"Nasty Little Truth About Spacetime Physics" (http://home1.gte.net/res02khr/crackpots/notorious.htm)
As far as I can tell, this is much hot air to the effect that "time travel is motion through spacetime, which is impossible because it already contains time, so all the physicists talking about time travel are crackpots".
This seems rather silly, because what physicists actually talk about when they say "time travel" is simply a configuration of an object's world-line (graphed in spacetime) such that the world-line can intersect itself (or that the "future" light-cone from the world-line crosses some "past" part of the world-line, allowing communication, or any of a number of similar scenarios). This does not involve "motion" of the hypothetical fabric of spacetime; it's just a class of paths that objects can take within it.
Possibly I have misinterpreted the document, but this seems unlikely, as it makes it abundantly clear that time travel involves "motion" of spacetime, which "is impossible".
..is it as fun [or messy] as the classic cornflour and water.
It's more fun and more messy, if my own experiences were any guide.
Fun because it will do _very_ neat things if you wave a magnet under the jar - you get a spike of material building up that follows the magnetic field lines. Use multiple magnets, and you can get even wierder-looking effects.
Messy because, being oil-based, you can't just wipe it up with a wet cloth. It also won't dry out when you spill drops of it, like cornstarch does. You'll have to wipe it up with a wad of paper towels and then thoroughly wash the area with soap before it'll come clean. Think back to the last time you spilled a cup of vegetable oil when baking for some idea of how much "fun" this is.
But seriously, could you imagine what kind of toys your children could play with if this stuff became commercial??
As the article points out, you can make fluids like this very easily - just dump a bag of iron filings into a glass of vegetable oil. It isn't pretty, and it dries out eventually, but wave a strong magnet around the jar and you can get all kinds of fun effects.
Beyond just playing with a jar of it, though, I don't really see how magnetically-sensitive fluids would be used in toys. If you see a miraculous application, feel free to enlighten me (or to run off and make a bundle on it:)).
Imagine if you had a system, where you used two movies, layed one on top of the other. It would look blury of course, but if the viewer were to wear a pair of expensive, high-tech filter glasses (having differently coloured lenses), a stereoscopic 3-D image could be achieved. I bet in 2 years time we could have a working proto-type, and then actually bring it to market within 10 years given the proper funding and agressive marketing.
3D movies have existed for quite a while. Ancient systems used colour-filter glasses to get 3D. Other ancient systems used various tricks to get limited 3D effects in full colour. The Right Way to show a 3D movie is to have two projectors running films shot for each eye, put polarized filters on the projectors, and use polarized glasses to look at the resulting image. My understanding is that this is the way 3D movies are shown now, though I don't keep up with the industry (and so could easily be mistaken).
For computers, the nicest way of doing 3D is to display alternate eyes on alternate frames, and use LCD shutter-glasses to decode it. You can buy packages for this off the shelf from several vendors; they work by replacing the rendering driver with one that renders two images and handles the synchronization of the glasses. These have existed for a while.
Now, the fact that both of these solutions have existed for a while, but that neither of these have really caught on, should tell you the most important thing:
Nobody really cares about true 3D for most entertainment or gaming applications.
If they did, stereographic glasses would have sold like hotcakes when they were first introduced.
A flat projection of a 3D world seems to be enough for most viewers, despite the industry's repeated attempts to provide something more.
If the CIA shut it down, I mean there's a reason why all the big telescopes are government operated, there's stuff out there they don't want us to see...
Really? Like what?
The US's rivals already have telescopes that are more than good enough to spot anything the US's telescopes could spot. What is there to hide?
The real reason why telescopes are government-funded is that no sane private company would build them - there's no revenue stream. They're second only to particle accelerators as an example of expensive blue-sky research.
It's too bad you never figured out that you could put the panels on both sides.
That would accomplish nothing, as it's the inside panels that matter when trying to keep a hamster in thet maze. I didn't care what the maze _looked_ like - I didn't want the hamster to escape.
The panels can be easily popped out from the back. This means that if I want walls to be effective, there _MUST_ be panels on _ALL_ inside surfaces. This is not easy to do, because the tabs on the panels stop you from building certain patterns of internal wall.
Your proposed solution isn't a solution, and has very little bearing on the problem at all.
The older Athlons, for example, run at a 200Mhz clock speed.
Nope they don't. The slowest Athlon was 500Mhz.
Um, it was pretty obvious from context that the poster was talking about FSB data sampling rate, which _was_ 200 MHz.
I write this as a person with a bachelors in Computer Engineering who is currently completing masters in EE
Please tell me your are kidding about the above.
If that one statement the original poster made was your sole basis for trying to cut him down like this... then I respectfully suggest that you take a deep breath and re-read before hitting "submit" next time.
I have a BASc in Comp. Eng. and am a bit over half way through a MASc in the same, and as far as I can tell, the original post was pretty much correct (with only one minor error that I spotted).
Now, let's do the same thing with CPI. Instead of "Megahertz GOOD!", let's all stomp our feet and say, "CPI BAD!" I'm thinking of that metallica parody here. Anyway, people understand golf scores, where lower is better -- they can be made to understand that lower CPI is better. So why doesn't AMD come out with an ad campaign saying, "The pentium 4's average CPI is 97, and ours is just 2. Therefore, our chip is FIVE TIMES as fast as a p4 at the same clock rate!!"
The problem with this is twofold.
Firstly, anyone remotely sane will pick out the "at the same clock rate" line and be suspicious. Machines with higher CPI ratings tend to have higher clock rates to compensate.
Secondly, CPI for _throughput_ is likely to be in the 0.5-1 range for almost all systems, due to pipelining (issue 1-2 instructions every clock on average, and no matter how long they stay in the pipe before being retired, your throughput is 1-2 IPC). In practice, stalls would kill this for really long pipes, but there will *always* be benchmarks that perform well. I've had to benchmark this kind of thing. You wind up with numbers all over the map.
In summary, I think the only benchmarks that will make sense will be those that more-or-less accurately represent the real workload the machines are going to be exposed to (gaming benchmarks for the gamers, office suite benchmarks for office workers, etc.). It isn't a surprise that these are the best benchmarks, but with the architectures being compared diverging, it looks like they're the _only_ valid benchmarks.
Video games are so fast paced and dynamic, they adapt to maintain kids infamously short attention spans.
Are you kidding? Do you know how long you have to sit at a computer to beat even the first act of Diablo II?:)
In all seriousness, I don't think that video games are much of a problem. Firstly, kids won't be playing them until years after they start playing with more physical toys, and second of all, most of them _do_ require dedication and focus to play (or at least to do well in).
Add to this the fact that television, with its 10-second advertisements and other fast pacing, has been around for years, and you'll have difficulty convincing me that fast-paced video games are making a difference.
I was never a fan of LEGOs (for whatever reason) but I really did like to build things w/Construx. god only knows how many times I built myself into a box and had to have my mom come and try to get me out w/o breaking the new creation I made.
I used to amuse myself by building towers that reached the ceiling (at least three times my height in those days):).
Through much deviousness, I also managed to build a working Construx pendulum clock at one point... even if the hand went a quarter-turn around with each tick. (Show your kids the guts of an old-fashioned pendulum clock some day; I was endlessly fascinated by those as a kid.)
The most challenging task, though, was to build construx mazes for my hamster in such a way that he couldn't push any of the panels out. The trick was to make sure that all of the panels attached from the inside of the tunnels, which imposed interesting design constraints.
I had the good fortune to be exposed to many building toys as a kid. Construx is still one of my favourites.
As a side note, two-by-fours and nails work too. Let your kids help build the tree-fort you're making for them:). Just take steps to ensure safety if the bottom floor is above ground level.
With many parallel threads and an internal cycle time much faster than external SRAM access time, won't this chip have a lot of trouble managing locks in parallel applications?
Will this restrict the set of applications for which this chip is useful, or have you come up with a clever solution to the problem?
Wait, I'm missing something here. What's the difference between acceleration and delta v? I seem to remember that acceleration = change in velocity. What gives?
Acceleration is the instantaneous change in velocity (derivative of velocity) at any given time. Delta-v is the integral of acceleration over time (actually of the magnitude of acceleration). For a ship accelerating in one direction outside a gravity well, it will be the total change in velocity (v_end - v_start).
Acceleration is how fast you can change your speed, and delta-v is the total amount you can change your speed by.
You state early on that mining Mars or asteroids is unlikely to be profitable, and then state that the moon is better because space colonies need radiation shielding, which can be gotten relatively cheaply by using lunar regolith. Nowhere do you state why people would want to build space colonies, rather than lunar or martian colonies. I'd be curious to know why people would rather build space colonies (which are more difficult to construct and supply) than planetside colonies.
They might not be built at all. I'm postulating that they will be, which leads to my conclusion about profitable space industries. If you assume no large space structures will be built, then I doubt that any space industries will be profitable.
The most immediate use for space stations and space colonies is as way-stations to lunar colonies and for interplanetary craft. This assumes that lunar colonies will be constructed. If we have no need for substantial interplanetary travel or colonization, then there is no need for space stations.
The safest method of building and supplying a moon base or moon colony would be to have two fairly large space stations, orbiting the moon and the earth, with solar-powered ion drive shuttles carrying cargo between them. Build the first station in Earth orbit, and use it as a testbed to work out all of the problems with building space stations and more-or-less self-sufficient environments. Build a second station in Earth orbit, and use ion drives to move it to lunar orbit. Then set up the supply line. Travel time for the ion shuttles is a few months, but they're in a constant stream and unmanned, so this isn't a problem. You now have a conveyer belt carrying food and supplies to the lunar-orbit station, and carrying waste back.
Send construction materials along this pipe, and you can build a lunar colony. Send food and supplies, making sure to keep a month or two of surplus dirt-side on the moon and/or in the lunar station, and your lunar colony can handle just about any disaster without a big, fast, expensive rescue ship being needed.
The earth-orbit station is an ideal launch station for ion-drive probes to other parts of the solar system. The lunar-orbit station is an excellent site to manage construction of other space stations or large craft from (lunar material would be sent to a nearby construction site). This is where you'd likely build a Mars-colonizing ship. The ship would have to be big, carrying all of the equipment needed for a self-sufficient Mars colony base, and would become Mars's orbiting station.
All of this presupposes a desire to build lunar or Martian colonies. Given that desire, this is probably the easiest, cheapest, and safest way of doing it. Without that desire, there's no real reason to go into space at all.
Could something else be used for acceleration? Maybe a rocket booster? Once you get up to a nice speed, let the nuclear drive take over to power the rest of the trip?
The problem is that when the nuclear drive kicks in, you drop to very low acceleration. This brings back the bone degradation problem.
If you're planning to use a mixed scheme for faster travel, the chemical stage doesn't buy you much. You need a certain delta-v for the trip. If the chemical stage gives most of it, it'll be huge (as would an all-chemical solution). If the nuclear stage gives most of it, it'll take a while to build it up (low acceleration). In practice, mixed solutions make sense only where you need short bursts of high acceleration (like takeoff and landing to/from a planet, or fast maneuvering).
For a trip longer than about a year, a nuclear-electric drive will shorten the total travel time. For trips much shorter than that, it doesn't help much.
Okay, but do you have any particular reason to believe this, or is it just a tenet of your faith? If you consider that fuel can be made relatively cheaply from local ingredients (just react some H_2 with the atmosphere, really) and that transport time isn't important for cargo, it might not be too expensive at all. Strap a booster onto your block-o-platinum and loft into Martian orbit (low gravity, so lots easier than for Earth).
You still have to loft the cargo out of the Martian gravity well, and cancel the (very large) gravitational potential energy difference between Mars's orbit and Earth's. This will be about as expensive as launching something into space from Earth - not cheap. Your fuel isn't free. It costs time and effort (read: money) to manufacture, even on Mars.
There's also no reason to believe that mining on Mars will be cheaper than mining on earth even if you *don't* transport the cargo anywhere. Why would we magically find rich veins of platinum on Mars? It has roughly earth-like composition.
If you're going to mine anything, then near-earth asteroids are your best bet, and even then, I'm skeptical of asteroid mining being worth the cost. Asteroid composition varies widely enough that you can find ones that are very rich in metal ore.
IMO, mining the moon for raw mass is probably the most practical operation that will go on in space. To build a space colony, you need a lot of mass just for radiation shielding. Moon dirt works well for that, and is a lot cheaper to loft than material from Earth. If you're building a spinning structure that has mostly tensile forces, then you can get structural material from the moon too (fiberglass cables).
Mars, on the other hand, has little that would be worth transporting back to Earth. In pretty much all cases, you'd be better off mining or manufacturing it on earth and avoiding transport costs.
OTOH, Mars is a great site for colonizing and possibly terraforming, once there are enough settlers willing to pay out of pocket for the trip.
In another book of Arthur Clack he proposes an entirely different way to get artificial gravity The spacecraft can constantly accelerate with acceleratin equal to 1g.
A possible solution would be to have a nuclear reactor and use superheated water or a gass of some sort as fuel. In this way we get very high acceleration with relatively little "reactive mass".
If we had enough delta-v to do this, we could get to Mars in less than a week, and the problem wouldn't exist.
It turns out that nuclear power doesn't help us do this.
If we're using a nuclear core to heat fuel directly (as with the NERVA project), we get efficiency comparable to a chemical rocket, because our core (and thus exhaust) temperature can't be greater than the core materials can handle without degrading.
If we're using a nuclear core to generate electricity to power an ion drive or a plasma drive or another class of electromagnetic drive, we have nice delta-v, but very low acceleration, which doesn't help either the bone problem or our total travel time (if we're just going to mars; it would help for destinations farther away).
Other styles of nuclear drive have similar problems. They're great for long-haul trips, but won't give high acceleration and high delta-v at the same time.
Fusion drives won't exist for a while, so they're not a solution candidate yet.
Really? Got references? I'd heard that they were still all mondo expensive, but that may just be Big Oil FUD.
One of the older types uses sintered nickel oxide powder as the catalyst. Nickel's cheap. This kind works fine for hydrogen processing; the problem is that if you use air as the oxygen source, the catalyst gets "poisoned" by the CO2 (stops working efficiently after a while).
Another kind used aluminum oxide.
Industry mainly uses a third type of fuel cell; I don't remember what the catalyst in it is offhand. The electrolyte is phosphoric acid.
I did a project surveying the types of fuel cells years and years ago, but my memory of it is fading.
In any case, I think solar energy is better suited to stationary or low power mobile devices, not transportation. I am a big fan of biomass energy [biomass.org] for cars. Biomass methanol has a very high net energy value, a closed carbon cycle, and is safer than compressed hydrogen.
You could also produce methanol directly from air, water, and power, which might have higher efficiency (as long as you have an efficient source of energy). I'm told that the solar conversion efficiency of plants is actually rather low (your linked page didn't list figures to check this).
Hydrogen comes by electrolysis, which is very efficient.
CO2 comes out of air by fractional distillation or by effusion (take your pick; I'd personally go with fractional distillation). Energy cost of producing the low temperatures needed will be much less than the cost of the hydrogen electrolysis, so efficiency of this step isn't very important.
Then you burn the CO2 incompletetly in a hydrogen atmosphere, and fractionally distill the results to get the methanol. The other products (water and some other simple compounds of carbon, hydrogen, and oxygen) can either be sold as solvents or for use in industrial processes, or burned (producing heat or power) and fed back into the system. Even the primary reaction (burning of CO2 in hydrogen) is exothermic, so you'll get some heat recovered from this stage too.
Cleanly powering the conversion plant is left as an exercise to the reader, but either a solar heat plant or a nuclear plant should be adequate and reasonably clean (compared to fossil fuels).
In the past I've been a huge fan of EVs, but am disolusioned by the slow rate at which battery energy density has improved, especially considering the toxicity and expense of the new materials -- even compared to lead.
Slowly, I'm warming up to the hybrids. Something must be done to cut down on fossil fuel usage.
Fuel cells work adequately as a solution to the fossil fuel problem, if you can live with less fuel or a bigger gas tank (hydrogen is the most often proposed fuel, and can't be stored at liquid densities). Many varieties of hydrogen-based fuel cells are made from cheap materials, so cost shouldn't be a problem. This skips the carbon cycle all together (source water -> hydrogen -> water vapour -> rain -> source water).
Another solution is to switch to burning methanol. You can either produce this by fermentation, or build it directly from air (for CO2), water (for H2), and power (solar, nuclear, or whatever). Both ways draw carbon back in from the environment, stopping the short-circuit of the carbon cycle that's causing problems with fossil fuels. Methanol can be burned (cleanly) in conventional internal combustion engines, and can also be burned in advanced fuel cells (which may be expensive; I'd just use a normal engine). It can be stored as a liquid, though you'd probably want to put it in a pressure vessel (like propane) to keep it from slowly boiling off.
In practice, neither of these solutions will be implemented until the cost of gasoline and diesel rises to a level high enough to justify the switchover cost.
How that thing ever got into orbit without being tested is beyond me.
My understanding was that the mirror was tested - the test was just miscalibrated (one piece of the test optics was a few centimetres out of place). They needed to test the mirror continuously while grinding it.
Tech support load varies with configuration count.
on
Dorm Storm?
·
· Score: 3, Insightful
I wouldn't be so harsh about most of your policies, if you didn't also mix in a number of shortsighted, non-benificial rules in there as well. What the hell do you care what the user does behind his/her dorm-room port? Are you filtering packets? Blocking ports? Yes? Then it doesn't matter if Joe User wants to set up a single windows PC, or establish a 10 computer NAT network in their room, hidden behind a linux firewall. Second, why would you want to alienate technically savvy users by requiring them to use hardware or software different from what they already have? If a Joe User can do his own install, do you care *what* he installs? Of course not!
Spoken like a person who's never had to do tech support.
Any user whose install doesn't go *perfectly* or who doesn't know how to install/configure network gear will be asking tech support for help. If there's one and only one allowed configuration, there's one and only one way to set up one's network card. Tech support is easy.
Allow arbitrary hardware and software to be used, and you have a geometrically increasing number of configurations that your tech support staff will be asked to troubleshoot.
Only give tech support for sanctioned configurations? That won't work very well. Joe Idiot will say, "But I paid to be on this network! Set up my machine!", or "But it's *almost* the sanctioned configuration! Now tell me why my FooCom 7 card is barfing!". Joe Linuxd00d will say, "Um, sure I'm using Windows. Help me debug my firewalling rules.". Even if you hang up on these people, you'll still get the calls.
The university's networking department has to deal with all of this crud on a budget that is almost certainly far too small. I have no problem at all with them restricting hardware and software for machines connected to the dorm network drops - they're paying for the network infrastructure and support, so they have every right to say what they'll let people do on the network.
The only way I can see to do it, is to integrate the RAM with the processor, hence you will not have a CPU in your system, just a lot of processing nodes on a bus or series of busses, need more CPU, add another PRAM (Processor-RAM) module. need to do lots of floating number work?, add some FPRAM, RAM with extra FP units.
This has been tried. It didn't work very well. There are a few problems:
Your inter-processor communications bandwidth is low. This is a serious bottleneck for many tasks.
N-way SMP performs worse than N-times-faster 1-way. Amdahl's Law and coherence operation overhead both conspire to bite you on this. Amdahl's law, especially - you can't parallelize all tasks.
And the main reason why processor+RAM modules haven't taken off:
A processor+RAM scheme is functionally equivalent to an ordinary SMP system with no main memory. An ordinary SMP box already has memory tied to processors - the processor caches. Add main memory to your multi-module machine, and you have something that looks suspiciously like an ordinary SMP box with big L3 caches made from DRAM.
For really, really large systems (hundreds of modules or more), this approach is still used (look up "NUMA" for more information), but for smaller boxes it doesn't make a lot of sense.
Eventually, to conquer the latency beast, we will need to move more memory closer to the CPU. To do that is going to take moving to serial interconnects for lower pin counts, and reducing the physical footprint on the mainboard.
I'm not sure that switching to a serial system would help enough. While you could clock it more quickly, you'd still have a hard time matching the bandwidth of a many-line solution. This could ironically result in longer latencies, because despite the higher clock speed, you'd have to sit there and wait for all 32+ bits of the missed word or 128+ bits of the cache line to be transferred before resuming operation.
IMO, a better approach might be running many shielded lines in parallel transmitting data with self-clocking codes. This allow faster clocking by removing the need to keep all lines in synch with each other; data could be rebuilt in buffers at the receiving end.
Regardless of the bus implementation, you'll still likely be limited by the speed of the RAM used.
The final solution to all of this will probably come when we can put a big enough L3 cache on a die to hold the entire working set of most programs. That will give us a short, fast, wide path to L3 memory. Main memory will only be accessed for streaming data or for random accesses to huge databases. In the first case, a high-bandwidth, high-latency bus is acceptable. In the second case, I doubt anything we do will overcome latency problems.
An interesting design problem to think about, in any event.
What is dynamic logic? How is it different from conventional logic wired together with different types of gates?
Both dynamic and static logic use logic gates or blocks that are wired together. The difference is in how the gates are implemented internally, and how they pass data back and forth.
CMOS is a good example of static logic. It uses pull-up and pull-down transistor networks to make sure that outputs are always strongly asserted. This makes CMOS gates big and makes input capacitance larger than it otherwise needs to be. But, it's well-understood, has a few attractive features, and has a whole slew of design tools built for it.
Precharge logic is a good example of dynamic logic. It uses the parasitic capacitance of the output line to store the output value. The output node is charged up on one half of the clock (precharge phase), and left floating on the other half (readout phase). During the readout phase, the inputs are asserted. Inputs are fed into a pull-down transistor network that drives the output low if it should be low, and leaves it alone if it should be high. This style of logic takes up half the space of CMOS logic, has half the input capacitance, and has stronger driving capability (NFETs pulling down typically drive 2x-3x more strongly than PFETs pulling up). This means that if you play your cards right, you can make precharge logic circuits that are faster *and* more compact than CMOS logic circuits. The downsides are that designing and verifying precharge logic is a royal pain, and that you have to have a clock input into the logic block.
The article describes a more complicated dynamic logic scheme with a four-phase clock. These kinds of schemes have been floating around in research literature for years, but are usually not used because of the greater complexity and fewer tools available.
Slashdot Mirror
Your sig:
s .htm)
"Nasty Little Truth About Spacetime Physics" (http://home1.gte.net/res02khr/crackpots/notoriou
As far as I can tell, this is much hot air to the effect that "time travel is motion through spacetime, which is impossible because it already contains time, so all the physicists talking about time travel are crackpots".
This seems rather silly, because what physicists actually talk about when they say "time travel" is simply a configuration of an object's world-line (graphed in spacetime) such that the world-line can intersect itself (or that the "future" light-cone from the world-line crosses some "past" part of the world-line, allowing communication, or any of a number of similar scenarios). This does not involve "motion" of the hypothetical fabric of spacetime; it's just a class of paths that objects can take within it.
Possibly I have misinterpreted the document, but this seems unlikely, as it makes it abundantly clear that time travel involves "motion" of spacetime, which "is impossible".
Can you clear up this apparent discrepancy?
..is it as fun [or messy] as the classic cornflour and water.
It's more fun and more messy, if my own experiences were any guide.
Fun because it will do _very_ neat things if you wave a magnet under the jar - you get a spike of material building up that follows the magnetic field lines. Use multiple magnets, and you can get even wierder-looking effects.
Messy because, being oil-based, you can't just wipe it up with a wet cloth. It also won't dry out when you spill drops of it, like cornstarch does. You'll have to wipe it up with a wad of paper towels and then thoroughly wash the area with soap before it'll come clean. Think back to the last time you spilled a cup of vegetable oil when baking for some idea of how much "fun" this is.
But seriously, could you imagine what kind of toys your children could play with if this stuff became commercial??
:)).
As the article points out, you can make fluids like this very easily - just dump a bag of iron filings into a glass of vegetable oil. It isn't pretty, and it dries out eventually, but wave a strong magnet around the jar and you can get all kinds of fun effects.
Beyond just playing with a jar of it, though, I don't really see how magnetically-sensitive fluids would be used in toys. If you see a miraculous application, feel free to enlighten me (or to run off and make a bundle on it
Imagine if you had a system, where you used two movies, layed one on top of the other. It would look blury of course, but if the viewer were to wear a pair of expensive, high-tech filter glasses (having differently coloured lenses), a stereoscopic 3-D image could be achieved. I bet in 2 years time we could have a working proto-type, and then actually bring it to market within 10 years given the proper funding and agressive marketing.
3D movies have existed for quite a while. Ancient systems used colour-filter glasses to get 3D. Other ancient systems used various tricks to get limited 3D effects in full colour. The Right Way to show a 3D movie is to have two projectors running films shot for each eye, put polarized filters on the projectors, and use polarized glasses to look at the resulting image. My understanding is that this is the way 3D movies are shown now, though I don't keep up with the industry (and so could easily be mistaken).
For computers, the nicest way of doing 3D is to display alternate eyes on alternate frames, and use LCD shutter-glasses to decode it. You can buy packages for this off the shelf from several vendors; they work by replacing the rendering driver with one that renders two images and handles the synchronization of the glasses. These have existed for a while.
Now, the fact that both of these solutions have existed for a while, but that neither of these have really caught on, should tell you the most important thing:
Nobody really cares about true 3D for most entertainment or gaming applications.
If they did, stereographic glasses would have sold like hotcakes when they were first introduced.
A flat projection of a 3D world seems to be enough for most viewers, despite the industry's repeated attempts to provide something more.
If the CIA shut it down, I mean there's a reason why all the big telescopes are government operated, there's stuff out there they don't want us to see...
Really? Like what?
The US's rivals already have telescopes that are more than good enough to spot anything the US's telescopes could spot. What is there to hide?
The real reason why telescopes are government-funded is that no sane private company would build them - there's no revenue stream. They're second only to particle accelerators as an example of expensive blue-sky research.
It's too bad you never figured out that you could put the panels on both sides.
That would accomplish nothing, as it's the inside panels that matter when trying to keep a hamster in thet maze. I didn't care what the maze _looked_ like - I didn't want the hamster to escape.
The panels can be easily popped out from the back. This means that if I want walls to be effective, there _MUST_ be panels on _ALL_ inside surfaces. This is not easy to do, because the tabs on the panels stop you from building certain patterns of internal wall.
Your proposed solution isn't a solution, and has very little bearing on the problem at all.
The older Athlons, for example, run at a 200Mhz clock speed.
Nope they don't. The slowest Athlon was 500Mhz.
Um, it was pretty obvious from context that the poster was talking about FSB data sampling rate, which _was_ 200 MHz.
I write this as a person with a bachelors in Computer Engineering who is currently completing masters in EE
Please tell me your are kidding about the above.
If that one statement the original poster made was your sole basis for trying to cut him down like this... then I respectfully suggest that you take a deep breath and re-read before hitting "submit" next time.
I have a BASc in Comp. Eng. and am a bit over half way through a MASc in the same, and as far as I can tell, the original post was pretty much correct (with only one minor error that I spotted).
Now, let's do the same thing with CPI. Instead of "Megahertz GOOD!", let's all stomp our feet and say, "CPI BAD!" I'm thinking of that metallica parody here. Anyway, people understand golf scores, where lower is better -- they can be made to understand that lower CPI is better. So why doesn't AMD come out with an ad campaign saying, "The pentium 4's average CPI is 97, and ours is just 2. Therefore, our chip is FIVE TIMES as fast as a p4 at the same clock rate!!"
The problem with this is twofold.
Firstly, anyone remotely sane will pick out the "at the same clock rate" line and be suspicious. Machines with higher CPI ratings tend to have higher clock rates to compensate.
Secondly, CPI for _throughput_ is likely to be in the 0.5-1 range for almost all systems, due to pipelining (issue 1-2 instructions every clock on average, and no matter how long they stay in the pipe before being retired, your throughput is 1-2 IPC). In practice, stalls would kill this for really long pipes, but there will *always* be benchmarks that perform well. I've had to benchmark this kind of thing. You wind up with numbers all over the map.
In summary, I think the only benchmarks that will make sense will be those that more-or-less accurately represent the real workload the machines are going to be exposed to (gaming benchmarks for the gamers, office suite benchmarks for office workers, etc.). It isn't a surprise that these are the best benchmarks, but with the architectures being compared diverging, it looks like they're the _only_ valid benchmarks.
Video games are so fast paced and dynamic, they adapt to maintain kids infamously short attention spans.
:)
Are you kidding? Do you know how long you have to sit at a computer to beat even the first act of Diablo II?
In all seriousness, I don't think that video games are much of a problem. Firstly, kids won't be playing them until years after they start playing with more physical toys, and second of all, most of them _do_ require dedication and focus to play (or at least to do well in).
Add to this the fact that television, with its 10-second advertisements and other fast pacing, has been around for years, and you'll have difficulty convincing me that fast-paced video games are making a difference.
I was never a fan of LEGOs (for whatever reason) but I really did like to build things w/Construx. god only knows how many times I built myself into a box and had to have my mom come and try to get me out w/o breaking the new creation I made.
:).
:). Just take steps to ensure safety if the bottom floor is above ground level.
I used to amuse myself by building towers that reached the ceiling (at least three times my height in those days)
Through much deviousness, I also managed to build a working Construx pendulum clock at one point... even if the hand went a quarter-turn around with each tick. (Show your kids the guts of an old-fashioned pendulum clock some day; I was endlessly fascinated by those as a kid.)
The most challenging task, though, was to build construx mazes for my hamster in such a way that he couldn't push any of the panels out. The trick was to make sure that all of the panels attached from the inside of the tunnels, which imposed interesting design constraints.
I had the good fortune to be exposed to many building toys as a kid. Construx is still one of my favourites.
As a side note, two-by-fours and nails work too. Let your kids help build the tree-fort you're making for them
With many parallel threads and an internal cycle time much faster than external SRAM access time, won't this chip have a lot of trouble managing locks in parallel applications?
Will this restrict the set of applications for which this chip is useful, or have you come up with a clever solution to the problem?
Wait, I'm missing something here. What's the difference between acceleration and delta v? I seem to remember that acceleration = change in velocity. What gives?
Acceleration is the instantaneous change in velocity (derivative of velocity) at any given time. Delta-v is the integral of acceleration over time (actually of the magnitude of acceleration). For a ship accelerating in one direction outside a gravity well, it will be the total change in velocity (v_end - v_start).
Acceleration is how fast you can change your speed, and delta-v is the total amount you can change your speed by.
You state early on that mining Mars or asteroids is unlikely to be profitable, and then state that the moon is better because space colonies need radiation shielding, which can be gotten relatively cheaply by using lunar regolith. Nowhere do you state why people would want to build space colonies, rather than lunar or martian colonies. I'd be curious to know why people would rather build space colonies (which are more difficult to construct and supply) than planetside colonies.
They might not be built at all. I'm postulating that they will be, which leads to my conclusion about profitable space industries. If you assume no large space structures will be built, then I doubt that any space industries will be profitable.
The most immediate use for space stations and space colonies is as way-stations to lunar colonies and for interplanetary craft. This assumes that lunar colonies will be constructed. If we have no need for substantial interplanetary travel or colonization, then there is no need for space stations.
The safest method of building and supplying a moon base or moon colony would be to have two fairly large space stations, orbiting the moon and the earth, with solar-powered ion drive shuttles carrying cargo between them. Build the first station in Earth orbit, and use it as a testbed to work out all of the problems with building space stations and more-or-less self-sufficient environments. Build a second station in Earth orbit, and use ion drives to move it to lunar orbit. Then set up the supply line. Travel time for the ion shuttles is a few months, but they're in a constant stream and unmanned, so this isn't a problem. You now have a conveyer belt carrying food and supplies to the lunar-orbit station, and carrying waste back.
Send construction materials along this pipe, and you can build a lunar colony. Send food and supplies, making sure to keep a month or two of surplus dirt-side on the moon and/or in the lunar station, and your lunar colony can handle just about any disaster without a big, fast, expensive rescue ship being needed.
The earth-orbit station is an ideal launch station for ion-drive probes to other parts of the solar system. The lunar-orbit station is an excellent site to manage construction of other space stations or large craft from (lunar material would be sent to a nearby construction site). This is where you'd likely build a Mars-colonizing ship. The ship would have to be big, carrying all of the equipment needed for a self-sufficient Mars colony base, and would become Mars's orbiting station.
All of this presupposes a desire to build lunar or Martian colonies. Given that desire, this is probably the easiest, cheapest, and safest way of doing it. Without that desire, there's no real reason to go into space at all.
Could something else be used for acceleration? Maybe a rocket booster? Once you get up to a nice speed, let the nuclear drive take over to power the rest of the trip?
The problem is that when the nuclear drive kicks in, you drop to very low acceleration. This brings back the bone degradation problem.
If you're planning to use a mixed scheme for faster travel, the chemical stage doesn't buy you much. You need a certain delta-v for the trip. If the chemical stage gives most of it, it'll be huge (as would an all-chemical solution). If the nuclear stage gives most of it, it'll take a while to build it up (low acceleration). In practice, mixed solutions make sense only where you need short bursts of high acceleration (like takeoff and landing to/from a planet, or fast maneuvering).
For a trip longer than about a year, a nuclear-electric drive will shorten the total travel time. For trips much shorter than that, it doesn't help much.
Still a fascinating topic to think about, though.
Okay, but do you have any particular reason to believe this, or is it just a tenet of your faith? If you consider that fuel can be made relatively cheaply from local ingredients (just react some H_2 with the atmosphere, really) and that transport time isn't important for cargo, it might not be too expensive at all. Strap a booster onto your block-o-platinum and loft into Martian orbit (low gravity, so lots easier than for Earth).
You still have to loft the cargo out of the Martian gravity well, and cancel the (very large) gravitational potential energy difference between Mars's orbit and Earth's. This will be about as expensive as launching something into space from Earth - not cheap. Your fuel isn't free. It costs time and effort (read: money) to manufacture, even on Mars.
There's also no reason to believe that mining on Mars will be cheaper than mining on earth even if you *don't* transport the cargo anywhere. Why would we magically find rich veins of platinum on Mars? It has roughly earth-like composition.
If you're going to mine anything, then near-earth asteroids are your best bet, and even then, I'm skeptical of asteroid mining being worth the cost. Asteroid composition varies widely enough that you can find ones that are very rich in metal ore.
IMO, mining the moon for raw mass is probably the most practical operation that will go on in space. To build a space colony, you need a lot of mass just for radiation shielding. Moon dirt works well for that, and is a lot cheaper to loft than material from Earth. If you're building a spinning structure that has mostly tensile forces, then you can get structural material from the moon too (fiberglass cables).
Mars, on the other hand, has little that would be worth transporting back to Earth. In pretty much all cases, you'd be better off mining or manufacturing it on earth and avoiding transport costs.
OTOH, Mars is a great site for colonizing and possibly terraforming, once there are enough settlers willing to pay out of pocket for the trip.
In another book of Arthur Clack he proposes an entirely different way to get artificial gravity The spacecraft can constantly accelerate with acceleratin equal to 1g.
A possible solution would be to have a nuclear reactor and use superheated water or a gass of some sort as fuel. In this way we get very high acceleration with relatively little "reactive mass".
If we had enough delta-v to do this, we could get to Mars in less than a week, and the problem wouldn't exist.
It turns out that nuclear power doesn't help us do this.
If we're using a nuclear core to heat fuel directly (as with the NERVA project), we get efficiency comparable to a chemical rocket, because our core (and thus exhaust) temperature can't be greater than the core materials can handle without degrading.
If we're using a nuclear core to generate electricity to power an ion drive or a plasma drive or another class of electromagnetic drive, we have nice delta-v, but very low acceleration, which doesn't help either the bone problem or our total travel time (if we're just going to mars; it would help for destinations farther away).
Other styles of nuclear drive have similar problems. They're great for long-haul trips, but won't give high acceleration and high delta-v at the same time.
Fusion drives won't exist for a while, so they're not a solution candidate yet.
Really? Got references? I'd heard that they were still all mondo expensive, but that may just be Big Oil FUD.
One of the older types uses sintered nickel oxide powder as the catalyst. Nickel's cheap. This kind works fine for hydrogen processing; the problem is that if you use air as the oxygen source, the catalyst gets "poisoned" by the CO2 (stops working efficiently after a while).
Another kind used aluminum oxide.
Industry mainly uses a third type of fuel cell; I don't remember what the catalyst in it is offhand. The electrolyte is phosphoric acid.
I did a project surveying the types of fuel cells years and years ago, but my memory of it is fading.
In any case, I think solar energy is better suited to stationary or low power mobile devices, not transportation. I am a big fan of biomass energy [biomass.org] for cars. Biomass methanol has a very high net energy value, a closed carbon cycle, and is safer than compressed hydrogen.
You could also produce methanol directly from air, water, and power, which might have higher efficiency (as long as you have an efficient source of energy). I'm told that the solar conversion efficiency of plants is actually rather low (your linked page didn't list figures to check this).
Hydrogen comes by electrolysis, which is very efficient.
CO2 comes out of air by fractional distillation or by effusion (take your pick; I'd personally go with fractional distillation). Energy cost of producing the low temperatures needed will be much less than the cost of the hydrogen electrolysis, so efficiency of this step isn't very important.
Then you burn the CO2 incompletetly in a hydrogen atmosphere, and fractionally distill the results to get the methanol. The other products (water and some other simple compounds of carbon, hydrogen, and oxygen) can either be sold as solvents or for use in industrial processes, or burned (producing heat or power) and fed back into the system. Even the primary reaction (burning of CO2 in hydrogen) is exothermic, so you'll get some heat recovered from this stage too.
Cleanly powering the conversion plant is left as an exercise to the reader, but either a solar heat plant or a nuclear plant should be adequate and reasonably clean (compared to fossil fuels).
In the past I've been a huge fan of EVs, but am disolusioned by the slow rate at which battery energy density has improved, especially considering the toxicity and expense of the new materials -- even compared to lead.
Slowly, I'm warming up to the hybrids. Something must be done to cut down on fossil fuel usage.
Fuel cells work adequately as a solution to the fossil fuel problem, if you can live with less fuel or a bigger gas tank (hydrogen is the most often proposed fuel, and can't be stored at liquid densities). Many varieties of hydrogen-based fuel cells are made from cheap materials, so cost shouldn't be a problem. This skips the carbon cycle all together (source water -> hydrogen -> water vapour -> rain -> source water).
Another solution is to switch to burning methanol. You can either produce this by fermentation, or build it directly from air (for CO2), water (for H2), and power (solar, nuclear, or whatever). Both ways draw carbon back in from the environment, stopping the short-circuit of the carbon cycle that's causing problems with fossil fuels. Methanol can be burned (cleanly) in conventional internal combustion engines, and can also be burned in advanced fuel cells (which may be expensive; I'd just use a normal engine). It can be stored as a liquid, though you'd probably want to put it in a pressure vessel (like propane) to keep it from slowly boiling off.
In practice, neither of these solutions will be implemented until the cost of gasoline and diesel rises to a level high enough to justify the switchover cost.
How that thing ever got into orbit without being tested is beyond me.
My understanding was that the mirror was tested - the test was just miscalibrated (one piece of the test optics was a few centimetres out of place). They needed to test the mirror continuously while grinding it.
I wouldn't be so harsh about most of your policies, if you didn't also mix in a number of shortsighted, non-benificial rules in there as well. What the hell do you care what the user does behind his/her dorm-room port? Are you filtering packets? Blocking ports? Yes? Then it doesn't matter if Joe User wants to set up a single windows PC, or establish a 10 computer NAT network in their room, hidden behind a linux firewall. Second, why would you want to alienate technically savvy users by requiring them to use hardware or software different from what they already have? If a Joe User can do his own install, do you care *what* he installs? Of course not!
Spoken like a person who's never had to do tech support.
Any user whose install doesn't go *perfectly* or who doesn't know how to install/configure network gear will be asking tech support for help. If there's one and only one allowed configuration, there's one and only one way to set up one's network card. Tech support is easy.
Allow arbitrary hardware and software to be used, and you have a geometrically increasing number of configurations that your tech support staff will be asked to troubleshoot.
Only give tech support for sanctioned configurations? That won't work very well. Joe Idiot will say, "But I paid to be on this network! Set up my machine!", or "But it's *almost* the sanctioned configuration! Now tell me why my FooCom 7 card is barfing!". Joe Linuxd00d will say, "Um, sure I'm using Windows. Help me debug my firewalling rules.". Even if you hang up on these people, you'll still get the calls.
The university's networking department has to deal with all of this crud on a budget that is almost certainly far too small. I have no problem at all with them restricting hardware and software for machines connected to the dorm network drops - they're paying for the network infrastructure and support, so they have every right to say what they'll let people do on the network.
This has been tried. It didn't work very well. There are a few problems:
This is a serious bottleneck for many tasks.
Amdahl's Law and coherence operation overhead both conspire to bite you on this. Amdahl's law, especially - you can't parallelize all tasks.
And the main reason why processor+RAM modules haven't taken off:
An ordinary SMP box already has memory tied to processors - the processor caches. Add main memory to your multi-module machine, and you have something that looks suspiciously like an ordinary SMP box with big L3 caches made from DRAM.
For really, really large systems (hundreds of modules or more), this approach is still used (look up "NUMA" for more information), but for smaller boxes it doesn't make a lot of sense.
Eventually, to conquer the latency beast, we will need to move more memory closer to the CPU. To do that is going to take moving to serial interconnects for lower pin counts, and reducing the physical footprint on the mainboard.
I'm not sure that switching to a serial system would help enough. While you could clock it more quickly, you'd still have a hard time matching the bandwidth of a many-line solution. This could ironically result in longer latencies, because despite the higher clock speed, you'd have to sit there and wait for all 32+ bits of the missed word or 128+ bits of the cache line to be transferred before resuming operation.
IMO, a better approach might be running many shielded lines in parallel transmitting data with self-clocking codes. This allow faster clocking by removing the need to keep all lines in synch with each other; data could be rebuilt in buffers at the receiving end.
Regardless of the bus implementation, you'll still likely be limited by the speed of the RAM used.
The final solution to all of this will probably come when we can put a big enough L3 cache on a die to hold the entire working set of most programs. That will give us a short, fast, wide path to L3 memory. Main memory will only be accessed for streaming data or for random accesses to huge databases. In the first case, a high-bandwidth, high-latency bus is acceptable. In the second case, I doubt anything we do will overcome latency problems.
An interesting design problem to think about, in any event.
Dynamic logic circuits can handle more complex functions with fewer steps than static logic circuits So does this mean specialized applications/OSes?
Short answer: No.
Programs see the chip's high-level design only. Low-level implementation is hidden.
What is dynamic logic? How is it different from conventional logic wired together with different types of gates?
Both dynamic and static logic use logic gates or blocks that are wired together. The difference is in how the gates are implemented internally, and how they pass data back and forth.
CMOS is a good example of static logic. It uses pull-up and pull-down transistor networks to make sure that outputs are always strongly asserted. This makes CMOS gates big and makes input capacitance larger than it otherwise needs to be. But, it's well-understood, has a few attractive features, and has a whole slew of design tools built for it.
Precharge logic is a good example of dynamic logic. It uses the parasitic capacitance of the output line to store the output value. The output node is charged up on one half of the clock (precharge phase), and left floating on the other half (readout phase). During the readout phase, the inputs are asserted. Inputs are fed into a pull-down transistor network that drives the output low if it should be low, and leaves it alone if it should be high. This style of logic takes up half the space of CMOS logic, has half the input capacitance, and has stronger driving capability (NFETs pulling down typically drive 2x-3x more strongly than PFETs pulling up). This means that if you play your cards right, you can make precharge logic circuits that are faster *and* more compact than CMOS logic circuits. The downsides are that designing and verifying precharge logic is a royal pain, and that you have to have a clock input into the logic block.
The article describes a more complicated dynamic logic scheme with a four-phase clock. These kinds of schemes have been floating around in research literature for years, but are usually not used because of the greater complexity and fewer tools available.