This video is by far theeeeeee best I seen on UA-cam in a long long long time. So there is nothing I can do but put among the tabs and watch it all over again. I done a lot of computer programming in my life. I watch all kind of documentaries every day on UA-cam physics, military, history u name it.. In this video they had the balls not to dumb it down. I loved it. BTW B+ -trees are from database sorting on really weak computers back in 1970s and 1980s. B+ trees dont get so deep as B-trees. Makes the data align in chunks in memory ready to be written in bulk to disk. In that applicaiton the B-tree page size usually match the system OS page size.
I finished chemistry in high school and successfully made my own chemical reaction algorithm to predict whether or not two chemicals will react to each other. It’s not much but it’s honest work.
If you study Electrical engineering and you used 3D modeling before, you know why this is an awesome breakthrough in technology. It offers unprecedented level of details right where it matters. Great video, very glad it was shared on the internet.
Wow! Not only the technology they're presenting is impressive but also their presentation skills are great! Seems like a very complicated problem and they present it in a way that one understands what's going on. I wish I could do that.
I'm wowed as well, linked in from spacex lounge, I wonder if other companies now use these simulations, a week rather than typically a year? I'm blown away!!
japrogramer I'm sure it won't be released for public use because this software lowered the barrier for fluid dynamics in hypersonic region. Basically, easier development for hypersonic cruise missile, ballistic missile interceptor, anything related to very fast weapon delivery and nuclear thingy.
I wonder the limitations of this technique. I tried for 3 months to think of a way to apply this to cellular mechanisms. Its hard to connect it. I love this video!
Not sure. It does seem like an obvious rocket to use given the need for massive payload from Earth. I don't think rockets are the way to go on the return leg. As noted above, you're at the equivalent of 100,000 ft on earth, and we typically jettison the payload protective shroud at that altitude. So if they are serious about colonization of Mars, build an electro-magnetic launcher on Mars and to launch people and product back to Earth.
because the unfortunate truth for now is that just because you build a superheavy rocket, demand for its payload capabilities don't just materialise overnight. Starship is probably at the iffy sweet spot where it's large enough to start exploiting large rocket efficiencies, but small enough to minimise the risk of being too large for the current market demand
cool presentation Redhat so TL;DW: a realtime CFD sim of that much data output with current GPU and storage tech (even if it's some kind of 10,000 disk RAID SSD array) anytime soon... is not happening
I have a feeling that we still could use a "catapult" type system in extremely large rockets, similar to what an aircraft carrier use. The amount of thrust needed for the "first" lift is significantly larger than what the engines need to for the rest of the flight.
A fully loaded Falcon 9 weighs something like 500 tons, realistically you are not capable of building a catapult that can store and release enough energy to make a difference. You would have to put a nuke under the rocket, but that would break it.
have partical accelerators been used to experment with hypergolic chemical thermal expansion ratios temperature after basic particle research for temperature pressure and raio of mix resaearch...also suggest a model of observation in sunspot rims from subnspot surface gases combine in extreme burns..number of reactions reviewed by xray defraction and spectromiters.
1% atmosphere... If my calculations are right, that means that it's roughly equivalent to the pressure on Earth at 30 km in altitude? (or at 100,000 feet for you non-metric boys)
I disagree about unstructured being 'far better' than structured and requiring 'yottabites of data.' Why? Because structured meshes allow for far, far more efficient simulations, potentially THOUSANDS of times faster for a given resolution. How? Because unstructured meshes have arbitrary connectivity, and they're limited to second-order accuracy at best.
Complex dynamics for sure, get thee impression none of it has been done, mostly 45 years ago. Now why cant they just start going to the Moon, helium3 paradise? If I were to volonteer I sure would like, USA that does'nt own the Moon has only forbiden to approche Apollo landing sites.
@@pasoundman SpaceX provides satellite launches to countless international customers and they're building a world wide satellite based internet service. They're actively doing this right now, how's that for return on investment?
@@TaliwhakerRotmg That's great, there's a healthy market for that and of course they sell rides to NASA, saving NASA the trouble of developing spacecraft (just look at the delays with the Boeing capsule). I was referring exclusively to the bonkers ideas of colonising Mars !
@@pasoundman The hope is that Mars exploration will invigorate science and technology the way the moon landing has. It's an ideological purpose, not a practical one.
Please more. It’s been 6 years!
This video is by far theeeeeee best I seen on UA-cam in a long long long time. So there is nothing I can do but put among the tabs and watch it all over again.
I done a lot of computer programming in my life. I watch all kind of documentaries every day on UA-cam physics, military, history u name it.. In this video they had the balls not to dumb it down. I loved it. BTW B+ -trees are from database sorting on really weak computers back in 1970s and 1980s. B+ trees dont get so deep as B-trees. Makes the data align in chunks in memory ready to be written in bulk to disk. In that applicaiton the B-tree page size usually match the system OS page size.
I finished chemistry in high school and successfully made my own chemical reaction algorithm to predict whether or not two chemicals will react to each other. It’s not much but it’s honest work.
Cool, how do you know if they will react or not?
If you study Electrical engineering and you used 3D modeling before, you know why this is an awesome breakthrough in technology. It offers unprecedented level of details right where it matters.
Great video, very glad it was shared on the internet.
Nino Dvoršak trekkie sve ti je to show i cgi. fora za debile. zemlja je ravna i nepomicna.
Fotošop
Elektrotehnik electrical? Not even! Try it's intended purpose, aerospace. I'm used to 48 simulations that produce 1/10 the data
Wow! Not only the technology they're presenting is impressive but also their presentation skills are great! Seems like a very complicated problem and they present it in a way that one understands what's going on. I wish I could do that.
Awesome video. Awesome people. They make something very complex, relatively easy to understand. 10 out of 10.
Been trying since I saw this video to recreate the simulation... It's been a few years but I'm getting there!
Physics is art.
Ah, 2015 ... when people thought SpaceX needed an introduction :-)
Impressive is almost an understatement ! Thanks for posting.
Yay for having played Kerbal Space Program and understanding what the heck he's talking about =)
I don't remember KSP teaching you wavelet compression...
I'm wowed as well, linked in from spacex lounge, I wonder if other companies now use these simulations, a week rather than typically a year? I'm blown away!!
Great presentation, awesome technology, beautiful modeling.
Does someone have a high res image of the flame front simulation at 38:56 that I can use as my wall paper?
So as far as I can tell at 29:46 he is talking about using a hash table for data indexing. Does anyone disagree? He doesn't say it explicitly.
at 4:28 Image credit kerbal space program (:
DahPurpleHippo Ahaha, you get very sharp eyes :'D
Did he say founded in 2012 at 1:48?
Yes
Enjoyed every second!
Is this fast enough for real-time applications such as a flight simulator game?
theinsekt lol no.
So no hyper realistic aerodynamics and heat simulation in kerbal space program :( Maybe it will be possible far in the future.
theinsekt
Maybe if you had a Cray XK7 or IBM BlueGene/Q.
Patchuchan Not even then!
+theinsekt Not now. It might be in the future as technology improves though.
where is the download link for the software?
japrogramer I'm sure it won't be released for public use because this software lowered the barrier for fluid dynamics in hypersonic region. Basically, easier development for hypersonic cruise missile, ballistic missile interceptor, anything related to very fast weapon delivery and nuclear thingy.
Lol, you developing rocket engines?
I wonder the limitations of this technique.
I tried for 3 months to think of a way to apply this to cellular mechanisms.
Its hard to connect it.
I love this video!
15:50 H264 is an HD video codec by the way
..."so anyone who has played kerbal space program has seen this..." lel I agree
37:35 you did that on a single gpu hahahaha what the fuck ?
Great video! Nostalgically recall the master degree times
which softwares do you use ?
When is the ICO? :D
Beautiful minds. 👌🏼
Brillant!
does anybody know why don't companies use sea dragon-like architecture for "supermassive" rockets?
Not sure. It does seem like an obvious rocket to use given the need for massive payload from Earth. I don't think rockets are the way to go on the return leg. As noted above, you're at the equivalent of 100,000 ft on earth, and we typically jettison the payload protective shroud at that altitude. So if they are serious about colonization of Mars, build an electro-magnetic launcher on Mars and to launch people and product back to Earth.
because the unfortunate truth for now is that just because you build a superheavy rocket, demand for its payload capabilities don't just materialise overnight. Starship is probably at the iffy sweet spot where it's large enough to start exploiting large rocket efficiencies, but small enough to minimise the risk of being too large for the current market demand
cool presentation Redhat
so TL;DW: a realtime CFD sim of that much data output with current GPU and storage tech (even if it's some kind of 10,000 disk RAID SSD array) anytime soon... is not happening
04:50 Potatoes. Lots of Potatoes...
I have a feeling that we still could use a "catapult" type system in extremely large rockets, similar to what an aircraft carrier use. The amount of thrust needed for the "first" lift is significantly larger than what the engines need to for the rest of the flight.
A fully loaded Falcon 9 weighs something like 500 tons, realistically you are not capable of building a catapult that can store and release enough energy to make a difference. You would have to put a nuke under the rocket, but that would break it.
have partical accelerators been used to experment with hypergolic chemical thermal expansion ratios temperature after basic particle research for temperature pressure and raio of mix resaearch...also suggest a model of observation in sunspot rims from subnspot surface gases combine in extreme burns..number of reactions reviewed by xray defraction and spectromiters.
1% atmosphere...
If my calculations are right, that means that it's roughly equivalent to the pressure on Earth at 30 km in altitude? (or at 100,000 feet for you non-metric boys)
Turbulence is not fractal in nature. A foundational property of turbulence is that it is NOT self similar.
Fractals don't have to be self-similar either, whatever that means
@@blinded6502 A figure is strictly self-similar if the figure can be decomposed into parts which are exact replicas of the whole.
I definitely recommend watching at 1.25x, great video though.
settings > speed
danthemanzizle I always watch this type of videos at 1.5x.
Michael Hart Demigod technically. You're welcome, I hope you enjoyed it as I did.
danthemanzizle if someone asks you if you are a god, you say YES!
+danthemanzizle Much better!!!
35:37
"Our current engines run on kerosene and on _EXPLOSION_!"
Selamat kerja team semua disana .
so is this the best graphics on earth?
SpaceX Gpu benchmark!
No more war!!!
I disagree about unstructured being 'far better' than structured and requiring 'yottabites of data.' Why? Because structured meshes allow for far, far more efficient simulations, potentially THOUSANDS of times faster for a given resolution. How? Because unstructured meshes have arbitrary connectivity, and they're limited to second-order accuracy at best.
Now, KSP should use this to calculate and render its physics.
Nasa wants back their supercomputer... They did't give to you to play ksp on it -_-
Benedek Nagy too late
Complex dynamics for sure, get thee impression none of it has been done, mostly 45 years ago. Now why cant they just start going to the Moon, helium3 paradise? If I were to volonteer I sure would like, USA that does'nt own the Moon has only forbiden to approche Apollo landing sites.
KSP for the win 😂😂😂
When he said the think of Ksp I smiled so much😄
i want a silicon valley episode about this
Quantum computer available anyone?
One simple question. Exactly who will pay for all this ?
Private investors. SpaceX just got a 70+ Billion dollar valuation.
@@TaliwhakerRotmg I hope they have deep pockets ! fail to see any route for generating a return on investment from going to Mars.
@@pasoundman SpaceX provides satellite launches to countless international customers and they're building a world wide satellite based internet service. They're actively doing this right now, how's that for return on investment?
@@TaliwhakerRotmg That's great, there's a healthy market for that and of course they sell rides to NASA, saving NASA the trouble of developing spacecraft (just look at the delays with the Boeing capsule). I was referring exclusively to the bonkers ideas of colonising Mars !
@@pasoundman The hope is that Mars exploration will invigorate science and technology the way the moon landing has. It's an ideological purpose, not a practical one.
"#FU nVidia!!!1!"
WRONG!
marshalcraft ...what?
I wonder if Elon Musk paid to this small detail engineering work.