Hydrogen fuses to Helium. Helium fuses to Beryllium then Carbon. Carbon can fuse to numerous other elements - Oxygen, Neon, Magnesium, etc.
The problem is that fusing heavier elements gives less and less energy, and require more and more energy to trigger. Sol, for instance, will never start fusing Carbon into heavier elements. This causes the star to form distinct layers, as only deep within large stars does sufficient pressure exist to trigger heavier and heavier fusion.
Once you hit Iron*, it's a net negative - the reaction no longer produces energy. When a star's core fuses all the Silicon it can into Iron, the star stops producing energy. Note that, previously, that fusion energy had been "inflating" the star, keeping the outer layers from collapsing inward under the massive gravitational pull.
So the star collapses. As it does so, it generates a tremendous amount of energy - it's converting gravitational potential energy into anything else. Some of the energy gets used in further fission - this is where most heavy elements are made. But most of it is released in a violent explosion that blows off most of the outer layers, leaving behind a neutron star, black hole, or other weird-ass remnant.
* Technically an unstable isotope of Nickel that decays into Iron, but the end product is Iron
Mozilla gives those options because users want them. Wikipedia may not be paying, but it's genuinely useful. Amazon and eBay might be paying, but honestly, they're well-known enough that Mozilla would probably include them even if they weren't paying.
Likewise, if Google stopped paying, they'd likely still be in the dropdown list since they're so popular for search, and easily a worthwhile search engine. They just would not be the default (or, since you can configure the default, the "default" default, as it were). That would go to whoever paid more, Yahoo, Bing, or DDG. Probably Bing.
Further, I think Mozilla would survive even if they lost all their money from Google and couldn't replace it. They wouldn't be nearly so large - they'd have to cut Firefox OS and dozens of other vanity projects - but I would think that may help Firefox in the long run.
Really, what was the Cray when it first came out? One vector processing unit. How many does this new NVidia board have? How much faster are they than the original Cray?
2,880 "cores", each able to do one single-precision FMA per clock (double-precision takes three clocks for this card, but 24 clocks for most gaming GPUs). These are organized into fifteen "SMX Units", which have 192 ALUs apiece (with four schedulers and eight dispatch units). The exact clock rate is variable, as it will boost the clock speed above "normal" levels, thermal and power conditions permitting, but 1GHz is a good enough approximation. This comes out to about 1.92TFLOPS, 128GFLOPS per SMX, or (interestingly) 666 (point six repeating) MFLOPS per core.
The Cray-1 worked on arrays of up to 64 units (each 64-bits wide) at a time, and it could execute (in optimal cases) two instructions per clock. At 80MHz, that comes out to 160MFLOPS on a single core.
By that math, a single Kepler core is about four times as powerful as a Cray-1, and a full SMX is eight hundred times as powerful.
(I won't be surprised if someone corrects me on something, either miscalculating the Cray-1 or maybe even Kepler - feel free to tell me I'm wrong)
Asteroids hit the ground at about room temperature on the surface, and often below ambient temperature internally. Anything being blasted into orbit would experience similar temperature increases on the way out.
The spot directly impacted would be heated considerably, but the crater would extend many miles beyond that. I find it entirely feasible that such "life-bearing" rocks could exist.
What I find less likely are the odds of these rocks being blasted well out of Earth orbit, and impacting another planet or moon. The delta-e between Earth and Mars, let alone Earth and Europa, is pretty significant, and more importantly, space is really goddamn big. Yet even that stands up, under further consideration. They're claiming only 3 parts in a million hit Europa, which seems reasonable. I would expect far more rocks to have hit Mars or particularly the Moon, which brings up some interesting questions.
And arguably a better weapon anyways. Build a Chinese repeating crossbow - fast as a bolt-action rifle, greater range and accuracy than current plastic printed guns, and far, FAR less likely to explode in your face.
Seems a bit information-inefficient. Four characters for the year, but only one for the model specifier? That's the inverse of what's needed - there's a 60-100% improvement in performance per year, but the range between the lowest-end GPU and highest-end GPU is closer to 10,000%.
Try coming up with a better one. Invent a new naming convention that still hits all the same price points (let's say $150, $170, $200, $250, $300, $350, $450, and $600). And also accounts for releasing a new batch every year, maybe every other year.
Seriously, I tried once. I ended up in about the same place AMD is right now. Nvidia's naming is a bit wonky, partially because they've never been the clearest, partially because AMD just forced them to drop prices VERY abruptly, and partially because they're in the middle of a transition. AMD's is actually very clear, at least through the gaming card range (the 250 and below are odd, but they're also basically useless for gaming).
The first number is the generation. We're on "2", even though they just started this new numbering scheme this year, but that's fine.
The next number is the "category". Best way to think of it is monitor resolution: _70 is for 1080p - you'll get 60FPS+ on most games at max settings, real killers may need a settings drop but you'll generally be fine. _80 is for 120Hz or 1440p monitors, and the _90 is for tri-monitor 1080p, 4K, or obscene multi-GPU rigs. And an _60 part is a lower-quality 1080p - think "high" or "medium", not "max".
An X suffix means it's the "full" part, the lack of an X means it's been binned in some way (reduced clockspeeds and/or some cores disabled). For example, the 290X has 44 "compute units", while the 290 has 40 at a slightly lower clockspeed. On the 270s, they're both 20 compute units, but the 270X is clocked about 10% higher.
Since both new consoles use AMD chips, it's worthwhile to compare to them. The PS4 is a bit weaker version of the 270, and the Xbox One is a slightly underclocked 260.
Nvidia's scheme is similar (add another 0 on the end for no reason, swap "Ti" for "X", and be generation 7 instead of 2), but they've complicated it right now by not rebadging old chips as new names. AMD's recent launches were basically "launch a new 9-tier chip, take all the old ones, up the clockspeeds, bump them down a tier and cut their prices accordingly". The 270s that just launched are essentially overclocked 7870s (think "180X").
Right now, Nvidia's lineup starts at the 650 and 650 Ti Boost (medium-end 1080p), 660 and 760 (high-end 1080p), 770 and 780 (1440p), and the 780 Ti (4K). Nobody's really sure whether they're going to launch more low-end 700-series parts. They're also looser with which ones are low bins of what - the 780 is a binned 780 Ti, but the 760 is a binned 770.
PS: Ignore the Titan. It's no longer a gaming card - the 780 Ti outperforms it (the Titan is a binned 780 Ti), at $300 less.
The other advantage of the Titan is the double-precision performance. Almost all of Nvidia's cards, including the 780 Ti, run double-precision floating-point calculations at 1/24th the rate of single-precision, but for the Titan and the Tesla pure-GPGPU cards, it's 1/3rd the rate.
While I'm not sure if that's an actual hardware difference, or if it's some software limitation, or a mix of both or whatever, it's definitely real. That's the main reason a Titan is still $1000 - it's being sold as a low-end compute card, not a high-end gaming card.
I browse everything with Adblock, just because it's saved my ass from malware injected into ads more than once. If I like a site sufficiently, I give it a chance. But if I see any of the following, I turn it back on: Flash animated ads (no video - small animations are a maybe) - this is what got/. back on the block list political/religious ads ads literally in the middle of the content anything disguised as a download button, or disguised as anything but an ad
And I also tend to turn it back on if I never see relevant ads. For instance, I read a *lot* of webcomics. What's advertised a lot on webcomics? Why, other comics. It's fairly common for me to go weeks seeing only ads for comics I already read, or ones that I've checked out and decided I don't like. At that point, I often turn blocking back on, if it's even mildly irritating.
Yes. The HHGG movie gets a lot of flak for not following the books, but it was a successful comedy, and it kept the general humor stylings of the books.
Really, I think it would have been worse if it had copied the books directly. Comedy relies heavily on surprise, the unexpected. If all of the jokes were ones that we had heard before, it would have fallen flat. In retrospect, the funniest parts were those that were new - or at least done differently than in the books (turning "So Long and Thanks For All the Fish" into a musical number, for example).
The Titan isn't positioned as a high-end gaming card as much as it is a low-end scientific computing card. It's the cheapest GPU that has reasonable double-precision floating-point performance. For whatever reason, most Kepler cards run DP operations at 1/24th the speed of single-precision, but the Titan and most of the Tesla cards are able to do so at 1/3rd the speed. There, the Titan runs thousands less than the similar Tesla cards (the K20 is listed on Newegg for $3500, and the K20X is on Amazon for $7700).
The fact that the Titan also gets some buys from gamers with way too much money is just a side bonus. Even since the 780 came out, it's been extremely wasteful to get a Titan for gaming. And Nvidia's own 780 Ti is likely to out-perform the Titan in games for $300 less. Really, I think the only reason they ever marked it as a gaming card was as a publicity stunt - they held the title of "fastest card ever" for quite a while, and they held it by an impressive lead.
Going faster than light in a vacuum would violate causality. Going faster than light from point A to point B is trivial if you can take a shortcut. For instance, via wormhole. Light following the same path you do would still beat you, but you still get there in a reasonable amount of time.
Yes, 1280x800 is a horribly small resolution. But at least it's a good aspect ratio. 4:3 is bad for entertainment use - movies, games, and the like. 16:9 is similarly weak in productive use - even putting two windows side-by-side, it's not tall enough, and rotating it to portrait mode is often laughable. But, IMO, 16:10 is a good compromise - it works well for anything you do with it.
If it weren't for the fact that 2560x1600 monitors are absurdly overpriced compared to 2560x1440, I'd have gotten one of those for my primary monitor instead.
They're enabling a lower power state on the WiFi chip. So if it's not being used, it can drop to a power state lower than the one it was using previously. All this would do is add extra wakeup time when you suddenly start network activity, and I can't imagine it would be enough extra time to be noticeable.
Silk Road failed through secondary methods. The actual mechanism was solid, it's just that DPR was sloppy while setting it up, enabling regular investigative work to find him. I will note that we haven't heard of Silk Road clients being arrested en masse, which means they probably haven't been able to track down many of them. They caught the operator, shut down the site, and seized a large amount of money, but the fact that they haven't been arresting left and right the drug dealers who used it means they aren't getting much information from it.
Basically, all someone needs to do is do the exact same thing Silk Road did, except avoiding leaking information on other bands. Use a secure email, don't even think about the site on unencrypted channels, establish a completely new identity for running the site, etc.
A self-fulfilling prediction. Once it was accurate for a few iterations, it moved from being a prediction to being a target, and engineers at every major foundry have worked to meet or even exceed it.
Slashdot Mirror
Well, it's common to learn from the mistakes of others, isn't it?
Hydrogen fuses to Helium. Helium fuses to Beryllium then Carbon. Carbon can fuse to numerous other elements - Oxygen, Neon, Magnesium, etc.
The problem is that fusing heavier elements gives less and less energy, and require more and more energy to trigger. Sol, for instance, will never start fusing Carbon into heavier elements. This causes the star to form distinct layers, as only deep within large stars does sufficient pressure exist to trigger heavier and heavier fusion.
Once you hit Iron*, it's a net negative - the reaction no longer produces energy. When a star's core fuses all the Silicon it can into Iron, the star stops producing energy. Note that, previously, that fusion energy had been "inflating" the star, keeping the outer layers from collapsing inward under the massive gravitational pull.
So the star collapses. As it does so, it generates a tremendous amount of energy - it's converting gravitational potential energy into anything else. Some of the energy gets used in further fission - this is where most heavy elements are made. But most of it is released in a violent explosion that blows off most of the outer layers, leaving behind a neutron star, black hole, or other weird-ass remnant.
* Technically an unstable isotope of Nickel that decays into Iron, but the end product is Iron
Mozilla gives those options because users want them. Wikipedia may not be paying, but it's genuinely useful. Amazon and eBay might be paying, but honestly, they're well-known enough that Mozilla would probably include them even if they weren't paying.
Likewise, if Google stopped paying, they'd likely still be in the dropdown list since they're so popular for search, and easily a worthwhile search engine. They just would not be the default (or, since you can configure the default, the "default" default, as it were). That would go to whoever paid more, Yahoo, Bing, or DDG. Probably Bing.
Further, I think Mozilla would survive even if they lost all their money from Google and couldn't replace it. They wouldn't be nearly so large - they'd have to cut Firefox OS and dozens of other vanity projects - but I would think that may help Firefox in the long run.
So that's why BTC jumped up to $900 today, after opening around $500. I bet some speculators made a ton of cash that way.
Really, what was the Cray when it first came out? One vector processing unit. How many does this new NVidia board have? How much faster are they than the original Cray?
2,880 "cores", each able to do one single-precision FMA per clock (double-precision takes three clocks for this card, but 24 clocks for most gaming GPUs). These are organized into fifteen "SMX Units", which have 192 ALUs apiece (with four schedulers and eight dispatch units). The exact clock rate is variable, as it will boost the clock speed above "normal" levels, thermal and power conditions permitting, but 1GHz is a good enough approximation. This comes out to about 1.92TFLOPS, 128GFLOPS per SMX, or (interestingly) 666 (point six repeating) MFLOPS per core.
The Cray-1 worked on arrays of up to 64 units (each 64-bits wide) at a time, and it could execute (in optimal cases) two instructions per clock. At 80MHz, that comes out to 160MFLOPS on a single core.
By that math, a single Kepler core is about four times as powerful as a Cray-1, and a full SMX is eight hundred times as powerful.
(I won't be surprised if someone corrects me on something, either miscalculating the Cray-1 or maybe even Kepler - feel free to tell me I'm wrong)
What, no direct link to the site?
Not really.
Asteroids hit the ground at about room temperature on the surface, and often below ambient temperature internally. Anything being blasted into orbit would experience similar temperature increases on the way out.
The spot directly impacted would be heated considerably, but the crater would extend many miles beyond that. I find it entirely feasible that such "life-bearing" rocks could exist.
What I find less likely are the odds of these rocks being blasted well out of Earth orbit, and impacting another planet or moon. The delta-e between Earth and Mars, let alone Earth and Europa, is pretty significant, and more importantly, space is really goddamn big. Yet even that stands up, under further consideration. They're claiming only 3 parts in a million hit Europa, which seems reasonable. I would expect far more rocks to have hit Mars or particularly the Moon, which brings up some interesting questions.
And arguably a better weapon anyways. Build a Chinese repeating crossbow - fast as a bolt-action rifle, greater range and accuracy than current plastic printed guns, and far, FAR less likely to explode in your face.
They're not going to keep it at $5. By this time next year it'll probably be above $40.
Seems a bit information-inefficient. Four characters for the year, but only one for the model specifier? That's the inverse of what's needed - there's a 60-100% improvement in performance per year, but the range between the lowest-end GPU and highest-end GPU is closer to 10,000%.
Try coming up with a better one. Invent a new naming convention that still hits all the same price points (let's say $150, $170, $200, $250, $300, $350, $450, and $600). And also accounts for releasing a new batch every year, maybe every other year.
Seriously, I tried once. I ended up in about the same place AMD is right now. Nvidia's naming is a bit wonky, partially because they've never been the clearest, partially because AMD just forced them to drop prices VERY abruptly, and partially because they're in the middle of a transition. AMD's is actually very clear, at least through the gaming card range (the 250 and below are odd, but they're also basically useless for gaming).
AMD's scheme right now is actually pretty easy.
The first number is the generation. We're on "2", even though they just started this new numbering scheme this year, but that's fine.
The next number is the "category". Best way to think of it is monitor resolution: _70 is for 1080p - you'll get 60FPS+ on most games at max settings, real killers may need a settings drop but you'll generally be fine. _80 is for 120Hz or 1440p monitors, and the _90 is for tri-monitor 1080p, 4K, or obscene multi-GPU rigs. And an _60 part is a lower-quality 1080p - think "high" or "medium", not "max".
An X suffix means it's the "full" part, the lack of an X means it's been binned in some way (reduced clockspeeds and/or some cores disabled). For example, the 290X has 44 "compute units", while the 290 has 40 at a slightly lower clockspeed. On the 270s, they're both 20 compute units, but the 270X is clocked about 10% higher.
Since both new consoles use AMD chips, it's worthwhile to compare to them. The PS4 is a bit weaker version of the 270, and the Xbox One is a slightly underclocked 260.
Nvidia's scheme is similar (add another 0 on the end for no reason, swap "Ti" for "X", and be generation 7 instead of 2), but they've complicated it right now by not rebadging old chips as new names. AMD's recent launches were basically "launch a new 9-tier chip, take all the old ones, up the clockspeeds, bump them down a tier and cut their prices accordingly". The 270s that just launched are essentially overclocked 7870s (think "180X").
Right now, Nvidia's lineup starts at the 650 and 650 Ti Boost (medium-end 1080p), 660 and 760 (high-end 1080p), 770 and 780 (1440p), and the 780 Ti (4K). Nobody's really sure whether they're going to launch more low-end 700-series parts. They're also looser with which ones are low bins of what - the 780 is a binned 780 Ti, but the 760 is a binned 770.
PS: Ignore the Titan. It's no longer a gaming card - the 780 Ti outperforms it (the Titan is a binned 780 Ti), at $300 less.
The other advantage of the Titan is the double-precision performance. Almost all of Nvidia's cards, including the 780 Ti, run double-precision floating-point calculations at 1/24th the rate of single-precision, but for the Titan and the Tesla pure-GPGPU cards, it's 1/3rd the rate.
While I'm not sure if that's an actual hardware difference, or if it's some software limitation, or a mix of both or whatever, it's definitely real. That's the main reason a Titan is still $1000 - it's being sold as a low-end compute card, not a high-end gaming card.
We can send them there, then give Gitmo back to Cuba.
Hard for them to blackmail us if anyone of rank is sitting in solitary confinement at Gitmo, awaiting military tribunal for treason.
Or dead.
Just putting that out there.
I browse everything with Adblock, just because it's saved my ass from malware injected into ads more than once. If I like a site sufficiently, I give it a chance. But if I see any of the following, I turn it back on: /. back on the block list
Flash
animated ads (no video - small animations are a maybe) - this is what got
political/religious ads
ads literally in the middle of the content
anything disguised as a download button, or disguised as anything but an ad
And I also tend to turn it back on if I never see relevant ads. For instance, I read a *lot* of webcomics. What's advertised a lot on webcomics? Why, other comics. It's fairly common for me to go weeks seeing only ads for comics I already read, or ones that I've checked out and decided I don't like. At that point, I often turn blocking back on, if it's even mildly irritating.
I've seen a car catch fire while parked.
Yes. The HHGG movie gets a lot of flak for not following the books, but it was a successful comedy, and it kept the general humor stylings of the books.
Really, I think it would have been worse if it had copied the books directly. Comedy relies heavily on surprise, the unexpected. If all of the jokes were ones that we had heard before, it would have fallen flat. In retrospect, the funniest parts were those that were new - or at least done differently than in the books (turning "So Long and Thanks For All the Fish" into a musical number, for example).
The Titan isn't positioned as a high-end gaming card as much as it is a low-end scientific computing card. It's the cheapest GPU that has reasonable double-precision floating-point performance. For whatever reason, most Kepler cards run DP operations at 1/24th the speed of single-precision, but the Titan and most of the Tesla cards are able to do so at 1/3rd the speed. There, the Titan runs thousands less than the similar Tesla cards (the K20 is listed on Newegg for $3500, and the K20X is on Amazon for $7700).
The fact that the Titan also gets some buys from gamers with way too much money is just a side bonus. Even since the 780 came out, it's been extremely wasteful to get a Titan for gaming. And Nvidia's own 780 Ti is likely to out-perform the Titan in games for $300 less. Really, I think the only reason they ever marked it as a gaming card was as a publicity stunt - they held the title of "fastest card ever" for quite a while, and they held it by an impressive lead.
Going faster than light in a vacuum would violate causality. Going faster than light from point A to point B is trivial if you can take a shortcut. For instance, via wormhole. Light following the same path you do would still beat you, but you still get there in a reasonable amount of time.
Yes, 1280x800 is a horribly small resolution. But at least it's a good aspect ratio. 4:3 is bad for entertainment use - movies, games, and the like. 16:9 is similarly weak in productive use - even putting two windows side-by-side, it's not tall enough, and rotating it to portrait mode is often laughable. But, IMO, 16:10 is a good compromise - it works well for anything you do with it.
If it weren't for the fact that 2560x1600 monitors are absurdly overpriced compared to 2560x1440, I'd have gotten one of those for my primary monitor instead.
You seem to be a bit lost. Gizmodo is over that way. Or perhaps you're looking for Wired?
They're enabling a lower power state on the WiFi chip. So if it's not being used, it can drop to a power state lower than the one it was using previously. All this would do is add extra wakeup time when you suddenly start network activity, and I can't imagine it would be enough extra time to be noticeable.
Silk Road failed through secondary methods. The actual mechanism was solid, it's just that DPR was sloppy while setting it up, enabling regular investigative work to find him. I will note that we haven't heard of Silk Road clients being arrested en masse, which means they probably haven't been able to track down many of them. They caught the operator, shut down the site, and seized a large amount of money, but the fact that they haven't been arresting left and right the drug dealers who used it means they aren't getting much information from it.
Basically, all someone needs to do is do the exact same thing Silk Road did, except avoiding leaking information on other bands. Use a secure email, don't even think about the site on unencrypted channels, establish a completely new identity for running the site, etc.
A self-fulfilling prediction. Once it was accurate for a few iterations, it moved from being a prediction to being a target, and engineers at every major foundry have worked to meet or even exceed it.