Gadgetory

I still use this when I'm talking to people with an example of getting so much power out of it now this is an example of a closet simulation cloth simulation right this is this is an endless simulation where the gravity between all of these different bodies is being modeled and this particular algorithm interacts everybody with every other single body it's an N squared algorithm the technical term everything is interacting with everything everything else that's right and we computed one frame at a time and we do it so fast and we're able to do a million particles per second and we see of course this callous email in motion so this is 20,000 particles but because it's N squared that's all hundred of a million interactions each image Wow okay and and we surely admire the work that we've done in the past though what's the big deal now let's let's see let's see Kepler let's see Kepler right so this is all right what are we looking at here this is a whole step this is instead of twenty thousand forty two hundred and eighty thousand volumes but I'm modeling it that the Banzai multipole code out of the observatory of Leiden and in collaboration with her own fedora and Danny have so much success in taking this code and applying all of these amazing pepper things to it now I reckon I recognize there's this I see this where are we right now or enough space on the lower left we have the Milky Way on the upper right we have the Andromeda galaxy the the year is up in the top left hand corner is about 3.8 billion years in the future server we simulated just pretty young we simulated from present day to 3.8 billion years from now so this is not a time machine oh I don't know what is like this perfect example of a time machine and we've reached the moment where really interesting stuff starts to happen the Andromeda galaxy is heading towards us and we are going to end up merging not the first pass later on so this code is based on a in currency code where instead of interacting everybody with every other body we create this tree hierarchy of interactions and we can see the hierarchy right here this is this grid that you see is showing with ever increasing intensity where we are focusing a nuclear power and so each box is grappling about 16 stars and out of 280,000 stars I'm axon offices and so we are using the dynamic parallelism is popped about to very efficiently Thrun out these boxes and generate this tree structure in a way that would not be possible and you can see the beautiful swirl that starts just being flung around there's a 50-50 chance the sun's going to be out there in part of that it's just pulling out of the galaxies well so it's three and a half billion years we are going to collide with dominant that's right it'll eat pastas terribly stas away and come back so I guess the advice to the audience is to put our personal matters in billion years we might end up in the diff reality now think that that's really curious is is to me is all these scars and all these all these stars on the outer fringes of the galaxy don't seem to be moving nearly as slowly as I would have expected I mean in the center where where there's greater mass I would have expected that the planets and the stars to be rotating a much more put they on the outside much more slowly oh boy why am I not seen oh there's something wrong with the simulation so you've actually had one of the great mysteries of astronomy based on exactly these kinds of simulations and observations and stars they realize that the stars in our galaxy and every galaxy we see are going way way to ours and without a good explanation of what this was without the computer models to show it there were they spend about 50 60 years trying to figure it out and finally in models approximately this size but we can now do in real time they figured out that there is in fact only 10 laughter in the galaxy is visible to us and 90% of it is invisible dark matter that lives in a big halo around the outside of the galaxy holding it together and driving in a ladies and gentleman we've identified Dark Matter I need someone on an issue of paper to write up a paper quit