Watching the simulations reminds me of water on a counter top or tile floor. The water clumps together and 'seeks' to form a round shape. Excellent works guys.
This video does a wonderful job of expressing the Joy of Learning and Discovery. Well done Roger, James, and Brady! their research could indeed have huge implications to industry and other sciences. understanding how to best mix (or separate) materials is critical to so many subjects and to the future of humanity.
If we can join a cloud computing network to share our computing power to these guys, just think what it could do for them. That way when I'm not gaming and at work I could lend the power of my dual lga2011 system with crossfire setup and in part double their computing power right there. Of course the up/down of the isp would limit things a bit but If I could devote specific time periods to them it would greatly improve their computing power.
They have. The shaved head dude is the World Champion Tetris player. He is developing new variations of Tetris, according to a previous video. I presume he has his own game development studio.
6:38 "But you did retire a year ago!" "I know, but I've been working on this ever since! This is fun!" This has got to be one of the greatest things I've heard in a while. Really shows how enthusiastic scientists are about their work, and really how everyone should follow this sort of work ethic. A very off-the-cuff quote, but nonetheless hats off to Dr. Bowley.
More questions: With our atmosphere why doesn't the various elements making it up not separate into pockets, why do they mix? It seems surface tension effects things on a smaller scale(paper clip on the surface of water). How does surface tension effect larger things?
The computer simulation replicates the experiment very well; the computer simulation has no water on the surface of the particles so that thee surface tension we obtain from the computer simulations is for dry grains. If the grains have a film of water, then the water forms liquid bridges between the grains which increases the surface tension. So you are correct to suggest that moisture has an effect. But that is the subject of work already done by Mike Swift and his German friends.
This is amazing! The beautiful thing about physics is that it is the same on both the micro and the macro. If they can come to understand this then we could gain a greater understanding of our universe as a whole and could even model the universe from beginning to end in superb detail!
Think about if the momentum of a grain that is in the middle of the clump, it goes outward into the surrounding grains and loses all it's momentum from that, while the others it effected lose the momentum because of friction. And to the opposite viewpoint of a grain on the edge of the clump, it very well might be thrown away from the clump but as I said before it will just roam around until it finds the clump again.The momentum is lost and turned into other forms of energy upon every collision.
Time? Particles frying around in a big universe have to sometime stick together right, not that I know anything about this stuff but that makes sense in my head
No. The system has to be in stable steady state equilibrium before that consideration applies, and if a volume of air is in equilibrium then the pressure is everywhere the same. If a system is in equilibrium and there are no potential gradients and *still* there are pressure differences, then there is surface tension involved (or so I gathered from this video). However, the Earth's atmosphere is never in a steady state equilibrium and so pressure fluctuations are down to dynamic variations.
This is really cool! Every paper should have such a video. :) This would be amazing!! Great stuff. Keep it coming brady. This is really making science come alive. Looking how passionate the researchers are makes me really want to study harder and more passionate! :D
the patterns emerge due to the fact that the shaking surface is not perfect. so the grains have the tendency to not just bump into each other but even get distracted from the surface, so they add up in the -> lower area of the surface..
I'm sure they thought about possibilities like this, also like if there are significant electrical forces generated by the surrounding equipment which affect the bronze particles, or vibration modes of the plate exciting particles more in some areas than others but notice that they did a computer simulation which wouldn't cause any of these effects and it produces similar results (though I don't claim to know the details)
A graphics card in itself has a number of units, that are calculated in parallel. Adding more graphics cards just increases the number of units. They are still calculated in parallel. To make an analogy with the grains, say one graphics card can calculate the relative velocities of 200 grains. 40 graphics cards can calculate it for 8000 grains. All in parallel. It's like zooming out and seeing the bigger picture.
Could this explain that stage of planet formation where we understand why particles started clumping on the atomic level and on a bigger, asteroid-sized level, but not in between?
there are zero da and ad conversions, because the 2 cards are not actually producing any images. If you watch the video again, you might notice the monitor cable goes to a third card in the middle
Wow. The way the grains separate remind me of the way astronomical bodies form and cluster together to form solar systems and galaxies. Perhaps the secret to gravity is something relating to this research. Keep up all the work
I'm wondering how they define "noncohesive", and also whether they controlled the humidity inside the box of particles. It's well known that most materials under ambient conditions contain water on their surface... and these 160um particles will have several layers of water on them under ambient conditions. Doesn't appear the manuscript is Google Scholar'able just yet.
Learning how to program will leverage any engineering skill set, or even business administration. The programming language doesn't matter much, but you'd do yourself a favor if you take something that's current in the field that you're interested in, and not too bare bones. Examples are Matlab (especially for physics/engineering, use Freemat if your budget is finite or nonexistent), Python (general purpose computing), Java (also useful for web development), R (statistics, check Rstudio).
Yeah but the point of the experiment is to test for surface tensions in things other than water, where the little grains would unite and actually lift something denser then themselves. Hence the graph and the little blob that they kept simulating over and over for results.
ATi/AMD actually do have something similar which is OpenCL and AMD Stream, but they are not as well developed and supported as CUDA. Anyways from the look of the shorud (and the fact that he stated he can fry an egg on it lol) looks like they are Fermi based Tesla cards.
They should set up a BOINC distributed computing project. I bet with all the RUclips viewers for these videos they would be able to get some quite quick computing done that way.
My thinking: When they are shook; a single grain would travel farther than grains clumped, the grains that hit other grains will distribute their momentum among the colliding grains, slowing it down. And if there were enough grains in the clump it would effectively stop the moving grain and wouldn't effect the clump enough. Hence it's only logical that the grains would fly about until gathering. Does that make sense? Hard to put my thoughts into 500 characters or less.
Let me see if I have this correct. So if I put two different elements granulated in a tray and vibrate the tray, say copper and nickle, because they have two different amounts of surface tension they will separate into clumps. These are my questions: What relationship does surface tension have to do with mass? How would this work in an environment without gravity? Does irregular sizes of the granules effect the experiments results, causing like masses to group rather than like elements?
Free energy in this system is added due to the shaking right? It's only because of the decrease in energy due to the collisions that the particles "cluster" together.
I'd be curious to know the results of altering the frequency and intensity of the vibrations. Wouldn't the surface tension disappear as soon as the vibration is stopped? Or is there a threshold below which there is no surface tension? It seems that the energy of the vibration would have defined wave characteristics that might be imparted to the material.
Is this why minerals seem to cluster when in the soil? I've always wondered why gold particles clump into nuggets, is it related to this idea of surface tension?
the black and white images of the particles clumping together looks very much like the microwave background images do you think this system could apply to the very early universe
how do you know that it's not the weight of the granules that cause the plate to have nodes of less movement? I don't think it would matter if it were done in space, because there would still be a force applied to the plates cause oscillation at a given frequency, and that would cause this seemingly random gathering. (in my opinion).... you know what I mean?
This was three years ago...I wonder what kind of rig they're running their simulations on now. I'm expecting either liquid cooling or mineral oil immersion at the very least....these are physicists after all.
Geographers, Mining Engineering, Analytical Chemistry and machinery in general have dry particle behaviour to deal with. The combination of research is interesting and useful. Is the property being studied being called "inertial segregation"? There's a lot of physics in the video.
Could it be that aerodynamics is when you push air it has this sort of energy loss and clumps together with other air particles forming complex spiral patterns and differentiation of pressure.
Perhaps these findings can contribute to the thus unsolved problem of mixing nonuniform particles and fluids. It may seem strange, but we still can't model what happens when you stir grains of different sizes and/or densities in a fluid. The act is ubiquitous in daily life (stirring your tea or coffee) and industrial processes (making solutions and suspensions), yet it's very poorly understood. I don't doubt that this phenomenon of surface tension in granular media is at least partially at play.
a questions. Are particles at less dense areas moving faster, and could this be one of the reasons surface tension is formed? (so that in tightly clusetred areas particles are basically hitting each other as often as highspeed particles in less dense area.
Heat-sinks aren't for "dissipating the power", they're for dissipating heat. Just a slip-of-the-tongue, but if people don't know a lot about a subject, and are trying to learn it from you, it's best to get it right.
Well, water at room temperature has a huge surface tension because: -thermal particle (molecules) movement due to the 300Kelvin temperature -hydrogen bonds -london dispersion(...) and other forces Liquid helium is a superfluid on the other hand, therefore it has no surface tension, because of : -lack of thermal (molecular/atomic) agitation -single He atoms don't have any electric charge -london forces are already using all of their strength to collapse the helium atoms from a gas into a liquid ... So liquid helium has no surface tension at all ! Also, if you induce a general kind of oscillation (which compress all the particles at a period of time=T, in this case: a bunch of bronze granules), they will clump together because of the friction between them and the movement produced by you... Try the same experiment with a cluster of glass and plastic granules - the 'surface' forces will be even stronger because you added an extra-force : the electrostatic attraction. ...physics works every time...
Can anyone point me in the direction of more videos done by Brady that are about computers? Perhaps Brady could do a a coding language episode (e.i. python, java etc.) on numberphile or one of his many channels if he hasn't already. I love these videos so much, and thanks.
If cooling seems to be a limiting factor, run water not fans...and your competitors could blow all that away with ASICs to hash out the values instead of GPUs, you might want to start looking into that --- physiii
imagine what they're can do with a TITAN Xp now.
Watching the simulations reminds me of water on a counter top or tile floor. The water clumps together and 'seeks' to form a round shape. Excellent works guys.
Oh I love that Roger is still with us, such an awesome guy, wish I could have had him as a teacher :)
This video does a wonderful job of expressing the Joy of Learning and Discovery. Well done Roger, James, and Brady! their research could indeed have huge implications to industry and other sciences. understanding how to best mix (or separate) materials is critical to so many subjects and to the future of humanity.
I love videos like this! Very inspiring. keep up the great work guys!
Good work Brady, and congratulations to Dr. James Clewett!
Fascinating to see the borders of our knowledge, not in astrophysics or quantum mechanics but in grains of sand.
If we can join a cloud computing network to share our computing power to these guys, just think what it could do for them. That way when I'm not gaming and at work I could lend the power of my dual lga2011 system with crossfire setup and in part double their computing power right there. Of course the up/down of the isp would limit things a bit but If I could devote specific time periods to them it would greatly improve their computing power.
This video is much clearer than the original one.
James Clewett is such an inspiring lad
8:25 i love his enthusiasm its so pure and raw
Yeah, I'd love to see some videos on the practical scientific implementations of CUDA.
I wish, I love learning about this stuff
looks like the first step to get adaptive/regenerative armor based on surface tension :D
I dislike that one person that dislike this video and the one that didnt install proper heat sinks in those processors and video cards.
They have. The shaved head dude is the World Champion Tetris player. He is developing new variations of Tetris, according to a previous video. I presume he has his own game development studio.
I'm glad I watched this. I was wondering what the paper actually said
6:38 "But you did retire a year ago!"
"I know, but I've been working on this ever since! This is fun!"
This has got to be one of the greatest things I've heard in a while. Really shows how enthusiastic scientists are about their work, and really how everyone should follow this sort of work ethic. A very off-the-cuff quote, but nonetheless hats off to Dr. Bowley.
More questions: With our atmosphere why doesn't the various elements making it up not separate into pockets, why do they mix? It seems surface tension effects things on a smaller scale(paper clip on the surface of water). How does surface tension effect larger things?
The computer simulation replicates the experiment very well; the computer simulation has no water on the surface of the particles so that thee surface tension we obtain from the computer simulations is for dry grains.
If the grains have a film of water, then the water forms liquid bridges between the grains which increases the surface tension. So you are correct to suggest that moisture has an effect. But that is the subject of work already done by Mike Swift and his German friends.
This is amazing! The beautiful thing about physics is that it is the same on both the micro and the macro. If they can come to understand this then we could gain a greater understanding of our universe as a whole and could even model the universe from beginning to end in superb detail!
I think I saw "C2050" as the camera panned up on them, meaning they're last generation's Fermi-based Teslas.
Think about if the momentum of a grain that is in the middle of the clump, it goes outward into the surrounding grains and loses all it's momentum from that, while the others it effected lose the momentum because of friction. And to the opposite viewpoint of a grain on the edge of the clump, it very well might be thrown away from the clump but as I said before it will just roam around until it finds the clump again.The momentum is lost and turned into other forms of energy upon every collision.
I never thought that I would find surface tension a fascinating subject... but it turns out it is.
Time? Particles frying around in a big universe have to sometime stick together right, not that I know anything about this stuff but that makes sense in my head
No. The system has to be in stable steady state equilibrium before that consideration applies, and if a volume of air is in equilibrium then the pressure is everywhere the same. If a system is in equilibrium and there are no potential gradients and *still* there are pressure differences, then there is surface tension involved (or so I gathered from this video). However, the Earth's atmosphere is never in a steady state equilibrium and so pressure fluctuations are down to dynamic variations.
This is really cool! Every paper should have such a video. :) This would be amazing!! Great stuff. Keep it coming brady. This is really making science come alive. Looking how passionate the researchers are makes me really want to study harder and more passionate! :D
the patterns emerge due to the fact that the shaking surface is not perfect. so the grains have the tendency to not just bump into each other but even get distracted from the surface, so they add up in the -> lower area of the surface..
I'm sure they thought about possibilities like this, also like if there are significant electrical forces generated by the surrounding equipment which affect the bronze particles, or vibration modes of the plate exciting particles more in some areas than others
but notice that they did a computer simulation which wouldn't cause any of these effects and it produces similar results (though I don't claim to know the details)
A graphics card in itself has a number of units, that are calculated in parallel. Adding more graphics cards just increases the number of units. They are still calculated in parallel. To make an analogy with the grains, say one graphics card can calculate the relative velocities of 200 grains. 40 graphics cards can calculate it for 8000 grains. All in parallel. It's like zooming out and seeing the bigger picture.
Could this explain that stage of planet formation where we understand why particles started clumping on the atomic level and on a bigger, asteroid-sized level, but not in between?
there are zero da and ad conversions, because the 2 cards are not actually producing any images. If you watch the video again, you might notice the monitor cable goes to a third card in the middle
Wow. The way the grains separate remind me of the way astronomical bodies form and cluster together to form solar systems and galaxies. Perhaps the secret to gravity is something relating to this research. Keep up all the work
You can say a lot about swallows flocking, bacteria colonies, and fundamental forces such as gravity, and the cosmic microwave background (CMB).
That computer would make any gamer cry.
Thank you kind person. From Roger Bowley. Professor
Cause those cards only calculate stuff, they are not used for video output at all. The middle card is just a simple one for the screen.
but can it run crisis?
but can it beat dark souls?
I'm wondering how they define "noncohesive", and also whether they controlled the humidity inside the box of particles. It's well known that most materials under ambient conditions contain water on their surface... and these 160um particles will have several layers of water on them under ambient conditions. Doesn't appear the manuscript is Google Scholar'able just yet.
Learning how to program will leverage any engineering skill set, or even business administration. The programming language doesn't matter much, but you'd do yourself a favor if you take something that's current in the field that you're interested in, and not too bare bones. Examples are Matlab (especially for physics/engineering, use Freemat if your budget is finite or nonexistent), Python (general purpose computing), Java (also useful for web development), R (statistics, check Rstudio).
Yes, but it has gone into the mathematics behind the formation of these patterns which gives a clue as to the underlying physics.
great video
The best part about that idea is that it's totally doable.
I'm really interested in full configuration of that computer.
Yeah but the point of the experiment is to test for surface tensions in things other than water, where the little grains would unite and actually lift something denser then themselves. Hence the graph and the little blob that they kept simulating over and over for results.
ATi/AMD actually do have something similar which is OpenCL and AMD Stream, but they are not as well developed and supported as CUDA. Anyways from the look of the shorud (and the fact that he stated he can fry an egg on it lol) looks like they are Fermi based Tesla cards.
love this, so exciting. 08:13 he's thinking of starlings. I don't think swallows do that, not in England anyway, but starlings do.
They should set up a BOINC distributed computing project. I bet with all the RUclips viewers for these videos they would be able to get some quite quick computing done that way.
It's awesome to see a physicist excited.
Very interesting!
I am working with simulations using CUDA (nvidia) in my physics course too.
I wonder how different this would turn out if you used a circular "box" to shake the grains...
My thinking: When they are shook; a single grain would travel farther than grains clumped, the grains that hit other grains will distribute their momentum among the colliding grains, slowing it down. And if there were enough grains in the clump it would effectively stop the moving grain and wouldn't effect the clump enough. Hence it's only logical that the grains would fly about until gathering. Does that make sense? Hard to put my thoughts into 500 characters or less.
I wholeheartedly support this idea. Give the science boys (and gals) the best tools to do what they do best.
Let me see if I have this correct. So if I put two different elements granulated in a tray and vibrate the tray, say copper and nickle, because they have two different amounts of surface tension they will separate into clumps. These are my questions: What relationship does surface tension have to do with mass? How would this work in an environment without gravity? Does irregular sizes of the granules effect the experiments results, causing like masses to group rather than like elements?
That would be awesome, and if I remember correctly James Clewett use to be a software engineer.
that computer is a beast
There are already some rather good CUDA programming tutorials out there...
Free energy in this system is added due to the shaking right? It's only because of the decrease in energy due to the collisions that the particles "cluster" together.
Will there ever be a colloboration with the 1,000,000 ARM Core project?
The simulation would be awesome... wouldn't it?
I'd be curious to know the results of altering the frequency and intensity of the vibrations.
Wouldn't the surface tension disappear as soon as the vibration is stopped?
Or is there a threshold below which there is no surface tension?
It seems that the energy of the vibration would have defined wave characteristics that might be imparted to the material.
Is this why minerals seem to cluster when in the soil? I've always wondered why gold particles clump into nuggets, is it related to this idea of surface tension?
Love this video, this was almost better than the main video. xD Really funny, too.
I was thinking exactly the same thing.
"Intel Xenon processors" might be the most adorable slip of the tongue by a scientist/science geek ever.
What an excitement :D 8:22 Lucky guy!!
Yeah but can you get it to pause there?
This phenomenon reminds me of Ricci flow
so exciting!!
Because most programs are made to run in serial computation and it's architecturally simply not possible to run it FASTER on a parallel processor.
Awesome.
the black and white images of the particles clumping together looks very much like the microwave background images do you think this system could apply to the very early universe
awesome
It's a bit hard to make out, but if you look at 2:56 you can see that they're two Tesla C2050.
ps...how do you know it's the grains of bronze grouping together and not the space between pushing them together as the space grows?
Can you share the specs for that workstation computer?
This work is great, i'd love to see updates in the future when you have more information on it.
how do you know that it's not the weight of the granules that cause the plate to have nodes of less movement? I don't think it would matter if it were done in space, because there would still be a force applied to the plates cause oscillation at a given frequency, and that would cause this seemingly random gathering. (in my opinion).... you know what I mean?
This was three years ago...I wonder what kind of rig they're running their simulations on now. I'm expecting either liquid cooling or mineral oil immersion at the very least....these are physicists after all.
Cool stuff!
Geographers, Mining Engineering, Analytical Chemistry and machinery in general have dry particle behaviour to deal with. The combination of research is interesting and useful.
Is the property being studied being called "inertial segregation"? There's a lot of physics in the video.
Could it be that aerodynamics is when you push air it has this sort of energy loss and clumps together with other air particles forming complex spiral patterns and differentiation of pressure.
Perhaps these findings can contribute to the thus unsolved problem of mixing nonuniform particles and fluids. It may seem strange, but we still can't model what happens when you stir grains of different sizes and/or densities in a fluid. The act is ubiquitous in daily life (stirring your tea or coffee) and industrial processes (making solutions and suspensions), yet it's very poorly understood. I don't doubt that this phenomenon of surface tension in granular media is at least partially at play.
a questions.
Are particles at less dense areas moving faster, and could this be one of the reasons surface tension is formed? (so that in tightly clusetred areas particles are basically hitting each other as often as highspeed particles in less dense area.
is it bad that this is the most interesting part about this.
Heat-sinks aren't for "dissipating the power", they're for dissipating heat.
Just a slip-of-the-tongue, but if people don't know a lot about a subject, and are trying to learn it from you, it's best to get it right.
I just really love these guys. Do NOT retire Roger! :D 6:29
You need to water cool that PC.
I had two GTX 480s hit the mid 90s under full load, now they stay in the low 50s.
maybe that force is analogous to what is keeping time going!?
Well, water at room temperature has a huge surface tension because:
-thermal particle (molecules) movement due to the 300Kelvin temperature
-hydrogen bonds
-london dispersion(...) and other forces
Liquid helium is a superfluid on the other hand, therefore it has no surface tension, because of :
-lack of thermal (molecular/atomic) agitation
-single He atoms don't have any electric charge
-london forces are already using all of their strength to collapse the helium atoms from a gas into a liquid ...
So liquid helium has no surface tension at all !
Also, if you induce a general kind of oscillation (which compress all the particles at a period of time=T, in this case: a bunch of bronze granules), they will clump together because of the friction between them and the movement produced by you...
Try the same experiment with a cluster of glass and plastic granules - the 'surface' forces will be even stronger because you added an extra-force : the electrostatic attraction.
...physics works every time...
Honestly I heard that Kepler isn't good with computing as good as Fermi was so if anything they'd be better off with 570/580's for their intended use
True. Maybe with an interface fluid, like liquid nitrogen. Or use the helium to cool the air.
You could water cool those cards fairly cheaply, you just need to find some compatible water blocks..
good to know
Godspeed!
in what way is the form of the output signal relevant? They need Calculationpower, Not 2 ms faster transport to the eye.
I have a hypothesis/question, if there was a certain gas in the 20x20cm square, then would the "tension" be different in the experiment?
Just curious as to why with a system running that hot they don't use a liquid cooled system?
Can anyone point me in the direction of more videos done by Brady that are about computers? Perhaps Brady could do a a coding language episode (e.i. python, java etc.) on numberphile or one of his many channels if he hasn't already. I love these videos so much, and thanks.
So there's surface tension in air between a high and low pressure area?
I think this calls for a new channel on computer science!
If cooling seems to be a limiting factor, run water not fans...and your competitors could blow all that away with ASICs to hash out the values instead of GPUs, you might want to start looking into that --- physiii
That graphic would be awesome in 3D, but in 2D it's like one of those face masks which seem to switch from an in to out view.
It would be interesting to see in a vacuum...