Sharing the Load

you guys are putting so much pressure on me. There's way too much I'm having to carry the whole load of explaining what the topic is. How can we help you, Jonathan? I think we should all kind of, you know, share the responsibility. Okay, let's do that thing, the improv thing where each person Okay, so I'll start and then where's it going to? From you to to Lauren. Today's topic is about distributed computing guys. Yeah, all right, there we go that. I think that was probably the most successful

word of the the time I've ever been part of. And nobody brought in like gussy anything about nazis. Yeah, there wasn't a head in a box, which is almost always what happens in an improv scene. Wait, what did we

say distributed computing? Yes, yes, so Joe back in the day, which was a Tuesday, Uh, there was this issue where you would have a big computational problem, really really complex, and you essentially had one computer defeat it through and you really just had to wait until the computer was able to complete that problem, and you were completely reliant upon that computer and that computer's limitations. Right, that was it. That was that. So if it didn't have enough punch

card slots, you were just out of luck. Yeah, that's pretty much it. Like if you if the program you wanted to run had a requirement that was so large that you didn't have a box big enough to hold

all the punch cards. It was too complex for that machine. Yeah, which is you know also how stuff like algorithms started to be written because people said, oh, hey, if we can create a shortcut, that says you know, yeah, this is the birth of hackers, right where hackers would say, I have an outcome that I need to have happen. How do I make that happen? It doesn't have to

be pretty, It just has to work. Well, you know, in science we actually have issues like this where scientists researchers are working on uh projects where you get massive, truly massive amounts of data and then you have to figure out, well, we've got all this information, but we can't really do anything meaningful with it. Yeah, you've got a problem that's computationally intensive, meaning you you sort of know how to solve it. You have a method to get there, but you just need the ability to do

all that work. And if you're going to run it through one computer. What if you're trying to do a research project where you're sitting here in front of your computer and it gives you an estimate it says, okay, it will only take forty five years to do all the calculations on this data, right. Yeah. Yeah. It's sort of the equivalent of saying like, well, I have this room full of fun sized candy bars. I know I can count them, but I only have you know, ten

fingers and ten toes. So once I get beyond that, it's going to start getting tricky. Well, here's the other way of doing this. I just want a room full of fun sized candy bars. I'm sorry, I'm gonna come back and run with that example interruption, what kind of candy bars? Oh, I'm going to say heath crunch bars. That's acceptable. I also would have accepted butterfinger. Okay, I shouldn't really have peanuts, So that's true. Yeah, you're having the issue with the peanuts. So let's say let's say

that all we have that issue with fun heaths. Okay, Lauren needs has got a room full of fun sized heath bars, and uh. And if she were going in there on her own to count them all, and was using her fingers and toes as reference, she would very quickly run out of fingers and toes, and then she'd have to start trying to remember things. It would take her time to get through all of that. I'd have

to make a little hash marks on paper individually. Right. So, if she ended up inviting a bunch of her friends over, and she had divided up the floor into a grid and assigned each person a part of that floor where they are not allowed to eat said candy bars, merely count them, and then together you all came up and wrote down the numbers you came up with, and then added all those numbers. You would then know how many candy bars were in that room you had divided the

problem up. Yeah, take a lot less time that way. So this is kind of what distributed computing is all about. So you could do the approach of getting a supercomputer that has an incredibly fast processor, or maybe several processors, but one those are rare too. Because they're rare, their time is often spoken for. So your project will probably be competing with multiple other projects that they're very expensive

to use. That's that's number three as well. They are very expensive to get that time and so and also you know they're going to tackle that problem. They might be really fast at it, but they're going to tackle that problem sequentially. Okay, but how about instead of one really powerful computer, you buy a whole bunch of less powerful computers and then you break the task up into small chunks like we were talking about. And so you've you've got a lot of sort of powerful computers computing

little bits of your task. That's often referred to as grid computing. It's uh, grid computing and distributing computing are are sometimes interchangeable. It really depends upon who's doing the talking,

is ultimately what it comes down to. Sure, but but what if you don't have the funding to buy a grid worth of computers, or what if even a reasonably sized grid it of computers isn't going to get the problem done faster, or if you don't have space to put a grid of computers because it's all taken up by fun sized heath bars. Yeah, so clearly this is not necessarily the best option either. There could be other

options out there, right, So problem meat wasted potential. Your computer and I mean your computer, you, the listener, and the three of us. Actually yea, it was confusing me because Joe was clearly pointing at me or your well, I was pointing at my microphone, so I was pointing at the listener and Knowle's computer and everybody in the office. I mean, what if your smartphones and your tablets, etcetera. Yeah,

things that have processors. We have tons of CPUs around the world and that are not Yeah, yeah, that are so they're they're capable of doing important processing work. They can do math, they can run simulations, they can catalog data, they can do all these computational tasks these big supercomputers do on a much smaller scale, and they're not helping.

They're just sitting there when you're not using Yeah, unlike unlike our brains, which which genuinely do use the whole brain to do stuff, that ten percent rule does not apply. As we've spoken about before, your computer is really only using a small fraction of its potential processing power at any given time, and especially when you're, for example, away

from home. Right. So here we now reach the idea of the popular distributed computing method also known as volunteer computing, where there's a project that says, hey, I wonder if we can get a bunch of people out in the world to donate some of their computer power to help us find a solution to our problem right now. Clearly, this only really works if the problems you are tackling can be divided, right. Not all computer problems are that way.

So there. We kind of talked about this with quantum computers as well, which sort of take this concept and boil it down to a single machine. It's the fact that the cubits can act as either zeros or ones and all values in between technically that allow them to to do all the parallel processing simultaneously. At least, that's

the concept. This is sort of the same thing, except instead of the conceptual doing them all simultaneously on one machine, you've divided that problem up across a network of machines, all of which are working on parts of it. If the problem can't be divided, then grid computing or distributed computing isn't going to work. You have to have that sequential approach. Think of it as if I have a a math problem that's completely distinct, then I can solve

that uh. And if I have a whole series of math problems, each of which are distinct, I can solve them in whatever order I want. Right. If problem one seems like it's too tricky, I can skip the problem to and solve that one, then go back to problem one. But if I have a series of math problems, each of which which is dependent upon the answer of the previous problem, that's more of an issue where I need to take a more sequential approach. So it all depends

on the nature of the problem. But assuming that it is one that can be divided up, this model is a fantastic solution, and it's really amazing how people have started to leverage it. Right, and so this is ongoing today. But the concept is not at all new it has been. The idea has existed for decades, and it's been employed in very powerful ways for at least like ten or

fifteen years now. But I think we should go back to the beginning of the idea of distributed or grid computing and take a look at how we got where we are today. Okay, all right, So if you go to the very beginning, we're talking about the basic, major, huge mainframe computers that didn't have any connection with any other computational device. So they were completely standalone. And if you wanted to divide up work, well you might want a lot of things, but here's the way life works.

It's not gonna all pan out for you. So that was the big and then you had the the early days of networking computers together, the first major one being our Pannette. Now that was in the sixties and seventies really, So in the sixties you get this team together that DARPA had wanted to get together, um, and they started trying to figure out ways to let computers talk to one another. Now that early communication didn't really facilitate distributed computing.

You could share work, but you couldn't divide a problem up at that point. There was no infrastructure to allow that sort of thing to happen yet, So think of it as people who are able to communicate with one another, but they can't collaborate actively on things. Yet you get to the nineteen seventy one you get the first distributed computing problem, which was not a true distributed computing challenge.

It was actually the first computer virus purposeful virus though I'm well, I mean, I guess all viruses areful viruses. It was an experiment. Really yeah, it's called Creeper. It's a self copying, self duplicating piece of code, which you can realize if you talk about self duplication, it could quickly get out of control and duplicate self in order to take up all remaining space. There was also Reaper, which was meant to diffuse and disarm and remove Creeper.

Then you in the nineteen eighties you get to the birth of the Internet. So this is the true network of networks. Urban net is kind of the predecessor to the Internet. It's not really the um the earlier version of it. Internet itself was a different thing, right, But this is when you get to the kind of net that Sandra bullet gets caught up right exactly. It was

very accurately portrayed in that movie. As I recall, Uh, well, distributed computing ends up becoming an area of research, even though no one has actually launched a distributed computing project the way we think of them today. All Right, Well, at this time, it was only government officials and university researchers, people on that level who had access to to this network, right right, You and I, assuming we were not employed by a government or military or university, didn't have access

to it. Uh. The Internet was something that probably most of us were largely unaware of at the time. It was it was still so um so expensive to obtain the kind of equipment that you would need that it was. It was really left up to these large organizations. Most of us were dealing with local networks like like Bolton board systems and stuff that had some connectivity but weren't

ultimately leading to the larger Internet. So by nineteen eighty two there was a Distributed Principles of Distributed Computing symposium, and then another one in International Symposium in nineteen five, So people were already thinking about the potential for computers to do this kind of thing and starting to lay the groundwork for the technology that would make it possible. By night we get the first Internet based distributing computing project at the d e C System Research Center. And

the way it worked is through email. We get a task through email. So you would volunteer to be part of this? Yeah, well, I mean there wasn't an automated system here yet. Yeah, so you will? You would, you would first, you would volunteer. You'd see like, you know, you go to the coffee shop. You see a little piece of paper with strips on it saying, hey, are you willing to lend your computer to help science? Tear

off a strip and send an email? And uh, and most people in the coffee shop and say, what's an email? But you knowing what an email is, You tear off a strip and you send off an email. They send you an email with a task to complete. You run it on your computer, complete the task, and you email the results back. So it's an It takes takes an active role for the volunteer in order for this system

to work. And the purpose of this project was to factor large numbers, which is an important role part of cryptography, computer cryptography and UM It was a relatively modest project by There were about a hundred volunteers using this and it was fueled in part by challenges issued by the r S, a security incorporated company. H They wanted to do lots of research on factoring of large numbers as a security company, and and the fact that this has

a lot to do with cryptography. It was very important to their business in order for them to make sure that their systems and other systems remain ahead of hackers. So if it turns out that a group of people can quickly factor a certain size prime number, or to find the largest primes of a certain number, then they wanted to make sure they could go to the next level in order to avoid issues with people immediately hacking through all sorts of security system Yeah. Yeah, greater encryption,

for greater protection against people who wanted decrypt YEP. And in NT you get distributed dot net, which is really the first project to use the Internet to distribute data for calculation and collect the results automatically, so it did not require the volunteers to receive and send email. It had a software package that did all the handling of

that automatically. Yeah, it's all in the background. So the user just had to install software onto his or her computer and then start up the software, have an Internet connection, and the program would would run itself. Yeah, essentially running in the background, taking you know, any unused cycles of processor and dedicating it to working on the problem. And that has pretty much been the model for distributed computing

or volunteer computing ever since. You know, I thought you were going to say, the first distributed computing project was like an x files used net group that was sending out task to get different people to decode the messages and learn Molder secret. I wish. I'm sure some of that was going on. I think, I mean not Molder secret, the secret of his his his family has passed. I think I think the real secret was that everybody but Molder and Scully knew. Everybody everyone in that show knew

more than the spoiler. Look, I think I think that's not a spoiler. That's saving you some time. Okay, okay, Well here we are in the present day. Yes, what are some of the big distributed computing projects or volunteer computing projects out there today? And how are they actually making a difference? Like is this just kind of a lark. You know, Oh, that's kind of fun, you can participate,

or does this actually impact science in a meaningful way? Well, let me answer that second question first, because it's certainly impacts it in a meaningful way. You know, like we were saying, access to supercomputers and grid computers, that's something that if you're incredibly well funded, might not be an issue you. But there are a lot of science projects out there that are doing amazing science on very limited budgets.

And so for this kind of model to come along for specific types of problems, particularly science projects where you're generating huge amounts of data and then you have to analyze it, it's an incredibly useful tool. And it also not only is it meaningful in that sense, it's meaningful in the sense it gets people involved in science, even

in a passive way. You start to learn more about what your computer is being used for one the purpose of that scientific project is, and that encourages scientific literacy, which I think is incredibly valuable. Yeah, so it's partially outrage. It's it's not just getting stuff done. It's also saying, hey, you can be a part of this, and also look at this cool thing that we're learning about Yeah, like, you know, the universe is mysterious. You want to help

but be less mysterious. Yeah, you know, it's kind of kind of the way it boils down. And a lot of these projects are all folded under the umbrella of the Berkeley Open Infrastructure for Network Computing, which has a great acronym. You want to say it, Joe, you know, you want to say it bow wink, blink blink. It doesn't actually it's not. The full thing isn't blink with a C, but it is blink that has a C at the end of it. Um. But it's just fun to say, and so I'll probably try and say it

as frequently as I possibly can. Well, this is sort of an umbrella program. Yeah. Yeah. It actually started as a specific platform for one of the programs will talk about in the second But what happened was it was such a useful approach that got adopted for other programs and now it serves as this umbrella where you can go to the Blink website and see all sorts of

different projects listed underneath. Yeah, there's a few of these kind of infrastructures out there, and and this one is really cool because it's definitely run by by Berkeley, who are good people and and don't really have much of a profit factor in it. This is more by night. It's it's definitely more about enabling people to do science right and less about we've made this incredible platform that you can use for one low price of five million

dollars or whatever. UH. So it's it's really cool. I blogged onto it the day that we're recording this podcast November four, two thousand and fourteen, and saw that there were about forty projects listed under Blink. I know that there that changes. There are times when there are many more uh and some of them actually were offline when I looked into them, probably for actually maintenance for for

several of them, because you're talking about a fairly complex system. Ultimately, the way Blink works is that you've got a master system that kind of determines what the job is, sends the various tasks out to the volunteer computers, uh, accepts the incoming tasks from computers that have completed whatever work you've given them, and then assimilates that into a meaningful way for researchers to look at. So all of this

is going on behind the scenes. And meanwhile the volunteers computers are doing actual scientific work without really, you know, without necessarily you being aware of it. Some of them come with fun like screen savers, so you can kind of watch a graphical representation of what's going on with your, you know, the work you're doing. One of the ones we'll talk about has one of those. That's pretty cool. Well, let's get to the examples. Sure, so here's a fun one,

steady at home. The search for extraterrestrial intelligence. Wait a minute, how can I search for extraterrestrial intelligence without a radio telescope. Well here's how, Joe. You can take data that was gathered by other radio telescopes and use your computers processor to analyze that information and look for anything that might

be meaningful. Right, Because the thing is is that when you know these telescopes are searching space, there's no like big red light that starts flashing when when something, oh, alien, it's right there, it's right there. We just found the Venusian elvis. No, I guess what we end up with is a lot of numbers, a whole lot of numbers that need to be crunched, and so your computer can crunch part of them. And so it wouldn't have a red light, but your computer might be the one that says, hey,

here is an anomaly in the data. Right, wonder what that means? Right, And that's it really is, looking for those kind of anomalies and trying to flag them so that other people can take a closer look and see if there's actually any signal inside the noise. Because as we know, there's lots of stuff out there in the universe that generates radio signals. So just because you pick up yeah, yeah, you might get up stuff from a pulsar,

which is already super cool. The fact that you know your your computer is working on data generated by an extrasolar body that could be light years across from us. But obviously that's not the same as Venusian Elvis, which

is what our real goal is here for CET. But that's the thing is that with all that kind of information, it's the same sort of thing that if you were to have a massive telescope look deep into space and slowly do uh pans across the night sky looking at every single teeny tiny bit of that space to look

for meaningful things, would take forever. Same sort of thing. Um, it's really dividing up this massive task among lots of different computers to make it more manageable, because otherwise you would just keep gathering data and it would pile up, and you know, you would always be playing catch up and always falling further and further behind. So that's the

purpose of SETI at home. Then you have other projects too, for other types of science, like like astronomy is a big one, but it's by no means the only one, right, And it's not always just that you will be analyzing data that comes in from some kind of collector. You might be running emulations of some kind or processing data from simulation, right, one of those being Atlas at Home, And Atlas is one of the big projects, one of

the scientific research projects connected to the Large Hadron Collider. Now, is that the one that shrugged that I keep hearing about. No, no, uh, John Galt has nothing to do with it either, so uh well, very little to do with it at any rate, you know, I mean, who is John Galt really really? So No, the the Atlas at Home project, you're just staring daggers at me, Joe, You're giving me this superior grand I don't know what it means it's not superior at all. It's not superior at all. It's just it's

just kind of shame. Actually, it's a shame grin. Now, Atlas at home is all about running these U simulations that you were talking about. So, uh, what Atlas is looking for. It's looking for the the outcomes of proton collisions, which is what happens at the Large Hadron Collider. Right. You're you're lighting streams of protons that near the speed of light and then looking to see what happens. And part of that is a search for lots of stuff that we just don't know if it's going to pan

out or not. You know, the boson was one of those things, and then it turned out that we found it. But other things include extra dimensions of space, or dark matter or the unification of the various fundamental forces. Yeah, you know we're talking about like the big questions about the universe from a physics perspective, right, well, from any perspective. I mean, these are like the most fundamental unsolved questions about reality that exists. Yeah, or at least the ones

that we know of. Who knows what other questions we're going to find in the future, which is also really explaining. Right, so you know what's your laptop doing right now? Right now, it's telling me that my battery is dying. But no, the the the idea is that we would use the Atlas at Home project. In fact, this is this is what's happening that to go through pedo bytes of data.

A petabyte is one thousand terabytes. Uh So, in order for you to get a handle on how big that is, first of all, I remember when a kilobyte was a lot because home hold. But kilobite is a thousand bits. Then you've got megabyte which is a million, gigabyte is a billion, terabyte is a trillion, and then peta bite which is one quadrillion. So it needs to go through

all this huge amount of data. The computers attached to Atlas at Home are using a computer program that run the simulations of the creation and decay of supersymmetric bosons and vermons, and it then uh sends that information back to the Atlas project. Now, the reason for this, the reason for running all these simulations is to look at which scenarios are the most likely to really happen in real life under the conditions of the pot proton collisions

and the large Hadron collider. So you look at the ones that look the most likely based upon these computer simulations, and if you find any evidence of a similar uh reaction after a real proton collision, then that shows that you might be on the right track, right, You might actually have seen something that could be lead to evidence to something like dark matter, which to this day we

think exists. It's kind of a kind of a placeholder name. Really, it's the to fill up the matter that must be out there for our our vision of the universe to be tracked. We observe its effects, but we can't detect it directly right exactly. So this could lead to evidence that would give us more of an idea of what dark matter actually is. And that's just one thing that

Allison Home could help lead to. Now, it's not the type of program that you're super old computer is going to be able to run in the background, because the simulations actually do require a bit of processing power. So they have some well some of these well, I mean, we're simulating particle collisions, y'all. Yeah, but some of the some of these programs require just you know, some unused cycles of a CPU and that's it, right, It doesn't.

It doesn't have to be particularly powerful. It may mean that you are a computer running this This program might take longer to solve a certain task than another person's computer, but it would still work. With the case of Atlas at home, there are actual system requirements. You need to have a sixty four bit computer with at least four gigabytes of memory, which is not unusual today. If you went out to buy a computer, more than likely fil

fill those requirements. But if you're talking about this desktop computer that's just collecting dust, and you think, oh, I can dedicate this to science, it's not. No, it'll you know, maybe you'll be able to play some some some wicked games from the nineteen nineties, but that's about all you're gonna manage to do as opposed to helping figure out whether or not dark matter exists. Um So, anyway, it's it's a neat example. And then we come to a

couple of brothers. Right. Yeah. One that I wanted to talk about is the Clean Energy Project, which is through IBM's World Community Grid, which is another one of these kind of infrastructures that allows different projects to come in and use this. Uh, you know, network of networks, so sort of sort of this similar role as boink right right, UM and and the Clean Energy Energy Project is looking for cheaper, more efficient, greener, more more flexible solar cell materials.

But we've talked a lot on the show previously about the issues with solar cells and what makes them so clunky and expensive and difficult to manufacture UM, so the research that they're doing could be huge for the future of energy. UM. The project itself has been around since about two thousand four, and they teamed up with Harvard University to create this distributed UH computing program around UH twenty ten or eleven, which is when phase one kicked off.

It just completed recently, and and during it, volunteers helped to sift through a database of over two million compounds that they were mathematically testing how these these molecules joined together to form solids, and then predicting whether those solids could have the right electrical properties to be useful as components and solar cells. UH. They isolated some thirty six thousand compounds that could be able to to double the

current efficiency of your average solar cell so UM. According to the project website, we can put the work of this in perspective by saying that it would have taken a single average PC some seventeen thousand years to have done this work. The Volunteer Computers joint effort took about three seventeen thousand years. Yeah, I think we might have different problems by then some time. That's quite some time.

It's a chunk. It's a chunk, um And the project has now moved on to phase two, in which there phase two. I know it does sound fair. We're moved on to final processing. I hear phase two. I just think, now what we we figured out the most efficient means for solar panels now the world domination. Yeah, it's very very pinky in the brain. Um but but but so, these these compounds, these thirty six thousand some compounds that they've identified UM as being potentially useful, are going to

be explored more thoroughly. You know, all of their physical and electrochemical properties from from optical ability to squish ability are are going to be identified. Like going down as squish ability is important and material science it's the most it's the most important, um you know, going down to the quantum level of what's going on with these things.

So the idea here is going to be to put together a database containing everything about these compounds and to provide um, you know, not only that, but also direct input to various researchers and inventors who are working on improving solar cell design. Cool. That's really awesome. Yeah, well, Joe, why don't you tell me a little bit of out proteins? Why do you ask, Jonathan, Because I'm pretty sure you're the one who wrote the notes on it, you'd be

right now. I'm gonna talk about proteins because one of the most interesting projects in distributed computing I think is called folding at Home. And you might have heard of this before because it's been around for years. This has been operating since two thousand. Yeah. Yeah, it's still going strong. And I think I think we've talked a little bit about protein folding before on the show. It was either on this her tech stuff. They'll get a little scrambled sometimes.

I think we've talked about folded before. I don't know. We did to fold it, which was the actual computer user program where you would video in order to to self self work out these problems self workout. That was so good grammar man, I need a good self work out. But yes, so proteins. Our cells are kind of made up of them. They do stuff in our bodies, right, Yeah, well,

proteins are what make your body interesting. Well, what makes you more interesting than a rock or a mud puddle there in general, not like my body in particular, your body in particular, but also everybody else's body. They're they're they're sort of the animal workhorses that they do all the molecular work in your body. You've often heard of d n A being referred to as like plans or blueprints. The blueprints are for making proteins. They make amino acids,

which chain up into proteins. So a protein is a huge long chain of amino acids in a particular sequence. So think of like a literal chain, except each link in the chain is of a different type that has different properties. Now, imagine you're building a machine, maybe a gigantic factory assembly line for making big mouth billy bass or terrible person for making heath bars. There you go, exact, bring it all back, but all you have to work with are these chains, and chains don't do much good

as machine parts. It sounds like a floppy machine. But have you ever taken a chain and twisted it and keep twisting it kind of curls up, makes these these

very like eventually uh immobile kind of shape. Yeah. And so if you can take a chain and twist it up into a tighter, rigid shape, you might be able to take those tight rigid shapes that are different because of the different amino acids and the sequences in the different ways that they twist up when you when you press them together, they might fit together like gears, or like wires and sockets, or like wheels and levers. And

this is sort of what happens in your body. For protein chains inside your body to do their job, they need to have the right shape, and to assume the right shape, they fold up. This is called protein folding. Um. Now, most of the time a protein folds up without a problem, right, It just goes on to do its important role and whatever that might be. It could be making up a body tissue like an artery or a strand of hair or a little piece of muscle, or it might become

an enzyme or who knows what I mean. They do call all kinds of stuff in your body. Yeah, these are these are the basic building blocks of like Joe said, of what make you electrically and chemically you right, But there's a downside because sometimes, in rare cases, a protein fails to fold correctly and the chain does not twist up into the correct shape for doing its job, and so a misfolded protein can be a really bad thing

in the body. A lot of really bad diseases are now believed to be caused by either the accumulations or the effects of misfolded proteins. So the examples given by the Folding at Home project are Alzheimer's disease, cystic fibrosis uh b SC, also known as mad cow disease. Uh They cite an inherited form of emphysema and some cancers.

So obviously, in order to cure these diseases, we need to know more about protein folding and about misfolding in particular, and unfortunately, there's a lot we don't know about how protein folding works because obviously, I mean, it's so tiny and it happened so fast you can't just like film it and run the game footage. Yeah, yeah, And there's so many potential combinations that they can they can mess up in a whole bunch of different exciting ways, exactly right.

So how are you going to solve this problem? Well, for more than a decade now, we've been trying to learn more about the whole process of protein folding by running computer simulations on this distributed network. Now, there are different kinds of simulations you can run about protein folding. You could just try to look at what the amino acid chain is and then predict the final shape based on that. So that's one kind of simulation, but that's

not what Folding at Home does. Folding at Home was designed specifically to study the whole process of folding, focusing on the intermediate states. So what's happening as this chain is curling up on it? So because they think that those intermediate states might be the source of the problem

with misfolding based diseases. This is really I mean, I find the whole concept of protein folding so fascinating because if you look at it with just a couple of amino acids like a small string, you could kind of conceptualize how those rules all work together, that certain sequences are going to fold in very specific ways. It's just

a set of rules. But then as you increase the length of that chain, it becomes more and more complicated, which rules end up being the ones that take priority, which ones are going to happen first, And then by the time you get to an actual protein length chain, it is almost unfathomably difficult for us to conceive this. This is truly what we were saying earlier, computationally intensive

to simulate, especially because of the time length involved. So from our point of view that they point out, and I think it's a good point to make from our point of view, a protein folds almost immediately, you might take a millisecond or a microsecond. But to run this simulation simulating every single force at the molecular level, that that that long of a transformation takes forever to compute. You know, it might take a forever just to do a few nanoseconds of computation. And so this is a

really serious computer problem. But they invited people to help out. In two thousand they launched this project, and since then it has really come a long way. In the past fourteen years, more than a hundred research papers have been published based on the simulations run by folding at home, and it's gotten obviously over time because computer, you know, because it has expanded in because computers have gotten faster, it's become a lot more powerful as a tool. And

it's also been used in combination with other similar tools. So, like those other types of simulations I was talking about the the in sequence predictors, you can sort of pair these two different approaches together to get some interesting information. Or you can pair the simulations done on the distributed network with other types of simulations done on supercomputers. And so this really is really important and and useful knowledge in medical science that could cure these things. I just

checked the stats on it today. Uh. They said around two pm on November fourteen that today we are one hundred and sixty three thousand, nine d and fifty eight computers strong, outputting thirty eight thousand, two hundred and twenty two TARA flops of computing power. That's a huge amount of computing. Yeah. Uh, it's it's pretty awesome. And flops is of floating point operations per second. So uh, and you're actually installed this on your computer. Oh yeah, I

was right. This is one of the coolest things about it. Just from a user perspective, they give you a little readout of whatever project your CPU cycles are currently contributing to. So mine is right now, it's my computer in the background is working on project nine zero zero eight, which is targeted at Alzheimer's and it's it's studying some of the natural features of Brian Staton, which it's got a whole description I think is a little too complex to

try to explain here. I don't really fully understand it, but uh, it's really cool that you can what would otherwise just be wasted potential. Just my computer sitting here while I'm talking into a microphone is actually contributing to something that could literally save lives in the future. Now, does your version of Folding at Home? Does that have the screen saver where it shows the the graphical depiction of the protein. I don't know. I haven't poked around

it on it enough. See when I when I installed Folding at Home on a work computer many years ago, that was the default, Like whenever my computer would go into you know, the low energy mode or whatever, that was the screen saver. So it actually showed the protein folding a representation of the protein folding as my computer worked on it. And uh, I remember it changing colors as well. To indicate sections that had completed versus ones

that were still working. But that was an early version of Folding at Home, and it may very well be that the current ones are very different because because just like any other software, this is software that gets uh patches, gets something. Yeah, it's continually updated, especially as the research moves forward. Now there are tons of other ones we

could talk about. Some of the other ones that are in boink include Climate Prediction dot net, which, as he would guess, studies the climate, Cosmology at home, which studies astronomy and cosmology, Constellation which is aerospace engineering, malaria control dot net which is it all about epidemiology, mind modeling at home, which was not for people with very attractive brains. I want to get into the magazine. Yeah, I thought it was that too, like you just see brains going

down the catwalk. But as it turns out, it's all about cognitive science. So I mean, still pretty cool, I guess. Yeah, there's skull. Yeah, I think we've got an idea for a reality show. Guys. Let's let's try and keep that on the d L till we can develop a full pitch. But now RNA World, which is all about molecular biology H and Quick Catcher Network which is all about seismology.

And like I said, when I looked, there were around forty projects, So this is just a it's really just to show you that they go across all sorts of areas of science. Also, I shouldn't mention, and I didn't mention it earlier, that they use lots of different platforms. So depending on the project, you might be able to run this on a PC, a MAC, tablet, UH, of smartphone UH. Some are specifically designed so that specific types

of graphic processing units can work on it. GPUs are fantastic because they are multi core processors, so they tend to be able to work on problems and divide those up into smaller sections that each core can work on independently. Who knew that your proclivity for violent video games could one day save the world. I certainly didn't. I did

not know that until you know fairly recently. But if you would like to get involved in any of these, you can visit blink by going to b O I n C dot Berkeley dot E d U and and get the hook up from there, or you could search for Boink on your your local app store of choice. I I believe that if you search it an Android or I I Apple, I stuff, yeah that thing, you

can come up with something. Um. Also, if you were interested in checking out World Community Grid, the four projects that they're featuring right now, um, well, so you've got the Cleaner Energy project. There's also geno mysteries, cancer and AIDS research up up up for for looking at, and so if you want to, if you want to check them out, you can go to World Community Grid dot org.

So yeah, lots of ways to get involved, and it's so easy, right you Just all you have to do is install a little bit of software and whenever your computer is going to be idle and still connected to the network, to the Internet, you'll be able to contribute to these scientific projects. So if there's something in particular that that tickles your fancy and you think I want

to be part of this, I want to contribute. And it may seem small, but it is significant, and it's definitely more significant than you saying that's a neat project but not doing anything. So there's that I recommend it. I haven't noticed really any performance difference on my computers since running these are these are designed so that they

are only supposed to take up the idle cycles. Like, so if you are actually actively using your computer, it's supposed to fade into the background and not take up the resources. Yeah, so if you have an older computer where you're having some sluggish issues already, that's probably not going to be ideal because it's just that suggests that there's probably some things that need to be cleared up on your machine, just because the various processes that are running.

But if you have a you know, a decent running computer and you think I wanna, you know, do my part and help science, and especially since it's going to be doing stuff when I'm not doing anything else, I recommend it too. I mean, it's a great way to get involved, so really exciting stuff. And who knows what we could see this used for in the future. Well, yeah, one of the things I was thinking about was just

pairing this with the continuation of Morse law. Sure, we don't know how long MOR's Law is going to continue, but that's the thing we've talked about plenty of times on here, that computer processing power multiplies at a pretty predictable rate. It's continually getting better, and we haven't seen it stop yet. So the faster computers get, the more and more you can do with a hundred thousand of them, or two hundred thousand of them, or or I don't know how many thousands of them you can you can

talk people into sharing their their computer power with you. Sure, and then also with multi core processors being able to divide problems up into smaller problems. As long as we continue that pathway where we're able to do that effectively, than this kind of approach could be useful in all areas of science. There's not there's not a specific one that it would be better for. I mean, anything that has huge amounts of data that needs to be analyzed is ripe for this kind of thing. So it's really

exciting stuff. I'm glad that we covered this topic. And Joe, this was something that you had suggested, and so it was a great suggestion. Um, some of you guys out there have had some great suggestions too, and I think you should continue to suggest great things and the best way to do it is to get in touch with us and let us know what those great things are. Because of the as we have to guess. So if you want to send us a message, do so at our email address that's FW thinking at how Stuff Works

dot com. Or drop us a line on Facebook, Twitter, or Google Plus. Our Twitter and Google Plus handle is FW thinking. Just search FW thinking in the Facebook search bar will pop right up. Let us know what you want us to talk about. If there's some topic that you want to hear more about about how that's going to be in the future. Tell us ask us your silliest questions. We read all of them, we really enjoy them,

and we will talk to you again really soon. For more on this topic in the future of technology, visit forward thinking dot com, brought to you by Toyota. Let's go Places,