news from orbit@home forum
26 Jun 2005 14:42:11 UTC
Topic 189419
(moderation:
To be more precise, I would say that you should start to crunch something by late August, with some sporadic test starting in late July.
The first one will probably be x86 Linux, since we develop on that platform, shortly followed by "32bit Windows", "x86_64 Linux", "Mac OSX".
news from orbit@home forum
)
July 5, 2005
Orbit@home status update: we didn't manage to fix the few website problems still around, and we didn't write the intro page yet. The main reason for this is that we're using all of our very limited resources to try to get orbit@home funded. This first funding wave will end on July 22. After that, a preliminary application and some work units should be available shortly. The subscription phase ended one week after the orbit@home website launch, and for the alpha phase we can count on 1036 users and about 1400 computers, with about [55% Intel, 43% AMD, 2% PPC], [92% Windows, 6% Linux, 2% Darwin] and [78% 1-CPU, 20% 2-CPUs, 2% 4-CPUs or more].
Hi Mr. & Ms. Harriet, What is
)
Hi Mr. & Ms. Harriet,
What is the size of the application for "Orbit@Home"?
Greg
RE: Hi Mr. & Ms.
)
This first funding wave will end on July 22. After that, a preliminary application and some work units should be available shortly.
This says to me that they don't have an application yet, just an idea in the early stages to becoming a project.
Jim
Jim
RE: RE: This first
)
Greg
RE: This says to me that
)
Actually, the project itself isn't new. Simply the BOINC aspect. Orbit is the BOINC incarnation of ORSA.
And Greg, developing a project such as this can take years of development. Look at Climate Prediction. Their climate model started development around 20 years ago.
There are people out there
)
There are people out there that could answer this question much much better than I can, so hopefully one of them will offer to answer. However, I am going to offer just a few things that I can think of, to get a project started.
1) An idea or an experiment that is very calculation intensive.
2) A source of raw data for the above mentioned experiment.
3) If the experiment is to be ran in the BOINC environment, one would have to download and install the boinc server on a computer or computers that can handle the load.
4) An appropriately written client (ie SETI, Einstein, CPDN, LHC, etc) that can download the work units created by the server, do the needed calculations, and upload the results.
5) Some method of analysing the results achieve the goal stated in the experiment.
6) A sufficent financial source to cover any and all expenses from the above mentioned steps.
I'm sure there are many many many more things that can be added to this list, but it's a start.
Do you have an idea that you want to put into production? If you do, I encourage you to talk to the people who have already been down this road. I am sure most of them would be more than happy to help you get started.
Jim
Jim
RE: RE: This says to me
)
You are quite correct. I guess I didn't make myself clear when I made that statement, but that was basically what I meant. Thanks for the clarification.
Jim
Jim
RE: Actually, the project
)
yes and no.
The main person behind ORSA, Pasquale Tricarico, is also the project scientist on O@h, and had hopes right from the start of ORSA that the software could be later applied to a public computing project - so yes you could say work started almost 4 years ago.
And it is equally possible to start the O@h project from the time that Pasquale moved to the Planetary Science Instute to apply ORSA to a distributed computing application sometime this year. THat way of looking at it says that ORSA was worthwhile work in its own right, whether or not it is later used in a BOINC project.
To BOINC-ise existing science still takes time - O@h is likely to take some 5 months writing and testing code. True, the science code exists, and true the BOINC code exists, but writing the specific code to link the two is still a big job. Current estimate is 4-5 months of alpha testing (that is time when code is written in small chunks and incomplete applications tested, so that the programmer knows if the parts that are written so far work). This will be followed by an unknown length of beta testing (testing supposedly complete, supposedly working applications that, inevitably, will reveal some more bugs once they are run intensively).
So it can take a good year, as a minimum, to turn existing code into a production project. Even longer if you need to start off by writing the science code from scratch!
~~gravywavy
Hi all, I have had the
)
Hi all,
I have had the pleasure in the past of working with "PI"(Primary Investigator) at the University of Hawaii. I just ask to see if a "public" project is the same as "closed" project that I hve had experiance with.
I'd like to think that with all the quastions I've been asking, at some point in the future if not sooner, I'll be able to insure a good-stable platform for these researcher.
Greg