Message boards :
Number crunching :
Seti needs expert advise on system setup........
Message board moderation
Author | Message |
---|---|
Losi Send message Joined: 29 Jan 04 Posts: 51 Credit: 86,125 RAC: 0 |
Im sorry to say but Seti does need expert advise so they can fix their problems. Dows any one here work in the IT industry that can offer their services? |
Chad@SETI.USA Send message Joined: 17 May 99 Posts: 61 Credit: 546,459 RAC: 0 |
I run a DataWarehouse that Performs 56 million SQL updates and deletes a day and we are running on a Quad Zeon 2Ghz 2MB cache 8MB Ram, Win2003 DataServer. The box still has CPU & Network left over to do reporting... The 4 Drive arrays are at 100% pretty much all day long though! Hard disks are my major bottleneck! This box communicates with 160,000 clients up to 20 times a day sending packets between 130K - 18K. It then takes the data that is communicated from the clients normalizes the data and then deletes/inserts into the live Database real-time, the same DB is also used for Production reporting of the data. I mean we do have some deadlocks but not many! Bionic only has 170,000 users and it is my guess that they are all not active? I would be happy to put in my 2 cents if anyone from BIONIC would care to contact me via my email! It took us almost 3 years to perfect my system so I know the headaches that they are probably fighting! Chad |
Toby Send message Joined: 26 Oct 00 Posts: 1005 Credit: 6,366,949 RAC: 0 |
I'm assuming that "8MB Ram" is supposed to be "8GB"? :) Seti has approximately 150,000 active hosts right now (370,000 total). Depending on their speed and user settings, they can contact the server anywhere between once and probably 30 times per day. Every time a unit comes back, it has to be checked against the other results. If credit is granted, the result, host, user and team credits have to be updated which also involves updating the recent average credit and a time field. Then after 14 days the file and all associated database rows must be purged. There are other details and many intricacies of course. I'm sure there is room for improvement so feel free to take a shot :) BOINC is open source so no need to wait for an email requesting help. The source and a lot of documentation are available at http://boinc.berkeley.edu/. There is a developer mailing list that you can subscribe to in order to discuss changes and issues. Knock yourself out :) A member of The Knights Who Say NI! For rankings, history graphs and more, check out: My BOINC stats site |
BarryAZ Send message Joined: 1 Apr 01 Posts: 2580 Credit: 16,982,517 RAC: 0 |
It isn't clear to me that the various issues that the Berkeley folks are coping with are BOINC specific. Though SETI BOINC *IS* the most active of the various BOINC projects, there are other BOINC projects running relatively large active host counts (Climate and Einstein) which, though only about a quarter of SETI are still quite large, and they seem to run quite solid. I figure I'm doing my bit to reduce the load on SETI BOINC by redirecting a large chunk of my CPU resources over to Climate and Einstein -- sort of sharing the workload (and reducing frustration with the fits and starts over here). |
Chad@SETI.USA Send message Joined: 17 May 99 Posts: 61 Credit: 546,459 RAC: 0 |
I'm assuming that "8MB Ram" is supposed to be "8GB"? :) You are correct on the Ram, I mistyped MB for GB... (Maybe I was thinking about my old 386...) As for my assistance, sorry I do not have the time to wade through thousands of lines of code... But if they are having a specific problem with a process I would be happy to be a brain to bounce ideas off of! Again I know what a chore it is to process Millions of transactions per day and please understand that I am not knocking anyone or any process, I am just offering my assistance if needed! BTW if I read the stats right there are 172,141 Seti Users but only 65.8% have been active in the last week... That means that only 113,269 have completed at least one work unit in the last week??? Again I am new at this Bionic stuff so I could be way off base? Per this page: http://www.boincstats.com/stats/project_graph.php?pr=sah Chad |
ThePhantom86 Send message Joined: 30 Jan 02 Posts: 268 Credit: 1,970,082 RAC: 0 |
Hate to correct ya Chad, but it's BOINC, not BIONIC. |
Saenger Send message Joined: 3 Apr 99 Posts: 2452 Credit: 33,281 RAC: 0 |
BTW if I read the stats right there are 172,141 Seti Users but only 65.8% have been active in the last week... That means that only 113,269 have completed at least one work unit in the last week??? Again I am new at this Bionic stuff so I could be way off base? Just about 12% ATM ;) From BOINCstats as well: Number of Hosts: 373,714 Active Hosts (credits within last week): 34.2% So number of active hosts: 127,810 And as the validators are off-line, the percentage is very low, so I expect this values a bit off-reality ATM. I think Tobys 150,000 are a good estimate, he maintains a stats site as well so he has IMHO enough experience. And adding to this, they have to maintain the old system as well, I expect a better performance once Classic is gone for good. Gruesse vom Saenger For questions about Boinc look in the BOINC-Wiki |
W-K 666 Send message Joined: 18 May 99 Posts: 19048 Credit: 40,757,560 RAC: 67 |
BTW if I read the stats right there are 172,141 Seti Users but only 65.8% have been active in the last week... That means that only 113,269 have completed at least one work unit in the last week??? Again I am new at this Bionic stuff so I could be way off base? But you should have looked lower at the number of hosts active and that is 360,000+. And connections/day going from my families hosts I would put at more like 50/day rather than the 30 suggested earlier. So that makes it more than 10M/day with files from 10k (results) to 350k (units). Want to do the sums again? ;-) Andy P.S. Einstein and climate both take much longer to crunch, therefore less connections * active hosts. P.P.S And Berkeley is still doing Seti Classic and the Beta Trial ( if this becomes the new Boinc/seti then crunch times will increase 10* longer) |
Ned Slider Send message Joined: 12 Oct 01 Posts: 668 Credit: 4,375,315 RAC: 0 |
Im sorry to say but Seti does need expert advise so they can fix their problems. If there is a problem (which actually I don't believe there is), then it's probably more a case of the EXCELLENT job they do with very limited resources. Resourses, both human and hardware, cost money so instead of whinging, why not stick your hand in your pocket and donate $100 - no expertise required ;) Ned *** My Guide to Compiling Optimised BOINC and SETI Clients *** *** Download Optimised BOINC and SETI Clients for Linux Here *** |
Mike Send message Joined: 17 Feb 01 Posts: 34255 Credit: 79,922,639 RAC: 80 |
Im sorry to say but Seti does need expert advise so they can fix their problems. Hi I did it twice in the past couple of month. And i have patience. greetz and respectful from Germany Mike Boincing since March 2003 With each crime and every kindness we birth our future. |
PhonAcq Send message Joined: 14 Apr 01 Posts: 1656 Credit: 30,658,217 RAC: 1 |
This community is doing Losi a disservice by rating him down so badly. He is making an objective observation that should be responded to objective, rather than what seems to have happened. I, for one, would like to see more objective banter on these boards, rather than questions like 'what computer should I buy?' (the latter has no answer but baits everyone who gets emotional about such things) May this Farce be with You |
Ned Slider Send message Joined: 12 Oct 01 Posts: 668 Credit: 4,375,315 RAC: 0 |
This community is doing Losi a disservice by rating him down so badly. He is making an objective observation that should be responded to objective, rather than what seems to have happened. I, for one, would like to see more objective banter on these boards, rather than questions like 'what computer should I buy?' (the latter has no answer but baits everyone who gets emotional about such things) I think the point is that BOINC/Seti IS the largest distributed computing project in the world. These guys ARE the experts - they concieved the idea and implimented it first. If anyone knows how to go about such an undertaking, it's them. I can't really imagine anyone out there with more experience - these guys aren't novices playing with a small home network on the weekends :) But I take your point ;) Ned *** My Guide to Compiling Optimised BOINC and SETI Clients *** *** Download Optimised BOINC and SETI Clients for Linux Here *** |
Prognatus Send message Joined: 6 Jul 99 Posts: 1600 Credit: 391,546 RAC: 0 |
I think Berkeley needs donations, not expert advice. They are doing a great job and have lots of skilled people, but little funds to buy hardware. Please donate some money. :) |
Spectrum Send message Joined: 14 Jun 99 Posts: 468 Credit: 53,129,336 RAC: 0 |
I think and this is my opinion not swayed by the previous posts in this thread that unless you where there at the time and know exactly what has happened or gone wrong you should not try to make a judgment, I do not know the level of expertees the people at the project have (but i am pretty sure they know more about it than me!), but what I do know is that the Idea is what matters, people from all walks of live and a multitude of countries sharing in a common interest with no political or religious ideals inferred, the idea of the search for something that is "out there" apart from our little blue planet shared by hundreds of thousands of "people" is something that brings us all closer together. The people running the project are doing so on a budget with the the donations and assistance of various organisations, if you would like to help please offer you time and expertees, as I am sure that if I was able to do more to support this project I would, but by the same token dont bad mouth the people who are trying to do the best they can with what they have. This post is not meant to offend or upset anyone, it is just my opinion and if you dont like it so be it. |
Thierry Van Driessche Send message Joined: 20 Aug 02 Posts: 3083 Credit: 150,096 RAC: 0 |
I think the point is that BOINC/Seti IS the largest distributed computing project in the world. These guys ARE the experts - they concieved the idea and implimented it first. If anyone knows how to go about such an undertaking, it's them. I can't really imagine anyone out there with more experience - these guys aren't novices playing with a small home network on the weekends :) Agree with this. Look at the other Boinc projects, like LHC@Home and E@H. These projects are working fine, no problems at all. Only that LHC@Home had some periods of WU's coming and going. These projects don't have the problems like S@H/Boinc. Why? They have more budget to run the project. It is clear that without funds you can't get what you want. The problem actually with S@H/Boinc is definitely a fund problem, not a question of incompetency. |
Riley Send message Joined: 30 Jan 00 Posts: 30 Credit: 6,362,824 RAC: 0 |
I think Berkeley needs donations, not expert advice. They are doing a great job and have lots of skilled people, but little funds to buy hardware. Please donate some money. :) I agree about the money. If Seti was a corporation it would have all sorts of redundancy and disaster recovery. Seti obviously does not have the funds at this time.(Hint, hint) Also, Once Seti Classic goes down more resources will be available for Seti BOINC. However, to be fair we would have to compare Seti to its competitors and look at budget versus up-time. Without this comparison there is really no way to objectively say that Seti BOINC is performing well given its budget. The question is whether this type of information is even available. Anyone? I definitely think Losi had the right to question the performance of Seti BOINC. It is always good to benchmark against others in industry to see exactly where you stand. |
Losi Send message Joined: 29 Jan 04 Posts: 51 Credit: 86,125 RAC: 0 |
So who has donated to Seti? It looks like that none of the Seti budget is used to upgrade the equipment that is used, and is completely reliant on donations by major companys. |
Losi Send message Joined: 29 Jan 04 Posts: 51 Credit: 86,125 RAC: 0 |
Negative feedback what is it? Im not on EBAY. Pathetic really. |
Losi Send message Joined: 29 Jan 04 Posts: 51 Credit: 86,125 RAC: 0 |
Im sorry to say but Seti does need expert advise so they can fix their problems. I am prepared to give $1.00 a day everyday of the year, if everymember does the same. That should solve the funding problem. They would receive, lets say $350000.00 US a day. |
BarryAZ Send message Joined: 1 Apr 01 Posts: 2580 Credit: 16,982,517 RAC: 0 |
I think you are on to something regarding comparisons with other BOINC projects, their funding and their performance. Regarding donations to Seti Berkeley -- perhaps it would be useful to get some information from Berkeley -- that is, what *THEY* see as performance contraints at this point in time (software, hardware, reliability, etc.). Else we might be thinking of contributions that have no performance related home. I don't know that I've seen a clear statement of status from the project folks regarding their view of what they see as current issues, what they see as solutions to those issues, and what outsiders can do to further that process. [/quote] I agree about the money. If Seti was a corporation it would have all sorts of redundancy and disaster recovery. Seti obviously does not have the funds at this time.(Hint, hint) Also, Once Seti Classic goes down more resources will be available for Seti BOINC. However, to be fair we would have to compare Seti to its competitors and look at budget versus up-time. Without this comparison there is really no way to objectively say that Seti BOINC is performing well given its budget. The question is whether this type of information is even available. Anyone? I definitely think Losi had the right to question the performance of Seti BOINC. It is always good to benchmark against others in industry to see exactly where you stand.[/quote] |
©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.