There are compilers that do this, e.g. Intel Fortran, but
for some reason you lose precision
when you activate that feature.
Lost precision can mean wrong results in this case.
Python and believe it or not, Fortran (I guess a lot of fluid dynamics types use Fortran still for some reason).
There are three reasons we still use Fortran in the CFD business. First, a lot of good old numerics code is written in Fortran, and interfacing between languages means overhead. (You see, we're the types that define and use onethird=1.0/3.0 if we have to divide by 3 more than 10 times in a tight loop since multiplication is faster than division, or loop over j in the outer and i in the inner loop because that's how arrays are stored...)
Second, for the type of stuff we normally do, Fortran is 10-20% faster than C and orders of magnitude faster than other languages. (C is faster at file I/O.) This is important when you measure runtimes in weeks. (With Python, the simulations I did for my thesis would literally take years.)
Third, there is still significant work by large companies to create even more efficient Fortran compilers (see Intel, PGI, NAG).
I think you should maybe check your math again. Kaliski used to work at RSA, and the quote above is from "Handbook of Information Security" by H. Bigdoli, page 577. I think they checked the math twice before putting it in print.
Yeah, when the US government learns what encryption is used, they'll crack your code in no less than it the time it takes to brute-force a 128 bit secret-key cipher. (Curve25519 is the algorithm used.) Seriously, learn some cryptography before commenting on cryptography. The following is Burt Kaliski's description of how hard it is to crack this encryption:
Imagine a computer that is the size of a grain of sand that can test keys against some encrypted data. Also imagine that it can test a key in the amount of time it takes light to cross it. Then consider a cluster of these computers, so many that if you covered the earth with them, they would cover the whole planet to the height of 1 meter. The cluster of computers would crack a 128-bit key on average in 1,000 years.
I've heard this "they'll scan my harddrives" a lot, but I don't think it will happen. I know quite a few industrial researchers who visit the US every now and then, and their laptops have 1TB harddrives that are fully encrypted. As long as there are people who want to travel to/from the US carrying proprietary business information, harddrive inspections will never happen.
To clarify: I meant a doubling every $time_period of the {average kWh produced in one year} per {dollar cost of the solar panel}. This would mainly happen because they get cheaper, yes.
This point may come one day, but for the foreseeable future (say 5 years), all advances in battery technology will be negated by consumers (and operating systems) demanding ever more from the processor, display, etc. The only way around this is to stop demanding so much from your phone.
Case in point: My LG Optimus One has a rated standby-time of 550 hours on 3G, Android 2.3. It's got an old ARMv6 processor running at around 0.4 GHz. When I don't use the wifi continuously, I get roughly 3 days between charges. I've seen people on the xda-forums underclocking it and removing bloatware etc. get battery life up to 6 days. You can't have your cake (newest iPhone/Galaxy with tons of apps and constant 3G) and eat it too (weeklong battery life)!
The real issue is that even if everyone agreed
on global warming, we haven't
the foggiest realistic solution.
All we can do is pray for photovoltaics
to follow Moores law for another decade,
or a breakthrough in fusion.
That's why people prefer to argue
whether there is a problem,
rather than admit that they're powerless.
I also study the Cosmic Microwave Background that seems to have a cold spot in it (Source: Through the Wormhole with Morgan Freeman). The cold spot is potentially another universe's gravity pulling on our universe.
I should point out that there is significant scientific debate on this point. Roger Penrose and colleagues claim that the rings in the CMBR could be evidence of multiverses, while others (e.g. Hajian, Wehus, Moss, etc.) claim that such rings are found in completely random, simulated CMBR data as well.
there is still a niche for traditional supercomputers left...
I don't think you know much about supercomputers. Sure, there are a few problems that are embarrassingly parallel, but most aren't. For those that aren't, the bandwidth and latency of the interconnect between different processors is more important, and often more expensive, than the speed of the processors themselves. This is why most supercomputers use exotic interconnects like Infiniband, Myrinet or 10GigE, and linking nodes together using complex topologies such as a torus.
Case in point: on the website of the QCDOC supercomputer, which was partially used in this study, they say that a highly optimized lattice QCD simulation achieves up to 50% CPU utilization, and this is considered very good. The rest of the time is mainly spent waiting for the interconnect.
Cloned your code and added a classical fourth-order Runge Kutta. Here. There isn't really any difference, presumably because you used 100 000 time steps. The difference between Euler and RK4 is tiny for non-stiff problems and small time steps.
So I guess either your interpretation is wrong, or he is wrong.
If you've ever tried to
install a compiler on OS/X,
you would know why it's not a proper
*nix. Fucking 2GB install!
That's like calling a Honda Civic
a drag racer. You can turn it into one,
but it's a lot of work.
As has been pointed out in earlier/. submissions, the logical conclusion of "The War On Piracy" is to ban the all-purpose computer. Only computers incapable of running software not from a MAFIAA approved app store should be allowed. What's that, you say, you have "research labs" and "universities" where they "write software themselves"? BAN THEM AS WELL! All those college kids are pirates, anyway.
Have you given Dolphin (the default KDE file manager) a shot? The only two things I see on the DirectoryOpus feature list that Dolphin doesn't have is CD/DVD burning and folder synchronization, but there are other good Linux tools for this. k3b (burning) and grsync (synchronization) come to mind.
However, after using Linux for a while, you will come to realize that file managers are unneccesary. If you learn to use the terminal (bash or preferably zsh) along with grep, find and friends, you will be able to do magic that leaves any file manager looking dumbstruck.
I assumed from your nick that you knew what upstream means. What I meant is that the bug should not be filed with Ubuntu (Launchpad), it should be filed with the audio subsystem of the kernel. Also, Ubuntu updates the kernel infrequently, especially if you're still on 10.04 or some other old version, so it could be that the bug is fixed in a newer kernel than yours.
This sounds like a driver bug, in which case you shold file a bug upstream. Both my wife's laptop (Ubuntu 12.04) and my laptop (Archlinux) do this without problems, plus you can turn this feature on/off in alsamixer.
Agreed, they don't prepare us very well. On the plus side, though, I learned to program on my graphic calculator, (Casio FTW!) in a weird cross-breed of apl and basic. GOTO FTW.
This is funny. Here in Norway, we all had to get graphing calculators in high school, but in university we're not allowed anything more advanced than a scientific calculator (+-*/, sine, cosine, etc.). Graphing and programmable calculators are expressly banned. As a consequence, I can freehand draw most functions you'll ever need.
I have quite a few useful little Octave scripts that would be neat to run on my phone. For instance RK4-integration of Newton's second law; just input the body forces, starting position and velocity, hit run, it calculates body trajectory in 2D or 3D as you wish. Or maybe you want to calculate the eigenvalues of some Hamiltonian you just thought up?
Slashdot Mirror
This should be modded (+6, TFA is wrong and sensationalist).
Oh, if only I had mod points for you...
I know you can make a regexp be greedy, but I had no idea it could be puritan as well! Perl example, please?
There are compilers that do this, e.g. Intel Fortran, but for some reason you lose precision when you activate that feature. Lost precision can mean wrong results in this case.
Python and believe it or not, Fortran (I guess a lot of fluid dynamics types use Fortran still for some reason).
There are three reasons we still use Fortran in the CFD business. First, a lot of good old numerics code is written in Fortran, and interfacing between languages means overhead. (You see, we're the types that define and use onethird=1.0/3.0 if we have to divide by 3 more than 10 times in a tight loop since multiplication is faster than division, or loop over j in the outer and i in the inner loop because that's how arrays are stored...)
Second, for the type of stuff we normally do, Fortran is 10-20% faster than C and orders of magnitude faster than other languages. (C is faster at file I/O.) This is important when you measure runtimes in weeks. (With Python, the simulations I did for my thesis would literally take years.)
Third, there is still significant work by large companies to create even more efficient Fortran compilers (see Intel, PGI, NAG).
I think you should maybe check your math again. Kaliski used to work at RSA, and the quote above is from "Handbook of Information Security" by H. Bigdoli, page 577. I think they checked the math twice before putting it in print.
Imagine a computer that is the size of a grain of sand that can test keys against some encrypted data. Also imagine that it can test a key in the amount of time it takes light to cross it. Then consider a cluster of these computers, so many that if you covered the earth with them, they would cover the whole planet to the height of 1 meter. The cluster of computers would crack a 128-bit key on average in 1,000 years.
I've heard this "they'll scan my harddrives" a lot, but I don't think it will happen. I know quite a few industrial researchers who visit the US every now and then, and their laptops have 1TB harddrives that are fully encrypted. As long as there are people who want to travel to/from the US carrying proprietary business information, harddrive inspections will never happen.
What are you doing writing sensible comments on an article whose headline includes the words "Oracle" and "Cloud"? I came here for the lulz, dammit!
To clarify: I meant a doubling every $time_period of the {average kWh produced in one year} per {dollar cost of the solar panel}. This would mainly happen because they get cheaper, yes.
This point may come one day, but for the foreseeable future (say 5 years), all advances in battery technology will be negated by consumers (and operating systems) demanding ever more from the processor, display, etc. The only way around this is to stop demanding so much from your phone.
Case in point: My LG Optimus One has a rated standby-time of 550 hours on 3G, Android 2.3. It's got an old ARMv6 processor running at around 0.4 GHz. When I don't use the wifi continuously, I get roughly 3 days between charges. I've seen people on the xda-forums underclocking it and removing bloatware etc. get battery life up to 6 days. You can't have your cake (newest iPhone/Galaxy with tons of apps and constant 3G) and eat it too (weeklong battery life)!
The real issue is that even if everyone agreed on global warming, we haven't the foggiest realistic solution. All we can do is pray for photovoltaics to follow Moores law for another decade, or a breakthrough in fusion. That's why people prefer to argue whether there is a problem, rather than admit that they're powerless.
I also study the Cosmic Microwave Background that seems to have a cold spot in it (Source: Through the Wormhole with Morgan Freeman). The cold spot is potentially another universe's gravity pulling on our universe.
I should point out that there is significant scientific debate on this point. Roger Penrose and colleagues claim that the rings in the CMBR could be evidence of multiverses, while others (e.g. Hajian, Wehus, Moss, etc.) claim that such rings are found in completely random, simulated CMBR data as well.
there is still a niche for traditional supercomputers left...
I don't think you know much about supercomputers. Sure, there are a few problems that are embarrassingly parallel, but most aren't. For those that aren't, the bandwidth and latency of the interconnect between different processors is more important, and often more expensive, than the speed of the processors themselves. This is why most supercomputers use exotic interconnects like Infiniband, Myrinet or 10GigE, and linking nodes together using complex topologies such as a torus.
Case in point: on the website of the QCDOC supercomputer, which was partially used in this study, they say that a highly optimized lattice QCD simulation achieves up to 50% CPU utilization, and this is considered very good. The rest of the time is mainly spent waiting for the interconnect.
http://abstrusegoose.com/281
Cloned your code and added a classical fourth-order Runge Kutta. Here. There isn't really any difference, presumably because you used 100 000 time steps. The difference between Euler and RK4 is tiny for non-stiff problems and small time steps.
So I guess either your interpretation is wrong, or he is wrong.
If you've ever tried to install a compiler on OS/X, you would know why it's not a proper *nix. Fucking 2GB install! That's like calling a Honda Civic a drag racer. You can turn it into one, but it's a lot of work.
If that's all she does, she can get an Asus/whatever at 1/4 the price of a Mac.
As has been pointed out in earlier /. submissions, the logical conclusion of "The War On Piracy" is to ban the all-purpose computer. Only computers incapable of running software not from a MAFIAA approved app store should be allowed. What's that, you say, you have "research labs" and "universities" where they "write software themselves"? BAN THEM AS WELL! All those college kids are pirates, anyway.
Have you given Dolphin (the default KDE file manager) a shot? The only two things I see on the DirectoryOpus feature list that Dolphin doesn't have is CD/DVD burning and folder synchronization, but there are other good Linux tools for this. k3b (burning) and grsync (synchronization) come to mind.
However, after using Linux for a while, you will come to realize that file managers are unneccesary. If you learn to use the terminal (bash or preferably zsh) along with grep, find and friends, you will be able to do magic that leaves any file manager looking dumbstruck.
I assumed from your nick that you knew what upstream means. What I meant is that the bug should not be filed with Ubuntu (Launchpad), it should be filed with the audio subsystem of the kernel. Also, Ubuntu updates the kernel infrequently, especially if you're still on 10.04 or some other old version, so it could be that the bug is fixed in a newer kernel than yours.
This sounds like a driver bug, in which case you shold file a bug upstream. Both my wife's laptop (Ubuntu 12.04) and my laptop (Archlinux) do this without problems, plus you can turn this feature on/off in alsamixer.
Agreed, they don't prepare us very well. On the plus side, though, I learned to program on my graphic calculator, (Casio
FTW!) in a weird cross-breed of apl and basic.
GOTO FTW.
This is funny. Here in Norway, we all had to get graphing calculators in high school, but in university we're not allowed anything more advanced than a scientific calculator (+-*/, sine, cosine, etc.). Graphing and programmable calculators are expressly banned. As a consequence, I can freehand draw most functions you'll ever need.
I have quite a few useful little Octave scripts that would be neat to run on my phone. For instance RK4-integration of Newton's second law; just input the body forces, starting position and velocity, hit run, it calculates body trajectory in 2D or 3D as you wish. Or maybe you want to calculate the eigenvalues of some Hamiltonian you just thought up?