Message boards :
Number crunching :
Projects to mix with LHC
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Send message Joined: 13 Jul 05 Posts: 55 Credit: 41,230 RAC: 0 |
Hey there River~~ Leiden Classical World Community Grid Do you want to get banned for 31 years, your account and credits deleted at a Boinc project ? Predictor@home is your best choice. |
Send message Joined: 13 Jul 05 Posts: 456 Credit: 75,142 RAC: 0 |
Hey there River~~ Has the ROSETTA program fixed their software where it will work well with others? I seem to remember a while back, they had major problems with starting and stopping in a shared environment. Thanks for bbcoding those links Nightbird :) Rosetta There are no reported problems with interworking between Rosetta and other projects on Win2k / XP or Linux or on the supported Mac machines. Some people were still claiming to see problems with win98 & WinMe last time I looked at the issue, but I have since run Rosetta on both without problems, and in a mix with Leiden and/or LHC. In fairness, the project team explicitly disclaim any support of win98 & Me. Rosetta is still, of course, a development project. It has its own alpha project for initial testing of new ideas, but they do produce new apps quite regularly and sometimes these have bugs that get through the alpha undetected. It is not as buggy as an alpha or beta project, but nor is it as reliable as say SETI or EInstein which run the same app for half a year at a time. Having said that, you will usually get credit for work that errors out on Rosetta, as they say that debugging the code is part of the science on their project. World Community Grid WCG is a not-for-profit outfit funded by IBM to run DC projects for the benefit of humankind. Some people wonder where the commercial angle is in that, others take IBM at their word. Certainly WCG is legally separate from IBM and is bound by its own rules not to sell its results for profit. They are a kind of mini BOINC, running more than one project at any one time. A technically interesting point is that they use two different clients. They have their own client that can only runs the WCG projects, and also have an interface that interworks with BOINC. So, for example, to run work for the WCG Fight Aids at Home project, you would tell your BOINC client you want to work on project WCG, and then go to the WCG website to tell their scheduler you want WU for FAAH. All three of the projects they have run to date are life sciences. One of them uses Rosetta code, so running WCG and ROsetta you canget two different versions of the Rosetta app going at once... The BOINC-related stats sites cannot tell which sub-projects you run within WCG - it all looks like a single project from the BOINC interface. They have clearly gone to a fair effort to attract BOINC people, exporting stats in BOINC format (and converting their own credit scoring into cobblestones for the BOINC stats). Personally I use them as my current last-resort project, I have some doubts about the IBM angle but would rather they were fighting aids than my cpus went idle. Leiden Classical No, classical dynamics actually. And as a theoretical tool in Chemistry. In science terms it is almost the opposite of QMC - QMC models quantum dynamics in order to understand chemistry, Leiden models movement of (say) water molecules as classical structures obeying classical force laws between the molecules. Many runs will be made for the same starting points but with different random number seeds, so that an ensemble of different results is used to predict the outcome - but the randomness is the kind of randomness you get in statistical mechanics, not quantum. Clearly quantum and classical dynamics will be useful in different areas of interest to Chemists, but I am not enough of a chemist to say any more. The project is sufficiently general that the various apps can also model movements of planets in the solar system - any system of classical dynamics is in its capability. Like Rosetta, Leiden is doing real work with the WU at the same time as developing the methods. Unlike Rosetta, Leiden is very much a one-man effort, run by Mark who clearly has other things to do as well as run a BOINC project. One nice innovation is that users can set up their own WU - you get to run a maximum of five personal jobs. They would not mean anything to me, so I have not tried it myself, but this looks to me like the first steps towards co-operative computing. I run other people's work most of the time, but when I want something done it can be spread over a lot of other people's boxes to make it (in Mark's words) like a desktop calculator composed of many other computers. It is this feature that attracted me to the project. The first users of his "desktop calculator" will be his students at Leiden U, so it is an educational project as well as a research one. I also like the fact that the project is in its early days and Mark is open to sensible advice when offered tactfully. Expect a higher error rate on Leiden than on say Einstein / SETI / LHC - and probably higher than Rosetta too. And, unlike Rosetta, no credit for routine errors, though Mark does grant manual credit if a mistake of his costs someone a lot of cobblestones. The project is still in development and has not yet got to the point where it has spawned a separate alpha or beta project. R~~ |
Send message Joined: 21 Jan 06 Posts: 46 Credit: 174,756 RAC: 0 |
I must agree with this whole-heartedly. Alpha and Beta projects are some of my favorite projects to read about and check into. RALPH (Rosetta Alpha) seems to be, at least was recently accepting new users, without any delays or quotas for signing up. The sentiments some mentioned for projects like Einstein and Rosetta are indeed just; i also think they are great, thankfully Rosetta has the kinks worked out. Give RALPH a shot if you like alpha projects and ease in jumping quickly up the point standings. I'm running Rosetta at 1, Einstein at 4, LHC at 90, Predictor at 1 (almost always suspended), Ralph at 4, seti at 1 (always suspended), Malaria Control at 4. Happy Crunching all, Mike Molzahn Postscript: It's great to see the spike in activity here recently; most importantly less attention focused toward complaints about work availability. blog pictures |
Send message Joined: 18 Sep 04 Posts: 38 Credit: 173,867 RAC: 0 |
Heres a few Projects Missed so far ( and maybe why )
|
Send message Joined: 13 Jul 05 Posts: 456 Credit: 75,142 RAC: 0 |
I think it is useful to be frank about the downsides of projects (all projects have their upsides and downsides). "If you don't mind" blah is exactly the way to put it to a newcomer. If they do mind those possibilities they will be warned in advance - and in the long run that will cut out a lot of whining. River~~ |
Send message Joined: 18 Sep 04 Posts: 38 Credit: 173,867 RAC: 0 |
River wrote:
We can but hope. btw fwiw i'v been reading the forums @ docking, having to manualy adjust # of stacks for a every different linux/unix host and the translation through wetware on how. it is interesting to note 1 thread 165 long was to help 1 person get their project up and running. if you always wanted to learn linux thats the project to join. sidebar {1} why quote not in a box |
Send message Joined: 13 Jul 05 Posts: 55 Credit: 41,230 RAC: 0 |
Heres a few Projects Missed so far ( and maybe why ) * Presently, Proteins@home doesn't support Linux but in the future, they will do. Officially, the project supports (only) XP but the application runs without problem also on Win2k and NT ; a little more difficult under Win9x/ME. ** Predictor has no work and nobody knows how long the offline period will be. Do you want to get banned for 31 years, your account and credits deleted at a Boinc project ? Predictor@home is your best choice. |
©2024 CERN