[Universe@home] on Youtube

News and Information related to Distributed Computing
Post Reply
StefanR5R
TAAT Member
Reactions:
Posts: 1661
Joined: Wed Sep 25, 2019 4:32 pm

[Universe@home] on Youtube

Post by StefanR5R »

Sorry, *old* news... but I only saw it now. This was posted last October in the "Cafe" section of the U@h message board.
Krzysztof Piszczek wrote:In June this year I recorded a short video about the project. Today I added subtitles to it, and Google translator made their transcription into English.
I invite you to comment and guess my mistakes :)
Universe@Home on YouTube.
It's a slideshow with Polish narration, and the actual info is in the latter, and thus in the subtitles.
Here is a straight copy of the subtitles:
krzyszp wrote: On May 13, 2014, I was contacted by Mr. Grzegorz Wiktorowicz, the then PhD student in astronomy at the University of Warsaw. Together with Professor Krzysztof Belczyński, they were looking for a way to increase the available computing power in the research project they were working on.

Well, in the research conducted by the team from the University of Warsaw, the gentlemen used a proprietary application written and developed by Professor Belczyński to simulate the evolution of stellar systems. However, despite many optimizations, this program required compilation at that time in the event of a change in the initial simulation conditions, and only thousands of repetitions of the program execution with each compilation allowed to build a database of results sufficient to draw research conclusions.

When I was invited to cooperate, I reacted enthusiastically, in the end it was one of the few Polish projects that had a chance to be launched on the BOINC platform, and also the first in the country to be created under the aegis of a research center. One cannot omit such a "trifle" as the opportunity to work with world-renowned scientists in their field.

We quickly agreed to port the application to as many system platforms as possible and to adapt it to work in a distributed computing environment. Just over a month after the first contact was made, the project's test server was launched, and the news about it began to spread around the Polish BOINC scene ... Of course, the test period could not last forever and on February 23, 2015, the project's production server was launched, which turned out to be very nice adopted by the international community. Since then, apart from very short (several hours) breaks caused by service works or temporary breaks due to traffic overload of the project's links, the server is still working, of course, after enriching with a data warehouse and a separate database server.

For the community, however, it is important what we calculate in the Universe@ome project? I will try to answer it in an accessible form.

Generally very broadly speaking, in the project we simulate the evolution of stellar systems from their inception to the end of their existence, constantly refining the models and comparing them with the observed results. This allows the models to be refined based on - necessarily fragmentary - observational data we currently have.

Nevertheless, to emphasize the importance of this research, it is worth using an example. On September 14, 2015, the detection of gravitational waves was performed in the LIGO detector for the first time in history. For this event to happen, the scientific community around gravitational wave detectors had to decide to observe a specific place in space and decide where there is the greatest chance of making a detection. Professor Belczynski, on the basis of his own simulations performed on his computer, was an advocate of the observation of double black hole systems. As we already know, this decision was right, and the program itself proved to be useful in real and practical application.

In 2017, the project and the entire crew were transferred to the Nicolaus Copernicus Astronomical Center of the Polish Academy of Sciences.

However, how can purely theoretical calculations lead to a researcher becoming convinced that it is such an object and not another that will be worth observing? Let us take another example. Let's deal with black holes for a moment. These objects can be described by three values: - mass - electric charge - spin (the speed at which the black hole rotates). For our needs, we can ignore the charge, because it is very small and does not affect our calculations. Also, the mass is relatively easy to calculate, as we can determine it based on the known parameters of the system - such as the mass of the accompanying star and its separation from the black hole. However, we still have a problem with determining its spin. There are at least three observational methods to estimate it. Neither is perfect, and the basic physical principles used in modeling are greatly simplified. Therefore, the obtained spin values ​​are very uncertain.

Project scientists want to investigate how our astrophysical evolutionary models fit with actual observations, and with this knowledge, gradually improve them. For this purpose, many models are included in the BHSpin application and are tested by mass calculations with a large number of initial values ​​for the initial parameters with which these models are made. Due to the very specificity of black holes, which we cannot see in any way, because even light cannot escape from them, the only way to find them is to observe stellar systems and try to find those where a nearby star is being torn apart by our object, or by its movements determine the proximity of an invisible object of great mass. Once we have the probable location of the black hole, we can start calculating its mass and spin using various models, but we still have no idea what exactly to substitute in the corresponding formulas in the simulation. At this point, we can use the distributed computing model. Currently, more than 100,000 independent simulations are performed every day.

There are several more applications in the project to simulate the evolution of various objects, such as simulating the formation and development of neutron stars, studying the possibility of observing quark stars, and the formation and evolution of black holes.

Over the years, the research team of the project has published over a dozen scientific papers, in which you can find thanks to the volunteers (in our community called "abacus") for the extremely valuable work put into the calculations. All publications are of course linked on the project website. Also, scientific members were appreciated by the astronomy community. AAS Nova, a website run by the American Astronomical Society, highlighted the research carried out on ultra-bright X-ray sources. The authors of the information also stressed that the data were obtained through volunteer calculations. Of course, you will find a link to this publication under the video.

I have been participating in the Universe@Home project for 7 years and I think that I will be dealing with its servers for a long time, and I invite you to count with over a dozen thousand other volunteers :)
(license: CC-BY)
Skillz
Site Admin
Reactions:
Posts: 1854
Joined: Sun Sep 15, 2019 3:03 pm

Re: [Universe@home] on Youtube

Post by Skillz »

That was a good read.

Really makes me wonder just how many projects out there that could benefit from the BOINC platform from whom the scientists have no idea even exists.
Post Reply