Addition is relatively light bignum math if the precision you're extending from is correctly rounded. GPU floating point usually isn't. Even if it is, you're usually talking about increasing your op count by a factor of 5 or more, which kind of blows the "4x performace speedup" out of consideration.
FFTs inherently work in the well-behaved arena of simple trig functions and three-function (add/subtract/multiply) math.
add/subtract/multiply math is the area that 754 has had the biggest effect on - in fact, the spec has very little to say about transcendental functions, but is almost entirely concerned with the basic arithmetic ops. prior to 754, floating point was, in general, not algebraicly closed under +-*/, nor were the results correctly rounded.
most highly parallel GPU-type chips lack support for gradual underflow, for example, one of those "ill-defined corners of the number space" where 754 has been a tremendous boon. flush-to-zero is fine if you're decoding MP3s or unpacking texture maps, but it causes a lot of problems when you start trying to do more general scientific computations. sometimes those low order bits matter a whole lot; sometimes they're the difference between getting an answer accurate to 4 digits and an answer with *no* correct digits.
"simple trig functions" have their own problems on these architectures; try writing an acceptable range-reduction algorithm for sin or cos without having correctly rounded arithmetic ops. sin and cos are, in fact, two the hardest operations in the math lib on which to get acceptable accuracy.
admittedly, none of these objections are an issue with FFTs. but the reason that FFTs will perform acceptably on such an architecture is that the operations are (usually) constrained to the domain in which you don't encounter the problems i mention, not because the operations themselves are inherently safe. the lack of support for gradual underflow will cause you to lose many, many bits in frequency components that have nearly zero magnitude, but you usually don't care about those components when you're doing FFTs, anyway.
My point was, and still is, Apple is part of a larger consortium that has betrayed the consumers.
If people didn't want the product, then they wouldn't buy it. Last I checked, iTMS was doing brisk business, ergo the consumers want the product, and Apple is giving it to them. Can you explain to me where "betrayal" figures in to that?
You may not like DRM; I don't like DRM. You may not like 160kbit AAC; I don't especially care for it either, but I find it listenable. But Apple failing to satisfy your personal version of "what the market should look like" hardly constitutes "betrayal of the consumers".
The dictionary definition of "monopoly" is NOT the legal definition of the word. This happens all the time, in every discipline:
To a mathematician, a "ring" isn't a piece of jewelry.
To a physicist, a "string" isn't a long piece of twisted interlocked fibers.
To a programmer, a "float" isn't a thing you attach to a fishing line.
To a lawyer, a "monopoly" isn't "a persistent market situation where there is only one provider..."
Assuming that the legal and dictionary definitions of "monopoly" coincide is like assuming that a float denotes a variable that does not sink in water. A market situation can definitely be a monopoly without meeting the dictionary/wikipedia/economics/whatever meaning of the word. I mean, it's not a board game either, but I don't see you commenting on that.
yes, but the nice thing is that for most encryption methods, the work to do the encryption grows linearly (at worst polynomially), whereas the work to break the encryption grows exponentially in key size. the larger the key gets, the bigger the gap between work to encode and work to decode.
and that would exert a pressure for AH prices to balance across servers... this is a bad thing?
Re:Undead Stratholme (with your Boss)
on
WoW the Next "Golf"?
·
· Score: 2, Informative
Baron: boss in an early endgame instance. Very easy to kill in a 10-player group, somewhat harder (though still easy to kill with good gear) in a 5-player group. It's actually much harder, imo, to get to him than it is to kill him as a group of five. You don't need to do anything fancy to take him down, just hit him as hard as you can and keep everyone alive.
i just got one in early december. am very happy with it. no first-gen foibles, technicians know how to diagnose and fix the damn things by now if something goes wrong, apple's had time to work out the kinks; you couldn't *give* me one of the new macbooks in exchange. besides, the powerbook runs more than fast enough for my portable computing (mostly writing code and running small simulations - i'm a numerical analyst). if anyone actually needs real computing speed (say, for *large* simulations), they're not (i hope) doing it on a laptop, anyway...
IANAL, but amendment XIV applies only to state laws, not federal. the federal government is free to grant protections to "journalists" and withold those protections from others, under this amendment. so the 14th amendment has no bearing on the issue at had.
that said, the 5th amentment has been interpreted by the supreme court as enforcing equal-protection guarantees on federal law, as well.
"very surprised" doesn't begin to touch on it. even assuming 720p HD content, you're talking about moving a sustained 83Mb/sec of data over the bus (the decoded video), never mind the issue of needing to read *in* the compressed data, or the actual process of decoding it.
Actually, this fits very tidily in a Chomskian context. The program has an internal, predetermined notion of "what a grammar looks like" (i.e. a class of allowable grammars sharing certain properties), and adapts that to the source text. The way all this is presented makes it seem like unsupervised learning that can find any pattern, but the best you can hope to do with a method like this is capture an arbitrary (possibly probabilistic) context free grammar (CFG).
Even then, Gold showed a long, long time ago (1967) that the task of inducing an arbitrary CFG using only generated strings from the language is basically hopeless [Gold, E. Mark. 1967. Language Identification in the Limit. Information and Control, 10:447-474].
That said, this doesn't even seem to be that novel (to me). Andreas Stolcke wrote a very nice PhD dissertation in 1994 on learning arbitrary PCFGs from langage strings [Stolcke, Andreas. 1994. Bayesian Learning of Probabilistic Language Models. PhD Dissertation. University of California at Berkeley.]
This is probably a better, more efficient method that Stolcke produced back in '94, but I would be *very* surprised if it revolutionized the way computers interact with language, or anything else of the sort. People working in computational linguistics have a nasty habit of making grand pronouncements, only to fall far short of what they claimed.
For the record: IANAL, but i play one on TV, by which i mean i'm an applied mathematician with a couple published papers in computational linguistics.
In general, dust is extremely flammable - anything with that much surface area is. On the other hand, there's not very much of it (mass-wise), and so it'll just flash burn in a split second, and usually wouldn't have the chance to light less-flammable materials on fire.
yeah, and even if it could display 1280x1024, it still wouldn't be proper HD (1920x1080). so many people get ripped off by "high definition" stuff these days.
you make the mistake of comparing BART to a subway system; it's not. it's really more of a commuter rail, which it's fairly successful at. you should be comparing it to NJ transit, LIRR, Metro North, the Boston commuter rail system, etc.
Now if only they'd make MUNI go somewhere useful...
if i punch you in the face, and that causes pain, the pain is YOUR reaction, not imposed on you. it's in your mind. pain is not a physical substance. just like your reaction when your best friend calls you a jerk.
Slashdot Mirror
Addition is relatively light bignum math if the precision you're extending from is correctly rounded. GPU floating point usually isn't. Even if it is, you're usually talking about increasing your op count by a factor of 5 or more, which kind of blows the "4x performace speedup" out of consideration.
most highly parallel GPU-type chips lack support for gradual underflow, for example, one of those "ill-defined corners of the number space" where 754 has been a tremendous boon. flush-to-zero is fine if you're decoding MP3s or unpacking texture maps, but it causes a lot of problems when you start trying to do more general scientific computations. sometimes those low order bits matter a whole lot; sometimes they're the difference between getting an answer accurate to 4 digits and an answer with *no* correct digits.
"simple trig functions" have their own problems on these architectures; try writing an acceptable range-reduction algorithm for sin or cos without having correctly rounded arithmetic ops. sin and cos are, in fact, two the hardest operations in the math lib on which to get acceptable accuracy.
admittedly, none of these objections are an issue with FFTs. but the reason that FFTs will perform acceptably on such an architecture is that the operations are (usually) constrained to the domain in which you don't encounter the problems i mention, not because the operations themselves are inherently safe. the lack of support for gradual underflow will cause you to lose many, many bits in frequency components that have nearly zero magnitude, but you usually don't care about those components when you're doing FFTs, anyway.
The dictionary definition of "monopoly" is NOT the legal definition of the word. This happens all the time, in every discipline: To a mathematician, a "ring" isn't a piece of jewelry. To a physicist, a "string" isn't a long piece of twisted interlocked fibers. To a programmer, a "float" isn't a thing you attach to a fishing line. To a lawyer, a "monopoly" isn't "a persistent market situation where there is only one provider..." Assuming that the legal and dictionary definitions of "monopoly" coincide is like assuming that a float denotes a variable that does not sink in water. A market situation can definitely be a monopoly without meeting the dictionary/wikipedia/economics/whatever meaning of the word. I mean, it's not a board game either, but I don't see you commenting on that.
yes, but the nice thing is that for most encryption methods, the work to do the encryption grows linearly (at worst polynomially), whereas the work to break the encryption grows exponentially in key size. the larger the key gets, the bigger the gap between work to encode and work to decode.
and that would exert a pressure for AH prices to balance across servers... this is a bad thing?
Baron: boss in an early endgame instance. Very easy to kill in a 10-player group, somewhat harder (though still easy to kill with good gear) in a 5-player group. It's actually much harder, imo, to get to him than it is to kill him as a group of five. You don't need to do anything fancy to take him down, just hit him as hard as you can and keep everyone alive.
i just got one in early december. am very happy with it. no first-gen foibles, technicians know how to diagnose and fix the damn things by now if something goes wrong, apple's had time to work out the kinks; you couldn't *give* me one of the new macbooks in exchange. besides, the powerbook runs more than fast enough for my portable computing (mostly writing code and running small simulations - i'm a numerical analyst). if anyone actually needs real computing speed (say, for *large* simulations), they're not (i hope) doing it on a laptop, anyway...
sure they do. powerbooks have DVI-out. and a reasonably good graphics card to run it. i have mine plugged into a 24" dell LCD.
that's 83MB/sec, not Mb. it's just not my day....
IANAL, but amendment XIV applies only to state laws, not federal. the federal government is free to grant protections to "journalists" and withold those protections from others, under this amendment. so the 14th amendment has no bearing on the issue at had.
that said, the 5th amentment has been interpreted by the supreme court as enforcing equal-protection guarantees on federal law, as well.
"very surprised" doesn't begin to touch on it. even assuming 720p HD content, you're talking about moving a sustained 83Mb/sec of data over the bus (the decoded video), never mind the issue of needing to read *in* the compressed data, or the actual process of decoding it.
Actually, this fits very tidily in a Chomskian context. The program has an internal, predetermined notion of "what a grammar looks like" (i.e. a class of allowable grammars sharing certain properties), and adapts that to the source text. The way all this is presented makes it seem like unsupervised learning that can find any pattern, but the best you can hope to do with a method like this is capture an arbitrary (possibly probabilistic) context free grammar (CFG).
Even then, Gold showed a long, long time ago (1967) that the task of inducing an arbitrary CFG using only generated strings from the language is basically hopeless [Gold, E. Mark. 1967. Language Identification in the Limit. Information and Control, 10:447-474].
That said, this doesn't even seem to be that novel (to me). Andreas Stolcke wrote a very nice PhD dissertation in 1994 on learning arbitrary PCFGs from langage strings [Stolcke, Andreas. 1994. Bayesian Learning of Probabilistic Language Models. PhD Dissertation. University of California at Berkeley.]
This is probably a better, more efficient method that Stolcke produced back in '94, but I would be *very* surprised if it revolutionized the way computers interact with language, or anything else of the sort. People working in computational linguistics have a nasty habit of making grand pronouncements, only to fall far short of what they claimed.
For the record: IANAL, but i play one on TV, by which i mean i'm an applied mathematician with a couple published papers in computational linguistics.
In general, dust is extremely flammable - anything with that much surface area is. On the other hand, there's not very much of it (mass-wise), and so it'll just flash burn in a split second, and usually wouldn't have the chance to light less-flammable materials on fire.
yeah, and even if it could display 1280x1024, it still wouldn't be proper HD (1920x1080). so many people get ripped off by "high definition" stuff these days.
you make the mistake of comparing BART to a subway system; it's not. it's really more of a commuter rail, which it's fairly successful at. you should be comparing it to NJ transit, LIRR, Metro North, the Boston commuter rail system, etc.
Now if only they'd make MUNI go somewhere useful...
well said. it's not so much that i disagreed with the grandparent, as i just wanted him/her to be more precise. your response addresses that nicely.
if i punch you in the face, and that causes pain, the pain is YOUR reaction, not imposed on you. it's in your mind. pain is not a physical substance. just like your reaction when your best friend calls you a jerk.
nah, not really:
http://ed.sjtu.edu.cn/rank/2004/top500(1-100).htm
even the british don't think so:
http://www.thes.co.uk/statistics/
is oxford very, very good? yes. i'm visiting now, and it's a great place to be. commonly held to be the finest in the world? that's a stretch.
but at least he's not drunk anymore!
They also didn't build a plane for their competition...
If that were the case, it would be 67, not 66.
Ok, so I'm being a smartass.
strange, i haven't noticed the sound of my computer (mac) in months.
what's the point of modding an AC a troll?
it's no fun unless it burns your karma.
Actually, it's pretty easy (to judge by video) and non warranty-voiding, though you will need to pay to repair any damage you do in the process.