Message boards :
Science :
How does Universe@H's BHSpin application differ from the new project BlackHoles@H?
Message board moderation
Author | Message |
---|---|
Send message Joined: 16 May 20 Posts: 7 Credit: 1,724,200 RAC: 0 |
Recently, I came across this new project, which is currently still under development. http://astro.phys.wvu.edu/bhathome/ This is the main scientific journal article linked at their website elaborating on the scientific modelling methodology. https://arxiv.org/abs/1712.07658 Since January the team behind this project are currently actively developing the apps and are set to launch it as standalone BOINC project by August 2021. Does anyone with a deeper scientific understanding know the difference in the 2 applications. From what I understand they will also study black holes but use a new holistic approach with an efficient coordinate system. The website states however, that the application will only require a few GBs of memory and that sounds as this new project will set way higher application requirements as Universe's BH-Spin application. Would love to discuss these two different project/application setups! |
Send message Joined: 10 May 20 Posts: 309 Credit: 4,733,484,700 RAC: 0 |
I've been reading the BlackHoles@Home project too. I too don't quite understand the difference between the two projects. The BH one seems to be a graphically oriented one. Not an actual gpu project but needing lots of resources to plot BH mergers graphically. I too think it will demand a lot more resources compared to Universe. Don't know if it's science or data output will be different or better in any way though. A proud member of the OFA (Old Farts Association) |
Send message Joined: 16 May 20 Posts: 7 Credit: 1,724,200 RAC: 0 |
Would be highly interesting though to understand how they differ. Once the new projects sets up shop with a new URL and a message board, I will likely ask this question there as I am sure that their project team is aware of Universe@Home. Probably will have to stack up on memory :) |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
BlackHoles@Home analyses actual data from BH collisions, as detected from the LIGO detectors in the U.S., and presumably others around the world. It attempts to recover various parameters of the BH, such a spin, with higher precision than is now possible. I believe Universe attempts to show how black holes form and evolve over time from interstellar gasses or whatever. So they are looking at quite different stages in the BH timeline. But that is about as far as I can take it. (I have had a Ryzen 3600 running Ubuntu 20.04.2 with 48 GB of memory waiting for BH@Home for the past year.) |
Send message Joined: 16 May 20 Posts: 7 Credit: 1,724,200 RAC: 0 |
Very interesting insights, thanks Jim for sharing! Your explanation hints at both projects becoming complimentary. You are indeed right that they seem to be modelling real-word data. Good to know! Having skimmed their technical paper, I still don't understand quite a lot, but the modelling approach really seems to be demanding (although Laptops are mentioned on their website). Excited to see how well these nested (adaptive mesh refined) Cartesian coordinate systems will perform. To my understanding, you can model these BH collisions/spins with less detailed coordinates exploiting near-symmetries. Upon further reading, it seems that the project will demand ~1.8 GB RAM per task. They state that by tuning certain parameters in their application (that I don't comprehend at all), the numerical error of this application can be reduced while increasing the required RAM. However this is just a preliminary ballpark number it seems and might be subject to future changes. In the end this app will use about 50 times less RAM to simulate these spins compared to the current standard procedure/application. - That will render my 32 GB on my 8 core machine a definite bottleneck. |
Send message Joined: 9 Sep 17 Posts: 14 Credit: 580,314,930 RAC: 0 |
BlackHoles@Home analyses actual data from BH collisions, as detected from the LIGO detectors in the U.S., and presumably others around the world. It attempts to recover various parameters of the BH, such a spin, with higher precision than is now possible. When two black holes collide they're travelling at about c/2. Only during the last quarter of a second of this collision is a gravitational wave of sufficient magnitude generated to be detected by LIGO. Has anyone installed this and run it yet? |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
Has anyone installed this and run it yet?TL;DR: Tremendous progress has been made, and BlackHoles@Home is on target for an August 2021 launch! http://astro.phys.wvu.edu/bhathome/ |
Send message Joined: 9 Sep 17 Posts: 14 Credit: 580,314,930 RAC: 0 |
I installed it but there wasn't anything to run. Best I can tell is it's for developers. http://astro.phys.wvu.edu/bhathome/download-en_US.html |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
Best I can tell is it's for developers.Isn't it just for simulations? The BOINC version will be using real data. |
Send message Joined: 16 May 20 Posts: 7 Credit: 1,724,200 RAC: 0 |
They just postponed the estimated start of the project and pushed it back one quarter. The revised start date is now fall 2021. Additionally, the project that was to my understanding affiliated with the West Virginia University, now changed its project's lead to the University of Idaho. |
Send message Joined: 15 Aug 20 Posts: 37 Credit: 8,554,094,169 RAC: 0 |
They just postponed the estimated start of the project and pushed it back one quarter. The revised start date is now fall 2021. As a WVU alumni, I was excited for this project's release. so it's unfortunate to hear that they switched affiliation. I don't think the WVU researchers have left WVU for Idaho, so I'm not sure the reason for the project's move. |
Send message Joined: 2 May 20 Posts: 19 Credit: 4,561,756,400 RAC: 0 |
The lead professor, Dr. Etienne, is associated with both universities according to his profile page on the site. I am looking forward to hear more about this project as well. Roger |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
The lead professor, Dr. Etienne, is associated with both universities according to his profile page on the site.Interesting. It appears that as an "Associate Professor", his main affiliation is now Idaho. The University of Idaho at Moscow is not that far from the Hanford LIGO Observatory. There may be a number of staff at both places. |
Send message Joined: 15 Aug 20 Posts: 37 Credit: 8,554,094,169 RAC: 0 |
Anyone have any additional info that’s not on the black holes page? At one point they projected a Fall 2021 launch, but that seems to have fallen through. Anyone know if it’s still a planned project? |
Send message Joined: 16 May 20 Posts: 7 Credit: 1,724,200 RAC: 0 |
I've recently seen a spike in activity on the website in late Marche/April. Seems like the conducted a lot of software development. RAM requirements stated in one of the recent news was ~3 GB per task and as such much less demanding than initially feared. This is taken from the website: After a year spent rewriting the software behind BlackHoles@Home, we're seeing beautiful results. In short, the software seems to do what it needs to do, but I'm still running it through some trials.Black Holes @ Home website Unfortunately still no concrete roadmap / launch date. Looking forward to it |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
I've recently seen a spike in activity on the website in late Marche/April. Seems like the conducted a lot of software development. RAM requirements stated in one of the recent news was ~3 GB per task and as such much less demanding than initially feared. I had some private communication with Prof. Etienne about six months ago, and he indicated that the work units could be only about 2 GB or so. That would make it possible for a lot of people to run at least one. But that may change of course, so I would not assume it until it actually appears. (And there could be some bigger work units during development too I think.) |
Send message Joined: 5 Sep 15 Posts: 3 Credit: 556,800 RAC: 0 |
New link. https://blackholesathome.net/ |
Send message Joined: 14 Jul 19 Posts: 6 Credit: 9,033,333 RAC: 0 |
|
Send message Joined: 17 Jan 21 Posts: 9 Credit: 65,626,633 RAC: 0 |
The installation manual contains 80 steps (!) |