like the backoff process? Change (or disable) it. Or script it to your liking with boinc.exe. It's not that hard, especially if you're remotely motivated and upset by something. We all pride ourselves in being among forward creative & scientific thinkers for the planet. So then, invest in what irks you, either financially (in Boinc) or sweat equity (recompile and run). Hell, wasn't that the idea behind Lunatics in the first place, folks were upset by performance, and motivated to do something about it. Sure, they could have instead whined in the wrong forum about it for years, but they learned to crawl, walk, run... and made a huge difference in the process for the entire project.


Fine if one has the time etc. but as one or two have noted I haven't been around, here or Lunatics, much recently.
I spoke a long time ago to JM7 about this, and he said it was put in at the request of another project, and also he did not seem to think it was much of a problem.
Well I for one think having to keep a computer on waiting for uploads to finish before it can get any downloads is STUPID.

The majority of BOINC users only connect to one project and also most do not come anywhere near these or any other BOINC project boards. So posting a solution is good but in all honesty you are not going to get many people to follow your suggestion.

The rule should not be there, there are other ways of limiting traffic density.
ID: 974771 · Report as offensive
Matthew S. McCleary
Avatar

Send message
Joined: 9 Sep 99
Posts: 121
Credit: 2,288,242
RAC: 0
United States
Message 974781 - Posted: 28 Feb 2010, 15:38:15 UTC

It's nice that you want change, but in the end it's going to come down to "no money, no resources" like most everything else around here.
ID: 974781 · Report as offensive
Profile Bill Walker
Avatar

Send message
Joined: 4 Sep 99
Posts: 3868
Credit: 2,697,267
RAC: 0
Canada
Message 974782 - Posted: 28 Feb 2010, 15:38:45 UTC - in response to Message 974770.  

I think you're missing the point. If you can't deliver SCIENCE due to the behaviour of the client, then the client needs an overhaul.

Sten-Arne


Do we know that the science is being delayed by BOINC? Or is the pacing item the ability of the scientests to generate new work and analyze the results delivered so far? From scanning SETI and Milkyway web pages, the limit seems to often be scientest time, not crunching time.

To put it another way, if SETI staff (or anybody else) suddenly got extra money or staff, would the project be better served with doing more science, or doing more crunching?

ID: 974782 · Report as offensive
Matthew S. McCleary
Avatar

Send message
Joined: 9 Sep 99
Posts: 121
Credit: 2,288,242
RAC: 0
United States
Message 974783 - Posted: 28 Feb 2010, 15:47:06 UTC - in response to Message 974770.  


I think you're missing the point. If you can't deliver SCIENCE due to the behaviour of the client, then the client needs an overhaul.

Sten-Arne


That pretty much sums up how I feel about things lately. Credits are nice, but after all the BOINC projects to which I am attached are those I feel do "good science" and are worthy of my computing resources.

Those who always toss out the argument "we are serving SETI@home, not the other way around" are correct, but that does not absolve the admins from the responsibility of keeping us informed about what is going on. Lately they have done a terrible job of that. When we complain about it, a few say "well what would you rather have, people fixing the stuff that's constantly breaking or people telling us about what's going on?" There's no reason we couldn't have both, but we don't.

I want to be clear: I don't care about the credits. I care about my contribution to the project. The project is down so much lately that I am having a hard time contributing.

There are other, equally worthy projects which are down far less frequently. Now, I'm sure you'll point out that no project has as many users as SETI@home, and therefore no project has the constant load on its servers as SETI@home. Well and good. But that changes nothing about the fact that I can't help SETI@home do its science because it's down 25% of the time!

Why do I keep bothering to contribute to SETI@home, I wonder?
ID: 974783 · Report as offensive
1 · 2 · 3 · 4 . . . 5 · Next

Message boards : Number crunching : BOINC needs a overhaul


 
©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.