 |
|
|
21)
Message boards :
climateprediction.net Science :
How many WATTS does CPN cost??
(Message 18710)
Posted 24 Dec 2005 by  old_user85254
Post: [...ten computers...] How much magnetism in the room! Magnetism? I don\'t have any CRT displays, which are the biggest source of magnetism as far as I know. Two TFT\'s. I connect to four boxes via a KVM and I control the others via Putty/SSH or BOINCview. It is also perfectly possible to drive a BOINC client on a linux box (linux gui or linux command line) from a windows box over a network and using only the standard BOINCmanager program, so if the BOINC-only boxes are set up to load BOINC at bootup you never need to have a monitor, keyboard or mouse attached to them at all. BOINCview gives the advantage of controlling them all from one screen. R~~ |
|
22)
Message boards :
climateprediction.net Science :
How many WATTS does CPN cost??
(Message 18707)
Posted 24 Dec 2005 by old_user85254
Post: Wow, ten computers...impressive :o) quieter than a 2kW fan heater, which is sold as a viable room heater ;-) R~~ |
|
23)
Message boards :
climateprediction.net Science :
How many WATTS does CPN cost??
(Message 18621)
Posted 22 Dec 2005 by old_user85254
Post: [quote]You are funny. There is no way a computer will heat a room! [quote] Exactly. These ten computers draw 550 Watts together, when the monitor is switched off. Except in the very coldest weather that is enough to keep my lounge warm. I position them reasonably low down in the room, below desk height, otherwise they end up only heating the top part of the room! I am thinking of buying a few more, to cover the very coldest weather! 1 x Dell Optiplex desktop, 700MHz deleron 2 x HP Kayak twin 665MHx cpu PIII 3 x Compaq sff 833 MHz 1 x Compaq sff 766 MHz 1 x Compaq sff 600 MHz 2 x Compaq sff 500 MHz All bought second/third hand, very very cheap. Obviously 7GHz in one box would draw less power - you\'d be looking for a twin core HT chip running at 3GHz per core, or a twin cpu twin core box with 2GHz chips & no HT. But i could not afford to buy a box like that, and I\'d still have to find a way to heat my lounge ;-) Together these ten slow boxes provide the same throughput as 7 x 2.8GHz hyperthreaded boxes that keep office hours. In summer most of the slow boxes are turned off as I don\'t need the heat. River~~ |
|
24)
Message boards :
climateprediction.net Science :
How many WATTS does CPN cost??
(Message 18619)
Posted 22 Dec 2005 by old_user85254
Post: First of all. Electricity is not the most efficient way to heat, but if you don\'t have an alternative, than I like the synergy in running computers for CPDN and using the waste heat to heat up the room. Yes. Even if electricity is not the most efficient form of heating, the combination with useful computing can make it a winner, \"Combined Heat and Crunching\" ;-) But, if you buy your electricity from a supplier that only uses wind and water power, in fact electric heat could *also* be the most environmentally efficient, even if not the cheapest. This is what I do. All my BOINCing comes from wind farms, and the cold wind also heats my lounge. River~~ |
|
25)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18594)
Posted 22 Dec 2005 by old_user85254
Post: Gravywavy, That does not help, but is not the only issue. It is clear from your comments that these issues were thought about, but that no access was forthcoming to allow changes to the FAQ, the minimum resources pages, and so on. This is a bureacratic failing to allow those with the knowledge of important changes to communicate them where people are most likely to see them. It is *shocking* that the climateprediction.net site still links to a page that says an 800MHz box can crunch a WU on 3 months, when in fatc not it is around 9 months. There has been time to document these things on the wiki and in the community pages (whether by staff or by volunteers), all it needed was appropriate permissions to allow the same writers to put the pages onto the central website.
Yes, thank you too for all your comments and support. River~~ |
|
26)
Message boards :
Number crunching :
latest trickles info stopped
(Message 18593)
Posted 22 Dec 2005 by old_user85254
Post: This thread was created by me to rpeort the temprary failure of the trickle facility. Since then it recovered, and even more recently was deliberately stopped. Please note that the issue of the deliberate removal of the trickle info is explained and discussed in this official thread and if you want your comments to be noticed, that would be a better place to mention them than here. But thank you for supportig my thread ;-) |
|
27)
Message boards :
Number crunching :
Database Sluggish -- Moved Trickle Info \'Down\'
(Message 18544)
Posted 21 Dec 2005 by old_user85254
Post: In Internet Explorer the user can add and name the different trickle pages for models that have to be checked on trickle update. hi again Kilcock yes, in IE you set the trickle pages as \'favorites\'; in Firefox you bookmark them - same thing really. It still means one bookmark/favorite for each model - with 8 boxes and 11 active model that still leaves 11 bookmarks to check. An even better way in the firefox browser is to open each of these pages in separate tabs, then use the firefox facility to save the entire set of tabs as a single bookmark. This means that one click gets all the pages, and you simply scan from one tab to the next to look at them all. Firefox is free and open source and available for all the platforms running CPDN, more info & downloads from www.firefox.com It is still 12 clicks to see all the info, so I\'d still prefer some way of getting it all onto one page. Also it means that when I click on the one bookmark it sends 11 queries to the database all together - a well designed single page might be less of a load on the database River~~ edits: speeling & typos ;-) |
|
28)
Message boards :
Number crunching :
Database Sluggish -- Moved Trickle Info \'Down\'
(Message 18543)
Posted 21 Dec 2005 by old_user85254
Post: Hi Carl, Thanks for the explanation Obviously that kind of load is unacceptable and you had to do something. It does not make sense to include expensive information on a page which may be loaded for other reasons. However with multiple computers, not all in the same place, it was very useful to have all recent trickles on the same page. Three useful alternatives would be 1) Have a separate \'recent trickles\' page that users could click on if they wanted it. The advantage to the project is that the heavy load is only produced when the user specifically wanted it, no for example when the user only went to the page to see their latest stats or whatever. 2) Introduce a \'last successful trickle\' date column to the page that lists computers (not quite the same as the existing \'last contact\' which always might be for some kind of err, manual update, even machine reboot sometimes) 3) Introduce a \'last successful trickle\' column on the page that lists a user\'s results. The advantage of any of these, from the viewpoint of a user with many machines, is that one page will give a snap overview of whether any of the machines / results needs some attention. Perhaps with option 1 you could extract a derived table once an hour with trickle info and userid in it, with an index on userid. The query would be an hour old but that is fine for a quick-look facility. With option 1, you could also make the script run at a lower priority than the rest of the web pages. This would mean two things, firstly, users asking that question would not slug the rest of the website, and secondly if those pages got overloaded the slowness of their response would deter their use for frivolous queries. The page would need to contain an explanation that \"at busy times this page may be slow\", and \"please do not repeatedly reload - this will only make it worse\". (One way I\'\'ve done this is for the page to call a separate script via a \'nice\' command -- it got round a similar problem we had of one particular page slugging a whole website) Any one of these options, or any other way of doing a quick \"is everyone happy\" check one to three times a day would be a welcome replacement. With 8 active machines I won\'t have the time to check on every machine every day without some way of doing it from a single bookmarked page. And to be clear, this is not a complaint, it\'s a request. You clearly did the right thing in the short term to de-slug the db. It is much better to have most of the old functionality running at a workable speed that have it all running as slow as it has been lately. The performance boost since your change is both obvious and welcome. |
|
29)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18501)
Posted 20 Dec 2005 by old_user85254
Post: While in Sulphur Beta test, we anticipated the necessity of expanded documentation for processing Sulphur Work Units. I started such a page and others added to it and made it better. Crandles honed it and put it in the Wiki. (We don\'t have access to the CPDN FAQ as far as I know.) No, I certainly don\'t fault Carl or Tolu. But I do think you should have access to the FAQ, and I do think the FAQ and the introductory comments about the project should be updated now that sulphur has come along. The poor handling of the documantation issue -- not by yourselves but by whoever decided not to give you access to edit the FAQ -- only strengthens my feeling that it is time to look for another project. Les suggested looking back every so often to see if there are any more \'smaller\' WU - but it is not going to be easy to do that if the intro and the FAQ are still based on slab sized work *regardles* of whther you have any. I\'m also a little disappointed not to see any \"it\'s been great to have you thanks for all your help\" kind of message to those users who honestly feel they csn\'t run the larger WU. It would not have taken much to make the end of the slabs a celebration for the initial CPDN project, for the slabs which paved the way for sulphur. I\'ve decided what I am doing: I have aborted some of my sulphurs, and will leave those currently renaining to run to completion. I will leave all my remaining slabs to run but in some cases with a lower project share than before. I\'ve set \'no more work\' on this project on all my boxes. In the short term most of this resource is going to Rosetta -- when orbit comes online at least half is going there. It\'s great science, slab was and sulphur is. Sulphur needs, in my opinion, more thought about the human aspects of up-sizing the science -- or if the thought has been there it needs more willingness from the project powers to allow it to feed into the website as well as the user forums. Best regards to everyone, to those who are staying and to others who are moving on to other projects. Happy holiday time and Merry Crunching to all. River~~ |
|
30)
Message boards :
Number crunching :
server state means ??
(Message 18292)
Posted 16 Dec 2005 by old_user85254
Post: Is there some kind of time-out that happens between trickles, or what?There is a 2-weeks time-out to the active machine count/status... Looking again at my results, the time out still does not explain this result, sent to me less than 14 days ago on the 6th, trickled twice since, and yet shown as \'over\' I am still curious why this result is over, when other trickling results are not. I wonder what makes the difference? |
|
31)
Message boards :
Number crunching :
server state means ??
(Message 18286)
Posted 16 Dec 2005 by old_user85254
Post: There is a 2-weeks time-out to the active machine count/status. In practice, I have twice had recycled WU that were aborted by previous user (shown as client error and code 197 if I remember rightly) but have personally never seen a recycled WU that had simply timed out. Doesn\'t mean they don\'t exist of course! Thanks for your responses Honza. River~~ |
|
32)
Message boards :
Number crunching :
Unsure whether to keep new sulphur WU?
(Message 18285)
Posted 16 Dec 2005 by old_user85254
Post: This thread is intended to help users who have just received a sulphur WU and just realised that it is a lot longer than their previous slab WU. You may be wondering whether to keep it or not, and this thread is to help you make up your mind on that. This thread may also help people who are currently crunching slabs to decide whether to set nomorework in order to avodi sulphur, or to allowmorework in order to accept sulphur when they get their next WU. There is an already-long thread here discussing what the project could/should offer as well as the new sulphur WU. If you want to comment about the project\'s decision to withdraw slab, please take those comments to that thread. This thread is simply to help you decide whether to abort the new sulphur WU or to allow it to run. First off, if you do decide not to keep the sulphur WU, it is kinder to the project to abort than to detach without an abort. The best procedure is 1. Set nomorework for CPDN 2. Abort the sulphur WU 3. In the work tab, wait for it to change from \"Aborted by user\" to \"computation error\" 4. Update CPDN 5. In the work tab, wait for the WU to disappear 6. Either stay attached but with nomnorework; or detach. I can\'t tell you how to make the decision -- like all donors to all BOINC projects you are volunteers and free to make your own choices. I do suggest that you think about the following points before making up your mind where you personally want to strike the balance. The points numbered A1, A2,... are points in favour of aborting sulphur, those numbered C1, C2, ... in favour of crunching it. A1. The run time will be about three times as long as a slab model A2. This might mean you run over the deadline A3. A1 might be an issue on a loan machine, or where you have permission from another person/organization to run the software on their box A4. A1 also means that the potential amount of lost science (after a machine failure, theft, etc) is up to three times as great A5. You are more likely to go into EDF / NWF modes A6. A5 means that CPDN will hog the box for months on end. For the next few of months your RAC will be skewed towards CPDN, followed by months when it is skewed away from CPDN while the long term debt mechanism hold CPDN back and lets your other projects catch up. A7. A5 also means that CPDN is suddenly less suitable for being a \'background\' project to cover gaps in work from projects like LHC which have frequent fallow times. When the work on LHC comes back, if your box is in NWF it will not look for it. Then when the sulphur finishes, LHC might be in another fallow period so you will download another sulphur A8. On an HT box you might be running two WU together as this gets around 30% more throughput at the expense of around 45% more runtime. On some boxes that are on part-time, you may want to abort one sulphur and fall back to single CPU running to keep the runtime more manageable. On the positive side C1. Yes the run is longer but the credits are in proportion. If you trickle regularly you should see around the same RAC as at present. C2. On this project the deadlines are advisory. You will not lose credit for runing over, and the scientists seem content to get the results late in preference to not at all. C3. The sulphur science is exciting science and an important part of the effort to understand how the atmosphere works, and if it takes longer then you may want to accept that it just does and stay with the project C4. If regretably something prevents the WU from finishing, you would keep the credit accumultaed in your trickles even if there is nothing for the science C5. Sulphur is all the project is currently offering (on BOINC anyhow) and you may feel strongly that you want to continue supporting this project. Those are my thoughts: please add any other A\'s or C\'s ? best regards and happy BOINCing whether you go or stay. And if you go, please check back every month or so as I for one am hoping the scientists & programmers will figure out something useful to do for those of us wanting slab-sized work units -- but those issues are for the other thread, please. |
|
33)
Message boards :
Number crunching :
server state means ??
(Message 18282)
Posted 16 Dec 2005 by old_user85254
Post: Perhaps a simple fix like: when newer trickle arrives, put to state from \"over\" to \"In Progress\" again. This may propitiate people\'s curiosity and make it more consistent with the actual situation of individual model progress. but what I am curious about is exactly what it is that is making it go to \'over\' in the first place? It does not seem to do it to all results, only to some; and it happens well before the deadline so it aint that. Is there some kind of time-out that happens between trickles, or what? Personally I am happy to leave the code as it is, I am saying it is the explanation that is buggy not the code, as users don\'t know what the various values mean. And some hackerish users, like me, want to know as much as we can about such things... |
|
34)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18281)
Posted 16 Dec 2005 by old_user85254
Post: All of the FAQ / tech info / front end of the project, were written in pre-BOINC days. **Exactly** And when increasing the minimum level of committment, the project *needs*, in my opinion, to update the words the describe the minimum committment
But on the BOINC incarnation of the project, slab is currently not being offered, so it makes no sense to still publish advice based on slab. By all means keep the classic advice for the classic platform, I agree that would be appropriate, but surely it is time for some separate words of advice for BOINC participants now that the minimum practical hardware is so different between classic-slab and BOINC-sulphur?
I\'ve successfully slabbed on 500MHz boxes, and on classic ran a model on a 200 MHz Celeron, but that one ended early due to instability and I always wondered if it would have done so on a box within the spec...
An issue, rather than a problem I\'d say. When updating the words, you address this issue by saying that the minimum spec machine is X GHz if on 24/7, and proportionally faster for machines that are only on part time.
The complication comes when people wwho have been running borderline machines successfully on slab, or even sub-spec machines, then meet a sulphur. Many users, myself included, take a great deal of effort to read up on a project just before and just after joining, then leave it running on autopilot once it seems happy. As the slab models come in you will continue to get people panicking when they see 6000 hour completion times, 22 million second completion times on the command line, etc etc. People will, as I did, detach such machines as a reflex action. Therefore an automated screen to prevent sulphur going to boxes below a certain threshold will save the project long waits for boxes that detach. Updated info will save users some anxiety even if they don\'t see it till after they have the first sulphur on board and wonder what is happening. In short, if sulphur is a new experiment, then new experimental guidelines are in order.
Agreed, and if the explanation relates to an expeirment (slab) that is no longer offered on the BOINC platform, then users/donors will \"know\" the wrong limits and will \"know\" only the outdated reasons.
I\'ve taken you at your word here... ;-) |
|
35)
Message boards :
Number crunching :
server state means ??
(Message 18274)
Posted 16 Dec 2005 by old_user85254
Post: ...I only want to be sure not to crunch for the paper-basket... Hi Frank: There is no problem with the \"over\" status, I see it often and the wu still accepts the next trickle. I am sure you do not need to worry about the waste paper basket. Hi Honza: agreed, but as a hacker myself, when I see more than one outcome from what looks lie the same situation I immediately get curious - what does this difference mean, or if it is meaningless, how does it arise? I am not worried, just curious River~~ |
|
36)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18270)
Posted 16 Dec 2005 by old_user85254
Post:
Agree absolutely. If 800 is a sensible limit for slab then 2GHz would be about right for sulphur I don\'t believe that an 800Mhz machine will be able to finish one before the deadline (based on the fact that I have a 1Ghz box that is estimating exceeding the deadline by a few days or so). Or perhaps an update to provide model specific requirements? Depends if you want to be safe in all cases, or to avoid excluding machines that can do the work. Boxes at the same nominal speed vary cosiderably in their throughput. My 700 MHz PIII does over twice the throughout of my 700 MHz Celeron for example, and AMD always outdo Intel at the same clock speed. The 800 limit seems to have been chosen so that all 800 boxes were very comfortable with slab, and the equivalent limit for sulphur might be around 2GHz. I would suggest the following for sulphur a) advise a 2GHz limit b) automatically enforce a 1GHz limit so that boxes slower than that do not get given sulphur WU (this project has never enforced its advice in the past but BOINC does offer the facility to do so) |
|
37)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18268)
Posted 16 Dec 2005 by old_user85254
Post: I have the situation where a sulfur model has put one of my comptuers into EDF: There was a suggestion to cancel sulphur wu using app version before 4.22 -- but I don\'t know if this was endorsed by the project team or just thrown in by a participant and I can\'t find th epost I was thinking of -- would Carl or Tolu comment please? |
|
38)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18089)
Posted 12 Dec 2005 by old_user85254
Post: As a short term solution I see is to get your machines to max specs, provided you have the resources and authorisation to do that. Neither. Most of the sub-GHz boxes are mine but can\'t be stretched to make sulphur viable, in my opinion. A couple of the subGHz boxes belong to friends and I can\'t presume to occupy a firend\'s computer 24/7 for half a year ahead. The 2.8 GHz boxes belong to a local charity and CPDN is running with the permission of one individual who has the discretion to allow it but the organization would not contemplate running those boxes outside the working day. And I won\'t ask for fear of rocking the boat and losing what we already have from them.
I\'m with Les on this one - if the new science needs 64 bit or needs super-minis then that is what the new science needs. You can only break down a given computer task into so many chunks before it goes unstable and science that borders on the unstable is no good to anyone. To take an extreme example, no amount of parallelization could get slab running on an infinite network of ZX spectrums. You might get the code to run but the results would be pure noise. So if sulphur and the oceans need the faster boxes, my suggestion is that the scientists also find some different science that can usefully be done on the slower boxes and alongside the sulphur and ocean runs. One option might be to run yet more slabs in the short term and in the longer term to find other interesting questions that can be split into slab sized chunks. Running two sets of science at once, one for fast boxes and one for slower boxes, will make the best use of the resources the project already has. It entails (sorry to keep coming back to this) some kind of selection process between boxes, whether it is user driven by prefs or automated via benchmarks and %on stats. If not than I can predict that CPDN is going to lose quite a lot of participants and crunchers. Agreed - and your comment suggests a lever for funding apps - \"we have a huge computing resource that is zero cost to public funds, and which we will lose if it is not re-deployed at the end of the slab runs\" could attract funding for a small, separate, research grant - all the scientists need to do is to identify some important question that could be answered in such a proposal. And the beauty of BOINC is that even if those boxes are lost to CPDN, very few will be lost to science. Predictor, for example, can do useful work on boxes too slow for the slab model (though it does need 128M ram). And in a year or so Orbit will be online and current estimates are that orbit WU will be very short and so very suitable for slower boxes. Orbit might be totally unnecessary, but if it finds something it would have more [pun] impact on the planet [/pun] than CPDN. If Predictor find a cure for mad cow disease or cancer then those potential benefits are not trivial either. Einstein & LHC benefit human curiosity and you can see from my stats that I personally rate that as important too. And there are now other projects on the BOINC front page that will attract other CPDN retirees. |
|
39)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18071)
Posted 11 Dec 2005 by old_user85254
Post: River Thanks.
and I guess I am saying the projects requirements have just increased a factor of three without warning and without thanks to those who were welcome a month ago. When we say \'my box can\'t cope with this\' it feels churlish when I am told to like it or leave. Ir does seem to me that a three-fold increase in the minimum acceptable committment to the project should be accompanied with, at least, a comment from the project recognising the fact that some users will inevitably therefore have to leave. The project\'s requirements are what they are - but they are not what they were a month ago.
Good point. And I guess I am saying that a separate experiment will be bound to have different requirements, and therefore it is unfair to simply port everyone across from the old experiment assuming they will accept the new requirements. Some kind of user opt in, or some kind of automated selection of the faster boxes was needed. Moore\'s law seems to apply to science - every three years we can expect the computing needs of the scientists to quadruple, and looked at in that light the increase from slab to sulphur is just what we\'d expect. I hope CPDN get a lot more grants in future; maybe they will need even faster boxes. If so, I hope the project team will put some thought into the responses from this time round, and figure out a more tactful way to break the news that the slower boxes are no longer considered helpful. What is exciting and new to those who can cope, is the end of participation for those who can\'t. And \"parting is such sweet sorrow\" as the bard put it...
I will, thanks Something may turn up. But finding where it is being talked about is the problem. Which is why my ideal solution is to have a preference setting saying I will accept work up to XXXX hours predicted run time based on my benchmarks. Then I can stay connected forever, accept that I may not get any work, but then one day I\'ll look at the GUI and say \"hey, I\'m helping CPDN save the planet again!\". But yes, I will keep checking back every few months or so. River~~ |
|
40)
Message boards :
Number crunching :
sulphur seems slower than slab
(Message 18070)
Posted 11 Dec 2005 by old_user85254
Post: River |
©2024 climateprediction.net