Which is still hard to program for anything but the most obviously data parallel applications
Unless you're suggesting we make it even lower level and have the compiler map individual instructions/basic blocks to processing elements.
In which case it's dataflow processing and has it's own equally large set of problems, most notably....that it's really hard to program in a way that maps well to the dataflow hardware.
The summary is rather skimpy on details, but a 5.2 is, all things considered, fairly significant for the region, even if it would be a blip on the map on the coast. I've been in the midwest (ohio and illinois) my whole life and this is the first quake I've ever actually felt.
I have no idea what it has to do with cognative dissonance. But the reason it's better to switch is the following:
You have a 1/3rd probability of choosing the car initially and a 2/3rds probability of choosing the goat. If you do not switch, you have a 1/3rds probability of having the car. After one of the other doors has been revealed to be a goat, however, the following is true: If you originally picked a car, you will get a goat. If you originally picked a goat you will get a car. Since you had a 1/3rd chance of originally picking the car and a 2/3rds chance of originally picking a goat, if you switch you end up with a 1/3rds chance of getting a goat and a 2/3rds chance of getting a car.
If you still don't understand, do this: Write out every possible combination of two goats and one car. CGG, GCG, and GGC. Now say you pick the first door. Elimiate one of the goats other than door 1. You are left with CG, GC, and GC. Notice that two of the three possible combinations have the car behind door 2, not door 1.
I really don't envision a computer creating perfect "backing tracks" in real-time to any vocals sung into it. It's more like, it'll sometimes/often make "passable" ones, fun for karaoke or practicing -- but not worthy of recording.
Given that pop music is already arguably not worthy of recording, I'm not exactly sure that there's any impediment to this being used for pop music.
The LHC is just a tourist trap. It's like Times Square. You go thinking it's gonna be all fun, then you realize it's just a bunch of bright colors and earth eating black holes, and there's nothing to do there but shop for overpriced bosons you could've picked up for half price at a more nondescript collider.
I'd be much less repulsed by this idea if I had any belief that the fees would be distributed in a fair fashion. As someone who listens to a fair deal of indie stuff (and virtually no major label stuff), I'm concerned that there's no way in hell anyone not on a major label would get to see a dime of the money. (Not that anyone ON a major label will see a dime of the money either, what with the soul-stealing contracts they make bands sign).
Ultimately then, it becomes about subsidizing an industry in a manner that provides absolutely no incentive for any major label to make desirable music. They can produce whatever they want and take the flat fee, preventing us from voting with our wallet. As a result, music would become even more controlled by the major labels than it already is now. And that's a particularly disgusting thought.
2) A significant number of applications can and do run on 1000+ cores. Sure, most are scientific apps rather than consumer apps, but there is a market for it nevertheless. Go tell a high performance computing guy that there's no need for 1k cores on a single chip and watch him collapse laughing at you.
There are some possibilities to use quanta (?) as signal carriers, but no encryption is involved. The theory is that if you wiretap such a signal, then the original receiver will find out. So it could maybe be called "Quantum Wiretap Detection" or the like. But since this is a physical thing that relies on theoretical models that are typically not exact, it is not actually known whether this is really secure. I seem to remember that there are actually possibilities to liesten in, found in te last few months.
The reason for the encryption in the name is that the idea is to exchange a private key over the secure (but very slow) channel, which will then enable encryption over an insecure channel. So you're correct that the name is misleading. To be more accurate, it should be called quantum key exchange, not quantum encryption.
I don't see it happening in the near-future, but perhaps near the end of my life-time (I'm 20-something). And it won't be like the first computer revolution, with guys in their garages and basements screwing around with computer hardware. The first quantum computers will be only really useful for large Monte-Carlo projects (like the Earth Simulator) that require tons of computing power.
Quantum computing is nigh worthless for Monte-Carlo. Yes, you can simulate a ton of inputs and get a ton of outputs in one run, but it all collapses into one waveform in the end anyway. Throw in the fact that Monte-Carlo simulations are classified as "embarrassingly" parallel and Monte-Carlo is the last thing you'll see on quantum computing.
The problem then becomes building a quantum computer that is faster than the supercomputers of the time. The first quantum computer prototype won't just start out as a powerhouse. After we get the first quantum computer working, it may be up to a decade before we see one actually being used.
The entire notion of faster or slower is thrown out the window with quantum computing. The power of a quantum computer is not limited by its speed, but the number of qubits. Furthermore, the first quantum computer prototype already exists. Indeed it is far from a powerhouse; it was used to factor the number 15. If we could expand the number of qubits arbitrarily we would have functional laboratory quantum computers, but it's our inability to increase the number of qubits because of decoherence and other physical limitations that prevents us from having useful quantum computers.
it seems painfully inefficient to 'redo' stuff that doesn't seem to be wrong just because a cosmic ray was detected.
1) The likelihood of a cosmic ray is ridiculously small. So small in fact that the cost of rewinding progress when they are detected would be completely unnoticeable.
2) We *do* have the ability to package CPUs such that they are protected by CPUs. The problem is that the packages are so large and expensive that no one would buy them given the current probability of soft errors.
So the solution is most definitely NOT to stop shrinking transistors. Even in 10 process technology generations, the mean time to a soft error actually affecting a bit on a CPU is something like 1 million hours. Never mind whether or not that particular soft error is critical.
You have a misconception about The Plague. killed two million per year at the most. And you'd have to count all victims in all three outbreaks centuries apart to reach the total of 137 million.
2 million a year is kind of a big deal when it comprises THREE TO FOUR PERCENT of the european population at the time. It would be equivalent to almost 22 million people dying in europe per year today.
You can't speed up the server or the network you are downloading from. They can claim it can do Gig speeds but that doesn't help with anything outside of Comcast networks.
Err...what kind of servers are you downloading from that use a home user connection? When you download something most of the lines the information passes through have far greater capacity than the user until you reach close to the home user as it's the "last mile" connection that most greatly limits bandwidth.
Or how about that person who gets Comcast 160Mb down only to have it run into a 10/100 Ethernet Card.
They spend $10-$20 to buy a gigabit ethernet adapter.
Not only has the problem of reliability not been solved, it is the primary barrier to effective quantum calculations. We simply cannot keep qubits in an entangled superpositioned state for long enough to perform calculations. And you are correct, every qubit you add decreases the time to decoherence, so as you increase performance you decrease reliability. This is however less a problem of computational time (i.e. requiring multiple runs of the algorithm) and more a problem of being able to get any answer at all. In many cases a solution is extremely easy to check.
Well if your only requirement for somethings to be "Moore's Law"-ey is that it follows an exponential curve, then there are thousands of things that follow Moore's Law. Just as there are plenty of things, even computer related, that do not follow, such as memory latency.
I guess I just have a general aversion to using the terminology of Moore's Law instead of just saying exponential growth. Moore's Law is a very specific observation on the cost effectiveness of transistor count per chip (not density specifically) where low transistor counts suffer from higher cost per transistor and high transistor counts suffer from low yield, and how that sweet spot of transistor count is increasing exponentially with time. Moore's Law is fundamentally an economic limit, not a fundamental physical or technical limit.
What they're referring to is quantum entanglement. There's a vast difference between having two independent qubits each in a superposition of two states and having two entangled qubits together in a superposition of four states. It's a subtle, but very important difference.
Probably not. Moore's law is a very specific and narrow-focused trend regarding the number of transistors that can be placed on a chip to achieve the highest cost effectiveness. Any major shift in technology (i.e. away from traditional transistors) will have it's own completely different trend. There's no inherent trait in computing as a whole that gives rise to Moore's law.
You can use ANY quantum mechanical system with two discrete states as a qubit, just as you can use any classical mechanical/electric system with two discrete states as a bit.
Typically with photons, it consists of the direction of polarization of the electro-magnetic field associated with the photon. Straight up and down represents one state, horizontal represents the the second state, and the photon can be in a superposition of both of these states.
Saying that photons get "destroyed" is irrelevant so long as we can measure the photon's polarization when it gets destroyed because as soon as we measure the polarization, the quantum state of the photon is destroyed anyway and becomes worthless to us. This is true of any quantum mechanical system, so whether the system representing the qubit sticks around or disappears after being measured (whether a photon, electron spin, or otherwise), is only a matter of logistics of the quantum computer, not of the actual computation.
Why would the velocity matter at all? It's the acceleration that is uncomfortable, but at the acceleration planes normally undergo 10,000 mph is easily reachable in a few minutes of constant acceleration.
What a lot of people apparently don't realize is that those people opposed to embryonic stem cell research are also opposed to the creation of embryos for fertility purposes. Claiming that they would be discarded anyway and thus should be used for scientific purposes begs the real and insufficiently addressed ethical question of whether or not they should have even been created in the first place.
There's two big draws I think. First off, look at the success of serialized shows like Heroes and Lost. Shows with ongoing plot lines, rather than completely episodic sit-coms and dramas like CSI. Rather than being the exception, shows with a single overarching plot line planned from the start of the series (or even earlier in the case of an anime based on a manga) are the norm in most genres. So you can have development, a real crisis, and a conclusion in 13 or 26 episodes. Compared to most american shows whose primary goal is to stay on the air as long as possible, anime provides a better storytelling experience.
Secondly, animated shows can tackle any subject matter. You don't need block buster CGI effects since everything is animated anyway. So anime shows can feature sci-fi, fantasy, or ridiculous action themes much more easily than an american tv show can.
There are some people who like it because it's Japanese and exotic and weird, but all in all I don't think that's the primary reason. It's simply that the animated medium allows more flexibility and creativity than live action, but is stigmatized in america as being childish.
Not that I don't agree with your sentiment, but man in the middle attacks are incredibly difficult to defend against without a secondary trusted channel, which is simply not feasible for a lot of internet traffic.
Wow released an expansion in January. Guess when the next expansion is coming out. Early next 2008. They've also said that they plan on one a year or so from now on. They added a lot of content between release and BC. How much have they added since BC came out? One 40 man instance that was supposed to be out at the same time as the xpac itself that less than 2% of the WoW population will ever even see.
So in conclusion, ditch the attitude if you can't even get your own fucking facts straight.
(Oh, and btw, phase and toxic grenades no longer bounce. One sticks to walls and the other explodes on contact. So not only is your entire complaint of gameplay on one skill, it's one they're tweaking because...SURPRISE SURPRISE ITS BETA)
Slashdot Mirror
So it's stream processing.
Which is still hard to program for anything but the most obviously data parallel applications
Unless you're suggesting we make it even lower level and have the compiler map individual instructions/basic blocks to processing elements.
In which case it's dataflow processing and has it's own equally large set of problems, most notably....that it's really hard to program in a way that maps well to the dataflow hardware.
The summary is rather skimpy on details, but a 5.2 is, all things considered, fairly significant for the region, even if it would be a blip on the map on the coast. I've been in the midwest (ohio and illinois) my whole life and this is the first quake I've ever actually felt.
I have no idea what it has to do with cognative dissonance. But the reason it's better to switch is the following:
You have a 1/3rd probability of choosing the car initially and a 2/3rds probability of choosing the goat. If you do not switch, you have a 1/3rds probability of having the car. After one of the other doors has been revealed to be a goat, however, the following is true: If you originally picked a car, you will get a goat. If you originally picked a goat you will get a car. Since you had a 1/3rd chance of originally picking the car and a 2/3rds chance of originally picking a goat, if you switch you end up with a 1/3rds chance of getting a goat and a 2/3rds chance of getting a car.
If you still don't understand, do this: Write out every possible combination of two goats and one car. CGG, GCG, and GGC. Now say you pick the first door. Elimiate one of the goats other than door 1. You are left with CG, GC, and GC. Notice that two of the three possible combinations have the car behind door 2, not door 1.
I really don't envision a computer creating perfect "backing tracks" in real-time to any vocals sung into it. It's more like, it'll sometimes/often make "passable" ones, fun for karaoke or practicing -- but not worthy of recording.
Given that pop music is already arguably not worthy of recording, I'm not exactly sure that there's any impediment to this being used for pop music.
The LHC is just a tourist trap. It's like Times Square. You go thinking it's gonna be all fun, then you realize it's just a bunch of bright colors and earth eating black holes, and there's nothing to do there but shop for overpriced bosons you could've picked up for half price at a more nondescript collider.
I'd be much less repulsed by this idea if I had any belief that the fees would be distributed in a fair fashion. As someone who listens to a fair deal of indie stuff (and virtually no major label stuff), I'm concerned that there's no way in hell anyone not on a major label would get to see a dime of the money. (Not that anyone ON a major label will see a dime of the money either, what with the soul-stealing contracts they make bands sign).
Ultimately then, it becomes about subsidizing an industry in a manner that provides absolutely no incentive for any major label to make desirable music. They can produce whatever they want and take the flat fee, preventing us from voting with our wallet. As a result, music would become even more controlled by the major labels than it already is now. And that's a particularly disgusting thought.
1) Quantum computing != parallel computing.
2) A significant number of applications can and do run on 1000+ cores. Sure, most are scientific apps rather than consumer apps, but there is a market for it nevertheless. Go tell a high performance computing guy that there's no need for 1k cores on a single chip and watch him collapse laughing at you.
There are some possibilities to use quanta (?) as signal carriers, but no encryption is involved. The theory is that if you wiretap such a signal, then the original receiver will find out. So it could maybe be called "Quantum Wiretap Detection" or the like. But since this is a physical thing that relies on theoretical models that are typically not exact, it is not actually known whether this is really secure. I seem to remember that there are actually possibilities to liesten in, found in te last few months.
The reason for the encryption in the name is that the idea is to exchange a private key over the secure (but very slow) channel, which will then enable encryption over an insecure channel. So you're correct that the name is misleading. To be more accurate, it should be called quantum key exchange, not quantum encryption.
I don't see it happening in the near-future, but perhaps near the end of my life-time (I'm 20-something). And it won't be like the first computer revolution, with guys in their garages and basements screwing around with computer hardware. The first quantum computers will be only really useful for large Monte-Carlo projects (like the Earth Simulator) that require tons of computing power.
Quantum computing is nigh worthless for Monte-Carlo. Yes, you can simulate a ton of inputs and get a ton of outputs in one run, but it all collapses into one waveform in the end anyway. Throw in the fact that Monte-Carlo simulations are classified as "embarrassingly" parallel and Monte-Carlo is the last thing you'll see on quantum computing.
The problem then becomes building a quantum computer that is faster than the supercomputers of the time. The first quantum computer prototype won't just start out as a powerhouse. After we get the first quantum computer working, it may be up to a decade before we see one actually being used.
The entire notion of faster or slower is thrown out the window with quantum computing. The power of a quantum computer is not limited by its speed, but the number of qubits. Furthermore, the first quantum computer prototype already exists. Indeed it is far from a powerhouse; it was used to factor the number 15. If we could expand the number of qubits arbitrarily we would have functional laboratory quantum computers, but it's our inability to increase the number of qubits because of decoherence and other physical limitations that prevents us from having useful quantum computers.
it seems painfully inefficient to 'redo' stuff that doesn't seem to be wrong just because a cosmic ray was detected.
1) The likelihood of a cosmic ray is ridiculously small. So small in fact that the cost of rewinding progress when they are detected would be completely unnoticeable.
2) We *do* have the ability to package CPUs such that they are protected by CPUs. The problem is that the packages are so large and expensive that no one would buy them given the current probability of soft errors.
So the solution is most definitely NOT to stop shrinking transistors. Even in 10 process technology generations, the mean time to a soft error actually affecting a bit on a CPU is something like 1 million hours. Never mind whether or not that particular soft error is critical.
You have a misconception about The Plague. killed two million per year at the most. And you'd have to count all victims in all three outbreaks centuries apart to reach the total of 137 million.
2 million a year is kind of a big deal when it comprises THREE TO FOUR PERCENT of the european population at the time. It would be equivalent to almost 22 million people dying in europe per year today.
You can't speed up the server or the network you are downloading from. They can claim it can do Gig speeds but that doesn't help with anything outside of Comcast networks.
Err...what kind of servers are you downloading from that use a home user connection? When you download something most of the lines the information passes through have far greater capacity than the user until you reach close to the home user as it's the "last mile" connection that most greatly limits bandwidth.
Or how about that person who gets Comcast 160Mb down only to have it run into a 10/100 Ethernet Card.
They spend $10-$20 to buy a gigabit ethernet adapter.
Not only has the problem of reliability not been solved, it is the primary barrier to effective quantum calculations. We simply cannot keep qubits in an entangled superpositioned state for long enough to perform calculations. And you are correct, every qubit you add decreases the time to decoherence, so as you increase performance you decrease reliability. This is however less a problem of computational time (i.e. requiring multiple runs of the algorithm) and more a problem of being able to get any answer at all. In many cases a solution is extremely easy to check.
Well if your only requirement for somethings to be "Moore's Law"-ey is that it follows an exponential curve, then there are thousands of things that follow Moore's Law. Just as there are plenty of things, even computer related, that do not follow, such as memory latency.
I guess I just have a general aversion to using the terminology of Moore's Law instead of just saying exponential growth. Moore's Law is a very specific observation on the cost effectiveness of transistor count per chip (not density specifically) where low transistor counts suffer from higher cost per transistor and high transistor counts suffer from low yield, and how that sweet spot of transistor count is increasing exponentially with time. Moore's Law is fundamentally an economic limit, not a fundamental physical or technical limit.
What they're referring to is quantum entanglement. There's a vast difference between having two independent qubits each in a superposition of two states and having two entangled qubits together in a superposition of four states. It's a subtle, but very important difference.
Mathematical of course. Because this is English. We don't remove suffixes, we ADD them.
(Pun only moderately intended)
Probably not. Moore's law is a very specific and narrow-focused trend regarding the number of transistors that can be placed on a chip to achieve the highest cost effectiveness. Any major shift in technology (i.e. away from traditional transistors) will have it's own completely different trend. There's no inherent trait in computing as a whole that gives rise to Moore's law.
You can use ANY quantum mechanical system with two discrete states as a qubit, just as you can use any classical mechanical/electric system with two discrete states as a bit.
Typically with photons, it consists of the direction of polarization of the electro-magnetic field associated with the photon. Straight up and down represents one state, horizontal represents the the second state, and the photon can be in a superposition of both of these states.
Saying that photons get "destroyed" is irrelevant so long as we can measure the photon's polarization when it gets destroyed because as soon as we measure the polarization, the quantum state of the photon is destroyed anyway and becomes worthless to us. This is true of any quantum mechanical system, so whether the system representing the qubit sticks around or disappears after being measured (whether a photon, electron spin, or otherwise), is only a matter of logistics of the quantum computer, not of the actual computation.
Why would the velocity matter at all? It's the acceleration that is uncomfortable, but at the acceleration planes normally undergo 10,000 mph is easily reachable in a few minutes of constant acceleration.
What a lot of people apparently don't realize is that those people opposed to embryonic stem cell research are also opposed to the creation of embryos for fertility purposes. Claiming that they would be discarded anyway and thus should be used for scientific purposes begs the real and insufficiently addressed ethical question of whether or not they should have even been created in the first place.
I don't think it should be written off as "vaporware" but this is obviously still very much in it's infancy.
I pretty sure that even if this technology is sufficiently developed it should still be classified as vaporware.
Not more of this low-carb propaganda bullshit. Calories make you fat, regardless of whether they come from fat, sugars, or starches.
There's two big draws I think. First off, look at the success of serialized shows like Heroes and Lost. Shows with ongoing plot lines, rather than completely episodic sit-coms and dramas like CSI. Rather than being the exception, shows with a single overarching plot line planned from the start of the series (or even earlier in the case of an anime based on a manga) are the norm in most genres. So you can have development, a real crisis, and a conclusion in 13 or 26 episodes. Compared to most american shows whose primary goal is to stay on the air as long as possible, anime provides a better storytelling experience.
Secondly, animated shows can tackle any subject matter. You don't need block buster CGI effects since everything is animated anyway. So anime shows can feature sci-fi, fantasy, or ridiculous action themes much more easily than an american tv show can.
There are some people who like it because it's Japanese and exotic and weird, but all in all I don't think that's the primary reason. It's simply that the animated medium allows more flexibility and creativity than live action, but is stigmatized in america as being childish.
Not that I don't agree with your sentiment, but man in the middle attacks are incredibly difficult to defend against without a secondary trusted channel, which is simply not feasible for a lot of internet traffic.
Wow released an expansion in January. Guess when the next expansion is coming out. Early next 2008. They've also said that they plan on one a year or so from now on. They added a lot of content between release and BC. How much have they added since BC came out? One 40 man instance that was supposed to be out at the same time as the xpac itself that less than 2% of the WoW population will ever even see.
...SURPRISE SURPRISE ITS BETA)
So in conclusion, ditch the attitude if you can't even get your own fucking facts straight.
(Oh, and btw, phase and toxic grenades no longer bounce. One sticks to walls and the other explodes on contact. So not only is your entire complaint of gameplay on one skill, it's one they're tweaking because