Superlink@Technion
Superlink@Technion
Last edited by rebirther on 24.12.2007, 01:13, edited 2 times in total.
Re: Superlink@home
Da war nix mit Konto-eröffnung, wohl aber bei der "Mutter" http://cbl-link02.cs.technion.ac.il/sup ... ttechnion/rebirther wrote:http://cbl-link02.cs.technion.ac.il/cplan/
lebe Dein Leben so wie Du wenn Du stirbst wünschen wirst gelebt zu haben
Wieder was neues mit Änderungen..
Testseite:
Testseite:
Hauptseite:This site is for Technion users only. Another project which will allow to connect your home computer or any computer outside of the Technion is under development.
Superlink@EGEE queue
This project is used to allow efficient invocation in EGEE. Contributors are not allowed to join
Aktueller Stand- Zusammenfassung:
Dear contributors,
Recently you have seen quite a lot of activity on Superlink@Technion. Part of it was related to our tests and benchmarks that we ran to investigate the properties of the system.
Our conclusions are promising: we were able to obtain the equivalent of ~350 modern CPUs running all the time, which is indeed great! the failures are rare, which was really surprising! You can see all the statistics on our graphs which are being constantly generated here.
There are some explanations (just click on the graph), but they are not always easy to understand, so feel free to ask.
Another part is related to the real science, Genetics. Unfortunately not too many good news to report here... I want to explain why. The nature of linkage analysis tasks is that at least 99% of them return false. Namely, for geneticists it means that in 99% of cases the gene mutation which causes the disease is not likely to be where it is being searched for. One may think that all the analyses are meaningless. Of course they are not, because the negative results are used to reject wrong hypotheses, gather more data, and eventually construct the right hypothesis. Thus, in the past month there have not been many successes, but there were some before, which can be found in our list of successes here
You can actually play with the linkage analysis by using our linkage analysis portal Superlink-Online
note, you will not be able to submit the runs (it is password protected), but you will see what data geneticists actually use to find the mutations.
If you want more explanations, feel free to use this forum.
Another thing is that the system becomes more mature and will soon be used for automatic submissions from Superlink-online. Meanwhile all the runs we ran were manually submitted to the system by us. They were real data, and the results were reported back to the geneticists. But all the geneticists will soon use Superlink-online to submit directly to Superlink@Technion. That means that the load may become sporadic: sometimes there will be no runs, and sometimes there will be tons of data to analyze. We always have some data to analyze, so we'll do our best to compensate on the lack of data coming from the users.
I hope I clarified most of the issues here.
Thanks really much to all who help us and contribute their CPUs!
Wieder ein Hilferuf wegen nicht gut laufender Anwendung. Wer helfen will sollte sich unter der oben genannten Adresse bewerben ohne Bezahlung, man kann sich nur einen Namen machen.Friends,
This is a plea for help
We are totally out of manpower to keep the development. The system is now in production, but obviously is not bug free, and these bugs make YOU - our only source of the computing power - very unhappy, and rightly so.
If any of you would like to spend some of your time (for free money-wise, but sharing the system fame and exposing your name on various boards, etc.) to help polish and develop BOINC-related things, we'd be very happy to start the collaboration. Yet, to make things productive we'd appreciate if you send your CV to marksATcs.technion.ac.il.