OK, I've waited in vain for clarifcation, so I have to ask. Why, again, exactly, is he a moron for having the beliefs that he does? Or, is it, perhaps, that you are the moron for failing to back up (or rescind) your attack?
Leakage current is a function of the transistor design only (mostly, the resistance of the gate oxide). Frequency of operation has no impact, since this current is cause by static (non-switching) current from voltage at the gate of the transistor. Dynamic (switching) power increases linearly with frequency, because every clock edge charges (or discharges) the capacitor that is the formed by the transistor gate.
Think of leakage (always-on) current as a leaky faucet. In this analogy, the switching current would be the force, or energy required to fill up (or empty) a pool. The whole while the pool fills or empties, leakage current drips precious water on the ground -- this wasted leakage does no work for your pool, but leaves water (heat) to be dealt with.
That said, the leaks in the faucet, so to speak, in any processor you have in your box now (0.13um and bigger) are still pretty negligible, and there's not much you can do about it anyway (except lower voltage, which makes over-clocking harder, but fits well with under-clocking). Most of your power (and heat) comes from switching current, so yes, lowering your clock speed will lower the heat dissipation requirements of your box. The reduction will be linear with clock rate, so 1/2 GHz = 1/2 power.
Oh, you're going to cover the licensing fees for GAIM? Cool! Thanks! Hey guys, didja hear that? Gherald is covering the MS license fees for GAIM to keep working with Messenger . . .
Uh, patents (the good kind)? Licensing fees? Higher profit margins from early-adopter sales? A head start on everyone? They know what they're doing (business wise) and are the best-prepared to do it technology-wise (though I have my doubts here), but they still probably won't get to try, not with the cash their asking from a tired economy that's been in recession 10x as long as the US has.
I don't believe strongly in anything except that in the freedom of being what you feel you are and expressing yourself in a way that does not physically hurt anyone.
You're preaching to the choir here; I heavily depend on the "do as thy will while harming none" way of thinking, but let me ask you this: did you ever consider that posibly, just maybe, there are some forms of harm you cannot anticipate accurately? Given that humans are complex (beyond our understanding, mostly) and that societies of humans are even more complex, how can you ever be sure that your (in-)action causes no harm?
If so, you'd know that both "hara-kiri" and "seppuku" are the same, with no distinction whatsoever in Japanese, not even in politeness levels, as is oft misreported.
In fact, they are Japanese palindromes. I'll spare you the mess of ASCII-art Kanji, but there is one character that can be read as "hara" or "se" depending on context, and another that is read as both "kiri" and "puku". Put them together one way, they are pronounced "hara kiri"; swap them and they are pronounced "seppuku", but the mean exactly the same in both cases (at least in Tokyo and Awaji, maybe in other areas there is a difference? But I don't think so.)
Besides, the original poster said "hari kari", and that's hard to confuse for "seppuku". I think he was going for "hara kiri":)
But how do you get "micron high" little gold studs to stick to the die in exactly the right places? How do you make sure each gold stud is exactly the same height (can't have a short one anywhere, even by a femto-meter)? Then, how do you physically/mechanically line them up exactly and keep them together perfectly for long priods of time under fairly wide ranges in vibration and temperature ranges? How do you prevent the dice from warping if each stud isn't 100% identical (such as if you try to tolerate some height variation by making the studs slightly compressible)?
Since you're using area-IO to connect the dice, how do you power them? Usually, in area-IO die the power comes from the top (like an IO buffer), but in a stacked die this would just lead to another die. You can't power from the "bottom", since that's not metal (it's Si substrate), and you need really big power wires to get all over the dice from somewhere. If from the top, do the tiny wires shoot out the sides between the two dice and then go to a power, uh, plug? Connector with tiny wires on one side, or what? And I do mean tiny -- the little metal studs would need to be placed every 200-300um apart, in a 2-D array, and some would have to connect to a power source, somehow.
If you got this far: how do you design a chip with identical, but mirror-image IO locations of another chip, which presumably does a different thing? It's a huge battle in system design these days to get a chip package pinout that makes both the PCB designers and the die designers happy. Making 2 die designers happy with one chipl-level pinout would be impossible.
Finally, given that gross simplifications and assumptios about near-perfect isolation used in modern chip design, how long until you can have the hardware and software ready for me to be able to calculate iterative solutions to 500 million simultaneous, co-dependent variations on Schroedinger's Wave Equation? (of course, it's intractable, so iteration is your only hope -- better pray for convergence too!) Oh, and I need that to take less than 6-12 hours each run to make a reasonable design schedule.:)
Right on a). well, mostly -- IBM has a new process that does allow transistors in some area-IO to be placed over logic gate transistors. It's more trouble than it's worth, though (unavoidable interactions are hard to calculate accurately).
And right on b) -- the distance between 2 dice stacked is much shorter than 2 side-by-side. But this is totally irrelevant, mostly due to previous posters:). See, it's not that it's further to go vertical from one die to the next, rather than packaging each individually and connecting them horizontally. The problem is it's hard to go vertical. This is true from design, manufacturability, and reliability points-of-view.
First, by area-IO I meed input/output (IO) drivers or receivers that can be placed anywhere in an area, rather than only around the circumference (preipheral IO). We have area-IO at the package level (such as BGA, or Ball-Grid-Array and FCBGA, or Flip-Chip Ball Grid Array [best for area-IO, and expensive]) and area-IO at the die level. Do we connect the dice before or after packaging?
Either way presents problems. Such as (for pre-packaging connections):
How do you electrically connect 2 area-IO dice? Usually, a die has little square landing pads, and these are only about 50um square, spaced every 200-250um or so on center in 2-D arrays of up to 70x70 and more. To be able to do anything with these tightly packed little signals, we drop special tiny drops of metal that stick to the pads, and press this up against a package substrate (ceramic), which includes routes to space those signals out more, like every 1.0mm or so. Even this is expensive and hard to mount to PCB, since it's hard to ensure both things are perfectly flat (package and PCB) so that all balls connect.
In fact, we rely on the package (often including an internal metal "stiffener") to keep the die nice and flat, which helps avoid de-lamination (layers peeling apart). Two dice pressed next to each other would require some space between them to make the connection (i.e., some bumps for the connection, and valleys for no connect areas), and this and the elasticity of the electrical connection medium would leave enough play to let the dice warp all over the place.
It'd be even harder to tell which ball(s) aren't connected. We do this now by confirming that the PCB is OK (usually pretty easy, so it makes a good reference), make the chip send specially-controllable data out (and take data in on inputs), then check to see what's right and wrong by measuring at the board level. If my board is another chip, how do I know which one I am debugging? This debugging (we call mfg testing) happens to all chips, not just some samples. If it isn't, failure rates will go up to unacceptable levels (like 20-50% or more).
Testability is hard if you stack dice before or after packaging. Design is a bizzotch too, since you can't very well even model one whole chip at a time (and how the circuit performs depends on process, voltage, and temperature), much less two chips stacked with an insulator and some kind of very short, very small, very fragile, very susceptible to noise and crosstalk hunk of 1000+ wires between them. One local hot spot at X,Y on die A can mess up operation at x,y on die B, and we'd never be able to practically predict that.
Most importantly of all, part of the reason chip design even works at all, and that we can churn them out for pennies each (after massive design and capital outlay for a fab), is that we can simplify the design dramatically by making assumptions, modelling the target device in isolation, verifying it in isolation, and then being able to safely assume this (truly wrong) assumption of isolation is close enough to true that the part will work in the system. Single packaged die are relatively infinitely insulated from everything except the I/O we carefully design. Stacked dice would not be -- they would interact strongly with each other in unpredicatab
I don't necessarily agree or disagree with any of you, but I do have to take issue with the ever-increasing tendency of (all kinds of) people to disparage anyone with firm personal beliefs. It seems like the only way to avoid this is to have completely relativity-based ethics and morals (no sure right or wrong -- "everything is relative"), and to adamantly avoid stating something one believes in firmly. I think this is a Bad Thing.
See, vacillating and changing your "beliefs" in order to appease or not offend others, IMHO, is disingenuous and dangerous. It's also subversive, and really no different than lying. Even failing to form an opinion and expressing it in the proper forum is somewhat irresponsible and lazy, IMHO. Such chameleons/chickens change their views (or at least downplay them) as is convenient, often to win support, avoid dissent, or just to be agreeable. Worse, it really seems to upset some of them when someone else expresses his or her own firm beliefs.
Frankly, I think the GP post was a non-flamebait, non-troll expression of the poster's beliefs. Just because you disagree doesn't make him a moron.
And about the only way to do this without sacrificing clockrate is by going to a smaller fabrication process.
Sorry, that's commonly believed, but wrong. There are lots of ways to reduce power consumption. Reducing gate widths (0.25um -> 0.13um -> 90nm) is commonly touted as a good way to reduce power, but in most cases that's more marketing pitch than reality.
First, there are two types of chip power to worry about (1) leakage, which happens all the time, just by being on, and which used to be always much much lower than (2) the switching power, or maximum dissipation when as many transistors as possible can switch at once (which, BTW, can never be all of them, and it's really, really hard to find the stimulus that makes maximum power happen. So, esitmates like the ones in the article for peak power are often made assuming a somewhat-arbitrary switching factor that may be low or high).
As gate sizes shrink, the effective capacitance of the gate shrinks, and voltage can be lowered (to a point). Capacitance varies with gate area and inversely with distance between "plates" of the gate (C = k*A / d). Reducing the gate width (space between the plates) actually increases capacitance, and this itself would increase power. But, you're also able to reduce the gate area (though not as much, but in 2-dimensions, so shrinking gates is usually a reduction in C). Most importantly, you can decrease voltage, since power varies with the square of voltage, this has much more impact on power than reducing gate capacitance (size). When we went from 0.25um (3.3V)to 0.13um (1.5V), we got a nice fat 1.8V drop in voltage. But 0.13um is 1.5V too, or 1.3V at best, and I've never heard of a 90nm (0.09um) process under 1.1V. The V isn't dropping as fast any more because the noise margins are getting too small.
Since p(switching) = 1/2*F*C*V^2 (F = clock freqyency, C = capacitance, and V = max voltage, lowering C (and moreso V which we can reduce some, but not much below 1.0V so far) will lower power a bit. Linearly with C. But unless we can reduce V, reducing C much more won't help a lot because we have more total C's (transistor gates) on the die, because they are smaller we can fit more.
But now, at 0.13um, and more at 90nm, it's not the switching power, but the leakage (always there) power that's getting worrisome. It used to be 1/20th of switching power or less, but now the gates are so small current of the same order of magnitude (almost) of switching leaks all the time.
So, the more you shrink, the more you have constant power, which is harder to deal with since you can't throttle it, and it's always cranking out. Worse yet, the more you shrink, the more gates you can fit on one tiny little die (the feasible mfg'able die size stays around 17-18mm max regardless of gate size once the process matures a bit, but bigger dice have ridiculous failure rates and thus silly high prices). And the gates shrink in 2 dimensions (L and W), so you get a squaring increase of the toal gate count, and only a linear decrease with C. Shrinking gates to save power doesn't work.
So, if we can't keep shrinking to save power, how can we? Lot's of ways. There are dozens of EDA companies with power-minded RTL coding, synthesis, and even place and route tools ready to help you reduce your power if you have a few $100k/seat/year. Or, you could use a SSC (Spread-spectrum Clock, where each clock edge is off by a bit to reduce power, but it slows down the max clock rate a bit too, of course). You can also try to use beneficial clock skew to reduce power after timing closure, or gate the hell out of all the clocks and only enable what you need (a la mobile chips). Or switch to asynchronous, or self-clocked design (every thing has it's own clock, which sends a clock to the next thing, etc. -- it's HARD to design!).
Anyway you look at it, it's a hard problem. And people who
It doesn't, of course, unless all of those spammed read (and buy) all the Dean supporters on/. who are trying to spin this into something positive for Dean. Otherwise we're all so vociferously anti-spam, it's sad to see so many pretend like this is OK. On the other hand, it's not the end of the world, and since Dean stopped it by terminating the relationship with the spamhaus, I think he's done a fine job handling it. Hopefully it will set a precedent for others to follow (well, don't spam in the first place really).
That's not really the point is it? I mean, does anyone think bigtime spammers spam for malicious reasons? Don't we all know they're in it for the gain only (money, in most cases)? And, since they will use annoying shotgun-approach methods (such as mass mailings with misleading headers and 0.001% conversion rates) to make that gain, we don't like them. Rightfully so.
Of course Dean wasn't being malicious. I have yet to read even one insinuation to that effect. He did, however, use spam tactics (either knowlingly or by negligent failure to check) to gain what he wanted -- campaign support (also money, in a sense). He's plenty rich, so it's not the cash I bet he's after, it's the exposure, but IMHO that's no better than any other spammer (x10, viagra, big penises, university diplomas, and now Dean for president.)
You're hyper-defending and trying to change the issue to something easily defendable: Dean's lack of malicious intent. Please stop, it makes you look even more desperate.
Cute, but see how the experts do it. Don't click there until you have some spare time and a wan for a riotous laugh. Father Will U. Tuchme is my favorite.
Because they didn't happen, at least not compared to wintel boxes. Proof.
It all comes back to what I call the lemming effect -- the willingness of people to follow blindly along, never questioning as they march in step with everyone else.
Riight. You don't like it, so everyone doing it is a lemming. If everyone we buying macs on the other hand, they would definitely not be lemmings.
Don't get me wrong: Conformity isn't all bad, especially when it comes to computers. A decade ago, most workplaces were a mess of different models, few of which could work together, let alone speak to one another.
Huh? Usually, when you use the construct "few could A, let alone B", B is harder than A. In this case, A = work together (doesn't that imply "speaking" to one another?), B = speak to one another. I don't get it. This reversal of sense is disturbing.
The whole article is disturbing, in fact, and I'm sorry I read it.
Anyone else tired of this oft-touted, but plainly false claim about Apple pricing?:
Today, PCs have little price advantage over Macs. Apple's eMac, designed as a low-cost model for school systems, costs $50 more, at most, than a comparable low-end PC.
The cheapest eMac I could find is $799.00. That includes: 800MHz PowerPC G4, 128MB SDRAM, 40GB Ultra ATA drive. CD-ROM drive.
I could get equivalent PC power for under $200, I am sure, but it is hard to find such a slow CPU to compare that to. Even a 2GHz AMD system, with more than the above, can be had at Best Buy for under $550 with a 2" larger monitor (under $500 with a 17" like the eMac). (Model: T2341 is $399, add a 19" monitor for $149, 17" for $100).
It's OK to love mac, and OSX makes me like them also, but don't lie and say they are $50 more. It just isn't true.
Whoa, that's very cool. I love it when I learn a new google goodie.
If you didn't try that 'gravity' link in the parent, check this out. Google calculator -- takes input in standard algebraic format, and knows some variables and units too (such as "G" being the universal gravitational constant, "mass of earth", and "radius of earth"), so you can just use the variable name and google fills in the values, converts units as needed, and gives a numeric result. Nice.
However, unless I'm doing something wrong or they're stil updating, the known variables seem rather limited. ( population of china ) / ( surface are of earth) didn't work. Neither did ( 1 barleycorn ) / (1 mm).
Well, not sure about what the OP through was funny, but I sure do think this is, from the article:
"It is feasible that an individual may include their social security number on copies of a resume sent to prospective employers, but delete it from the version put online to guard against identify theft," Byers writes.
Who in their right mind puts their SSN in any version of a resume??!
Apple said in Tuesday's court filing that it "continues to vigorously deny all of the material allegations" of the lawsuit but is willing to settle to avoid the costs of continuing to fight the legal action. An Apple representative declined to comment further.
. . . but . . .
"If you are completely dissatisfied (with Mac OS X), you can return it and get your money back," Ferlauto said. "If you want to keep OS X, but are kind of annoyed that you don't have full support, you can get (a $25) coupon."
. . . and of course, the winner is . . .
Apple has also agreed to pay up to $350,000 in legal fees to King & Ferlauto.
What gets me is the number of people who quibble over the details (the war isn't about oil, the WMD are really there, though noone can find them etc etc) and ignore the bigger picture.
We're too far OT here, but you happen to be the only one quibbling over these datails. Most of "the other side" just thinks we did the Right Thing(TM). In some sense and in some crowds it's an unpopular thing that we've done, but WMD found now or not, backing from France and Germany or not, attacks from the un-electable hardcore left notwithstanding, we did the Right Thing. It sucks that some people don't like us because we did what we think is right (as if they wouldn't do what they think is right?), and that we happen have the capacity to do what we think is right (aha -- the key!). Frankly, it seems to me that most of those down on the US as a result of this conflict are the same ones who have been pretty down on us for a while now.
Anyway, sometimes you have to do that -- the right thing, and if it's important, it will be hard and you will face adversity and detractors with arguments against. Such as yours. It's OK when you voice your real reasons against all this -- I can still respect you and at worst we can disagree. But, when you tout patently false or even dubious assertions as unquestionable fact, you lose my respect, and with it the chance to convince me that you are right.
A good argument is like a handful of sand -- the more tightly you hold on to it, the faster it slips through your fingers.
While you may be correct in reference to the PS2 (I have no idea), and you're are certainly justified in slapping the AC a bit, I must point out that "1-chip" != "1-processor".
Nowadays we often put several independent CPUs (processors) on one die. A CPU such as an ARM or NEC VR requires only 20-100k gates (depending on options used), and a whole x86 can be done easily in less than a million gates. A sound or networking core will use a custom engine rather than a general-purpose CPU, and will usually be much smaller. Each can have its own cache, memory, IO, etc. and still I can have more than one on a chip since 2-10 million-gate designs are pretty common now, even in the ASIC world. Custom designs can pack 20M, 30M or more gates onto one affordable die (read: up to about 18-19mm per side, larger than which causes incredible low yield rates and thus high prices.)
Conversely, as the old Pentum Pro showed, you can have one processor (if you include cache, which I do) on more than one chip (die). The just put the chips all on a little PCB and covered them up so as not to be too unconventional-looking. Also to deemphasize the fact that the CPU-cache bus was on slow PCB instead of fast wires inside one big chip.
Long story short, if it costs me the consumer the same amount of money, I'd prefer more chips please.
I'd consider more than just those two criteria if I were you. Such as performance and quality. And if you have to stick to those only, I'd seriously recommend the one-chipper, especially in mature technologies.
In general, if I can put things that talk to each other on one chip (assuming I can keep each thing's perfomance the same, which I can), I will get better overall performance from the integrated (one chip) system than the multi-chip because on-chip interconnect is fast and cheap (wide) compared to interconnect on PCBs or any other chip glue medium you can think of. On chip I need a small, fast buffer (10 pico-seconds, or 10^(-12) seconds, ten millionths of a millionth of a second) and a few mm of wire at most (200-500ps). OFF-chip, to another chip, I need a output driver (500-1000ps) and a really slow PCB trace (several 1000ps or more) to wherever the nearest place (often far for routability) they could find to put the other chip. On a chip 1000 wires connecting two IP-core blocks with less than 1000ps delay between them (or up to ~1GHz) is a lot faster than an off-chip bus of 32-256 wires (more is just too costly and big) with a delay of more than 5000ps (200MHz, maybe 400-500MHz effective with DDR signals).
As chips get smaller and faster for the buck (Moore's Law), PCBs stay at pretty the same density (routing pitch), speed (FR4 sux0rs), size, and cost. And, when new tech comes out people hurry to get first working silicon to market so they don't mind using a whole chip on the thing (such as with 5.1 sound logic core, PCI bus logic core, USB 2.0 logic core, IEEE1394, etc.). Then, by the time the tech really gets accepted in the mainstream, the phsyical technology (chip making process) has already advanced so much that you could have fit 2-10 copies of it in the new process technology were it available back then. So next generation you see core1 + core2 + core3 etc all in one die, along with some nice reduction in system size, power, I/O count, and cost. And usually at least a small (and often a large) performance boost from the interconnect latency reduction of integration.
A mildly-transparent troll (note my "Uh oh. IHBT?" comment. IHBT = I Have Been Trolled), but worthy of refute nonetheless. Not because of insight or accuracy, or even the rare-but-valuable troll effect of uncovering ignored relevant perspectives or information. No, only because it's so fashionable nowadays to bash the US and its motives and flippantly toss about factual-sounding allegations that, if close to true, would result in another shot heard 'round the world by now.
Remember when we were the underdogs and everyone wanted us to do well? Well, now all that has backfired I'm afraid. We did too well, and now everyone wants us to lose again. Human nature sucks sometimes.
I don't, google does, and their exact algorithm is a tightly-guarded secret. You can easily see the google pagerank of any webpage if you install the google toolbar (I use the beta 2.0, since it also has a nice popup stopper, form-filler, and other handy features), but changing a pagerank is a much harder task. In general, a higher pagerank means the page more sites (each weighted by its pagerank, of course) linking to it.
I see the potential confusion here, since it can appear to be a bit of a vicious loop (though it really isn't) -- you have to get highly-ranked sites to link to you in order to increase your pagerank. Or, less-effectively, get lots and lots of not-so-well-ranked pages to link to you. Either way, you have to get people to link to you (or cheat and make fake sites whose purpose is only to link to your other sites, but google catches on to that eventually). The way you do that is by having good information. Which is why the system works so well.
No. The US doesn't invade it's neighbours. It invades complete strangers on the other side of the planet.
Complete strangers? Hmm, I thought one of the arguments against the war in Iraq was that the US was in bed with them beforehand (similarly the Afghanis and Taliban). I think if we were really complete strangers they'd have been much safer. Don't you agree?
It's easier to get the funding through Congress that way.
Huh? Source? Basis? Clue? I think Congress usually wants a reason (and in the case of Iraq, apparently got a good enough one, since Repubnocrats and Demolicans both approved the action.) It's hard to give a reason to invade a complete stranger. Uh oh. IHBT?
Of course, if South America were overflowing with oil, well I'm sure someone would come up with a nice excuse. It's not like South America doesn't have its share of dictators and civil rights violations we couldn't use as justifcation.
Here's where you kill your own argument in-utero. The US gets more of its oil from Venezuela (13%) than it does from Iraq (8%). This is true now, and was true before the war. Check out that site for more surprising facts about the US and oil (some of which provide much better ammo for assaults like the one you attempted and failed at so miserably.)
My favorite (doesn't help you at all, and is quite OT): In 1996, the USA was the world's leading [oil] producer, with about 7.5 million barrels per day
Whether or not your intentions are good (I think so) doesn't aid your incredibly weak attempt at an argument.
Slashdot Mirror
OK, I've waited in vain for clarifcation, so I have to ask. Why, again, exactly, is he a moron for having the beliefs that he does? Or, is it, perhaps, that you are the moron for failing to back up (or rescind) your attack?
Leakage current is a function of the transistor design only (mostly, the resistance of the gate oxide). Frequency of operation has no impact, since this current is cause by static (non-switching) current from voltage at the gate of the transistor. Dynamic (switching) power increases linearly with frequency, because every clock edge charges (or discharges) the capacitor that is the formed by the transistor gate.
Think of leakage (always-on) current as a leaky faucet. In this analogy, the switching current would be the force, or energy required to fill up (or empty) a pool. The whole while the pool fills or empties, leakage current drips precious water on the ground -- this wasted leakage does no work for your pool, but leaves water (heat) to be dealt with.
That said, the leaks in the faucet, so to speak, in any processor you have in your box now (0.13um and bigger) are still pretty negligible, and there's not much you can do about it anyway (except lower voltage, which makes over-clocking harder, but fits well with under-clocking). Most of your power (and heat) comes from switching current, so yes, lowering your clock speed will lower the heat dissipation requirements of your box. The reduction will be linear with clock rate, so 1/2 GHz = 1/2 power.
So, there's still some hope.
;)
Oh, you're going to cover the licensing fees for GAIM? Cool! Thanks! Hey guys, didja hear that? Gherald is covering the MS license fees for GAIM to keep working with Messenger . . .
Uh, patents (the good kind)? Licensing fees? Higher profit margins from early-adopter sales? A head start on everyone? They know what they're doing (business wise) and are the best-prepared to do it technology-wise (though I have my doubts here), but they still probably won't get to try, not with the cash their asking from a tired economy that's been in recession 10x as long as the US has.
I don't believe strongly in anything except that in the freedom of being what you feel you are and expressing yourself in a way that does not physically hurt anyone.
You're preaching to the choir here; I heavily depend on the "do as thy will while harming none" way of thinking, but let me ask you this: did you ever consider that posibly, just maybe, there are some forms of harm you cannot anticipate accurately? Given that humans are complex (beyond our understanding, mostly) and that societies of humans are even more complex, how can you ever be sure that your (in-)action causes no harm?
Ar-e? Nihongo wa hanasemasu ka? Nihonjin desu ka?
:)
If so, you'd know that both "hara-kiri" and "seppuku" are the same, with no distinction whatsoever in Japanese, not even in politeness levels, as is oft misreported.
In fact, they are Japanese palindromes. I'll spare you the mess of ASCII-art Kanji, but there is one character that can be read as "hara" or "se" depending on context, and another that is read as both "kiri" and "puku". Put them together one way, they are pronounced "hara kiri"; swap them and they are pronounced "seppuku", but the mean exactly the same in both cases (at least in Tokyo and Awaji, maybe in other areas there is a difference? But I don't think so.)
Besides, the original poster said "hari kari", and that's hard to confuse for "seppuku". I think he was going for "hara kiri"
We're trying.
:)
But how do you get "micron high" little gold studs to stick to the die in exactly the right places? How do you make sure each gold stud is exactly the same height (can't have a short one anywhere, even by a femto-meter)? Then, how do you physically/mechanically line them up exactly and keep them together perfectly for long priods of time under fairly wide ranges in vibration and temperature ranges? How do you prevent the dice from warping if each stud isn't 100% identical (such as if you try to tolerate some height variation by making the studs slightly compressible)?
Since you're using area-IO to connect the dice, how do you power them? Usually, in area-IO die the power comes from the top (like an IO buffer), but in a stacked die this would just lead to another die. You can't power from the "bottom", since that's not metal (it's Si substrate), and you need really big power wires to get all over the dice from somewhere. If from the top, do the tiny wires shoot out the sides between the two dice and then go to a power, uh, plug? Connector with tiny wires on one side, or what? And I do mean tiny -- the little metal studs would need to be placed every 200-300um apart, in a 2-D array, and some would have to connect to a power source, somehow.
If you got this far: how do you design a chip with identical, but mirror-image IO locations of another chip, which presumably does a different thing? It's a huge battle in system design these days to get a chip package pinout that makes both the PCB designers and the die designers happy. Making 2 die designers happy with one chipl-level pinout would be impossible.
Finally, given that gross simplifications and assumptios about near-perfect isolation used in modern chip design, how long until you can have the hardware and software ready for me to be able to calculate iterative solutions to 500 million simultaneous, co-dependent variations on Schroedinger's Wave Equation? (of course, it's intractable, so iteration is your only hope -- better pray for convergence too!) Oh, and I need that to take less than 6-12 hours each run to make a reasonable design schedule.
Right on a). well, mostly -- IBM has a new process that does allow transistors in some area-IO to be placed over logic gate transistors. It's more trouble than it's worth, though (unavoidable interactions are hard to calculate accurately).
:). See, it's not that it's further to go vertical from one die to the next, rather than packaging each individually and connecting them horizontally. The problem is it's hard to go vertical. This is true from design, manufacturability, and reliability points-of-view.
And right on b) -- the distance between 2 dice stacked is much shorter than 2 side-by-side. But this is totally irrelevant, mostly due to previous posters
First, by area-IO I meed input/output (IO) drivers or receivers that can be placed anywhere in an area, rather than only around the circumference (preipheral IO). We have area-IO at the package level (such as BGA, or Ball-Grid-Array and FCBGA, or Flip-Chip Ball Grid Array [best for area-IO, and expensive]) and area-IO at the die level. Do we connect the dice before or after packaging?
Either way presents problems. Such as (for pre-packaging connections):
How do you electrically connect 2 area-IO dice? Usually, a die has little square landing pads, and these are only about 50um square, spaced every 200-250um or so on center in 2-D arrays of up to 70x70 and more. To be able to do anything with these tightly packed little signals, we drop special tiny drops of metal that stick to the pads, and press this up against a package substrate (ceramic), which includes routes to space those signals out more, like every 1.0mm or so. Even this is expensive and hard to mount to PCB, since it's hard to ensure both things are perfectly flat (package and PCB) so that all balls connect.
In fact, we rely on the package (often including an internal metal "stiffener") to keep the die nice and flat, which helps avoid de-lamination (layers peeling apart). Two dice pressed next to each other would require some space between them to make the connection (i.e., some bumps for the connection, and valleys for no connect areas), and this and the elasticity of the electrical connection medium would leave enough play to let the dice warp all over the place.
It'd be even harder to tell which ball(s) aren't connected. We do this now by confirming that the PCB is OK (usually pretty easy, so it makes a good reference), make the chip send specially-controllable data out (and take data in on inputs), then check to see what's right and wrong by measuring at the board level. If my board is another chip, how do I know which one I am debugging? This debugging (we call mfg testing) happens to all chips, not just some samples. If it isn't, failure rates will go up to unacceptable levels (like 20-50% or more).
Testability is hard if you stack dice before or after packaging. Design is a bizzotch too, since you can't very well even model one whole chip at a time (and how the circuit performs depends on process, voltage, and temperature), much less two chips stacked with an insulator and some kind of very short, very small, very fragile, very susceptible to noise and crosstalk hunk of 1000+ wires between them. One local hot spot at X,Y on die A can mess up operation at x,y on die B, and we'd never be able to practically predict that.
Most importantly of all, part of the reason chip design even works at all, and that we can churn them out for pennies each (after massive design and capital outlay for a fab), is that we can simplify the design dramatically by making assumptions, modelling the target device in isolation, verifying it in isolation, and then being able to safely assume this (truly wrong) assumption of isolation is close enough to true that the part will work in the system. Single packaged die are relatively infinitely insulated from everything except the I/O we carefully design. Stacked dice would not be -- they would interact strongly with each other in unpredicatab
I don't necessarily agree or disagree with any of you, but I do have to take issue with the ever-increasing tendency of (all kinds of) people to disparage anyone with firm personal beliefs. It seems like the only way to avoid this is to have completely relativity-based ethics and morals (no sure right or wrong -- "everything is relative"), and to adamantly avoid stating something one believes in firmly. I think this is a Bad Thing.
See, vacillating and changing your "beliefs" in order to appease or not offend others, IMHO, is disingenuous and dangerous. It's also subversive, and really no different than lying. Even failing to form an opinion and expressing it in the proper forum is somewhat irresponsible and lazy, IMHO. Such chameleons/chickens change their views (or at least downplay them) as is convenient, often to win support, avoid dissent, or just to be agreeable. Worse, it really seems to upset some of them when someone else expresses his or her own firm beliefs.
Frankly, I think the GP post was a non-flamebait, non-troll expression of the poster's beliefs. Just because you disagree doesn't make him a moron.
FYI, it's "hara kiri". Harry Carey (which you oddly spelled "hari kari") was a baseball announcer.
And about the only way to do this without sacrificing clockrate is by going to a smaller fabrication process.
Sorry, that's commonly believed, but wrong. There are lots of ways to reduce power consumption. Reducing gate widths (0.25um -> 0.13um -> 90nm) is commonly touted as a good way to reduce power, but in most cases that's more marketing pitch than reality.
First, there are two types of chip power to worry about (1) leakage, which happens all the time, just by being on, and which used to be always much much lower than (2) the switching power, or maximum dissipation when as many transistors as possible can switch at once (which, BTW, can never be all of them, and it's really, really hard to find the stimulus that makes maximum power happen. So, esitmates like the ones in the article for peak power are often made assuming a somewhat-arbitrary switching factor that may be low or high).
As gate sizes shrink, the effective capacitance of the gate shrinks, and voltage can be lowered (to a point). Capacitance varies with gate area and inversely with distance between "plates" of the gate (C = k*A / d). Reducing the gate width (space between the plates) actually increases capacitance, and this itself would increase power. But, you're also able to reduce the gate area (though not as much, but in 2-dimensions, so shrinking gates is usually a reduction in C). Most importantly, you can decrease voltage, since power varies with the square of voltage, this has much more impact on power than reducing gate capacitance (size). When we went from 0.25um (3.3V)to 0.13um (1.5V), we got a nice fat 1.8V drop in voltage. But 0.13um is 1.5V too, or 1.3V at best, and I've never heard of a 90nm (0.09um) process under 1.1V. The V isn't dropping as fast any more because the noise margins are getting too small.
Since p(switching) = 1/2*F*C*V^2 (F = clock freqyency, C = capacitance, and V = max voltage, lowering C (and moreso V which we can reduce some, but not much below 1.0V so far) will lower power a bit. Linearly with C. But unless we can reduce V, reducing C much more won't help a lot because we have more total C's (transistor gates) on the die, because they are smaller we can fit more.
But now, at 0.13um, and more at 90nm, it's not the switching power, but the leakage (always there) power that's getting worrisome. It used to be 1/20th of switching power or less, but now the gates are so small current of the same order of magnitude (almost) of switching leaks all the time.
So, the more you shrink, the more you have constant power, which is harder to deal with since you can't throttle it, and it's always cranking out. Worse yet, the more you shrink, the more gates you can fit on one tiny little die (the feasible mfg'able die size stays around 17-18mm max regardless of gate size once the process matures a bit, but bigger dice have ridiculous failure rates and thus silly high prices). And the gates shrink in 2 dimensions (L and W), so you get a squaring increase of the toal gate count, and only a linear decrease with C. Shrinking gates to save power doesn't work.
So, if we can't keep shrinking to save power, how can we? Lot's of ways. There are dozens of EDA companies with power-minded RTL coding, synthesis, and even place and route tools ready to help you reduce your power if you have a few $100k/seat/year. Or, you could use a SSC (Spread-spectrum Clock, where each clock edge is off by a bit to reduce power, but it slows down the max clock rate a bit too, of course). You can also try to use beneficial clock skew to reduce power after timing closure, or gate the hell out of all the clocks and only enable what you need (a la mobile chips). Or switch to asynchronous, or self-clocked design (every thing has it's own clock, which sends a clock to the next thing, etc. -- it's HARD to design!). Anyway you look at it, it's a hard problem. And people who
how does this help your compaign?
/. who are trying to spin this into something positive for Dean. Otherwise we're all so vociferously anti-spam, it's sad to see so many pretend like this is OK. On the other hand, it's not the end of the world, and since Dean stopped it by terminating the relationship with the spamhaus, I think he's done a fine job handling it. Hopefully it will set a precedent for others to follow (well, don't spam in the first place really).
It doesn't, of course, unless all of those spammed read (and buy) all the Dean supporters on
That's not really the point is it? I mean, does anyone think bigtime spammers spam for malicious reasons? Don't we all know they're in it for the gain only (money, in most cases)? And, since they will use annoying shotgun-approach methods (such as mass mailings with misleading headers and 0.001% conversion rates) to make that gain, we don't like them. Rightfully so.
Of course Dean wasn't being malicious. I have yet to read even one insinuation to that effect. He did, however, use spam tactics (either knowlingly or by negligent failure to check) to gain what he wanted -- campaign support (also money, in a sense). He's plenty rich, so it's not the cash I bet he's after, it's the exposure, but IMHO that's no better than any other spammer (x10, viagra, big penises, university diplomas, and now Dean for president.)
You're hyper-defending and trying to change the issue to something easily defendable: Dean's lack of malicious intent. Please stop, it makes you look even more desperate.
Cute, but see how the experts do it. Don't click there until you have some spare time and a wan for a riotous laugh. Father Will U. Tuchme is my favorite.
Anxious is a polite way to put it. He's not only vociferously wrong, his zealotry is annoying.
Why haven't steep price cuts stemmed Apple's market fall?
Because they didn't happen, at least not compared to wintel boxes. Proof.
It all comes back to what I call the lemming effect -- the willingness of people to follow blindly along, never questioning as they march in step with everyone else.
Riight. You don't like it, so everyone doing it is a lemming. If everyone we buying macs on the other hand, they would definitely not be lemmings.
Don't get me wrong: Conformity isn't all bad, especially when it comes to computers. A decade ago, most workplaces were a mess of different models, few of which could work together, let alone speak to one another.
Huh? Usually, when you use the construct "few could A, let alone B", B is harder than A. In this case, A = work together (doesn't that imply "speaking" to one another?), B = speak to one another. I don't get it. This reversal of sense is disturbing.
The whole article is disturbing, in fact, and I'm sorry I read it.
Anyone else tired of this oft-touted, but plainly false claim about Apple pricing?:
Today, PCs have little price advantage over Macs. Apple's eMac, designed as a low-cost model for school systems, costs $50 more, at most, than a comparable low-end PC.
The cheapest eMac I could find is $799.00. That includes: 800MHz PowerPC G4, 128MB SDRAM, 40GB Ultra ATA drive. CD-ROM drive.
I could get equivalent PC power for under $200, I am sure, but it is hard to find such a slow CPU to compare that to. Even a 2GHz AMD system, with more than the above, can be had at Best Buy for under $550 with a 2" larger monitor (under $500 with a 17" like the eMac). (Model: T2341 is $399, add a 19" monitor for $149, 17" for $100).
It's OK to love mac, and OSX makes me like them also, but don't lie and say they are $50 more. It just isn't true.
Whoa, that's very cool. I love it when I learn a new google goodie.
If you didn't try that 'gravity' link in the parent, check this out. Google calculator -- takes input in standard algebraic format, and knows some variables and units too (such as "G" being the universal gravitational constant, "mass of earth", and "radius of earth"), so you can just use the variable name and google fills in the values, converts units as needed, and gives a numeric result. Nice.
However, unless I'm doing something wrong or they're stil updating, the known variables seem rather limited. ( population of china ) / ( surface are of earth) didn't work. Neither did ( 1 barleycorn ) / (1 mm).
Anyone have tips on this new google gem?
Well, not sure about what the OP through was funny, but I sure do think this is, from the article:
"It is feasible that an individual may include their social security number on copies of a resume sent to prospective employers, but delete it from the version put online to guard against identify theft," Byers writes.
Who in their right mind puts their SSN in any version of a resume??!
so . . .
Apple said in Tuesday's court filing that it "continues to vigorously deny all of the material allegations" of the lawsuit but is willing to settle to avoid the costs of continuing to fight the legal action. An Apple representative declined to comment further.
. . . but . . .
"If you are completely dissatisfied (with Mac OS X), you can return it and get your money back," Ferlauto said. "If you want to keep OS X, but are kind of annoyed that you don't have full support, you can get (a $25) coupon."
. . . and of course, the winner is . . .
Apple has also agreed to pay up to $350,000 in legal fees to King & Ferlauto.
What gets me is the number of people who quibble over the details (the war isn't about oil, the WMD are really there, though noone can find them etc etc) and ignore the bigger picture.
We're too far OT here, but you happen to be the only one quibbling over these datails. Most of "the other side" just thinks we did the Right Thing(TM). In some sense and in some crowds it's an unpopular thing that we've done, but WMD found now or not, backing from France and Germany or not, attacks from the un-electable hardcore left notwithstanding, we did the Right Thing. It sucks that some people don't like us because we did what we think is right (as if they wouldn't do what they think is right?), and that we happen have the capacity to do what we think is right (aha -- the key!). Frankly, it seems to me that most of those down on the US as a result of this conflict are the same ones who have been pretty down on us for a while now.
Anyway, sometimes you have to do that -- the right thing, and if it's important, it will be hard and you will face adversity and detractors with arguments against. Such as yours. It's OK when you voice your real reasons against all this -- I can still respect you and at worst we can disagree. But, when you tout patently false or even dubious assertions as unquestionable fact, you lose my respect, and with it the chance to convince me that you are right.
A good argument is like a handful of sand -- the more tightly you hold on to it, the faster it slips through your fingers.
While you may be correct in reference to the PS2 (I have no idea), and you're are certainly justified in slapping the AC a bit, I must point out that "1-chip" != "1-processor".
Nowadays we often put several independent CPUs (processors) on one die. A CPU such as an ARM or NEC VR requires only 20-100k gates (depending on options used), and a whole x86 can be done easily in less than a million gates. A sound or networking core will use a custom engine rather than a general-purpose CPU, and will usually be much smaller. Each can have its own cache, memory, IO, etc. and still I can have more than one on a chip since 2-10 million-gate designs are pretty common now, even in the ASIC world. Custom designs can pack 20M, 30M or more gates onto one affordable die (read: up to about 18-19mm per side, larger than which causes incredible low yield rates and thus high prices.)
Conversely, as the old Pentum Pro showed, you can have one processor (if you include cache, which I do) on more than one chip (die). The just put the chips all on a little PCB and covered them up so as not to be too unconventional-looking. Also to deemphasize the fact that the CPU-cache bus was on slow PCB instead of fast wires inside one big chip.
Long story short, if it costs me the consumer the same amount of money, I'd prefer more chips please.
I'd consider more than just those two criteria if I were you. Such as performance and quality. And if you have to stick to those only, I'd seriously recommend the one-chipper, especially in mature technologies.
In general, if I can put things that talk to each other on one chip (assuming I can keep each thing's perfomance the same, which I can), I will get better overall performance from the integrated (one chip) system than the multi-chip because on-chip interconnect is fast and cheap (wide) compared to interconnect on PCBs or any other chip glue medium you can think of. On chip I need a small, fast buffer (10 pico-seconds, or 10^(-12) seconds, ten millionths of a millionth of a second) and a few mm of wire at most (200-500ps). OFF-chip, to another chip, I need a output driver (500-1000ps) and a really slow PCB trace (several 1000ps or more) to wherever the nearest place (often far for routability) they could find to put the other chip. On a chip 1000 wires connecting two IP-core blocks with less than 1000ps delay between them (or up to ~1GHz) is a lot faster than an off-chip bus of 32-256 wires (more is just too costly and big) with a delay of more than 5000ps (200MHz, maybe 400-500MHz effective with DDR signals).
As chips get smaller and faster for the buck (Moore's Law), PCBs stay at pretty the same density (routing pitch), speed (FR4 sux0rs), size, and cost. And, when new tech comes out people hurry to get first working silicon to market so they don't mind using a whole chip on the thing (such as with 5.1 sound logic core, PCI bus logic core, USB 2.0 logic core, IEEE1394, etc.). Then, by the time the tech really gets accepted in the mainstream, the phsyical technology (chip making process) has already advanced so much that you could have fit 2-10 copies of it in the new process technology were it available back then. So next generation you see core1 + core2 + core3 etc all in one die, along with some nice reduction in system size, power, I/O count, and cost. And usually at least a small (and often a large) performance boost from the interconnect latency reduction of integration.
A mildly-transparent troll (note my "Uh oh. IHBT?" comment. IHBT = I Have Been Trolled), but worthy of refute nonetheless. Not because of insight or accuracy, or even the rare-but-valuable troll effect of uncovering ignored relevant perspectives or information. No, only because it's so fashionable nowadays to bash the US and its motives and flippantly toss about factual-sounding allegations that, if close to true, would result in another shot heard 'round the world by now.
Remember when we were the underdogs and everyone wanted us to do well? Well, now all that has backfired I'm afraid. We did too well, and now everyone wants us to lose again. Human nature sucks sometimes.
I don't, google does, and their exact algorithm is a tightly-guarded secret. You can easily see the google pagerank of any webpage if you install the google toolbar (I use the beta 2.0, since it also has a nice popup stopper, form-filler, and other handy features), but changing a pagerank is a much harder task. In general, a higher pagerank means the page more sites (each weighted by its pagerank, of course) linking to it.
I see the potential confusion here, since it can appear to be a bit of a vicious loop (though it really isn't) -- you have to get highly-ranked sites to link to you in order to increase your pagerank. Or, less-effectively, get lots and lots of not-so-well-ranked pages to link to you. Either way, you have to get people to link to you (or cheat and make fake sites whose purpose is only to link to your other sites, but google catches on to that eventually). The way you do that is by having good information. Which is why the system works so well.
No. The US doesn't invade it's neighbours. It invades complete strangers on the other side of the planet.
Complete strangers? Hmm, I thought one of the arguments against the war in Iraq was that the US was in bed with them beforehand (similarly the Afghanis and Taliban). I think if we were really complete strangers they'd have been much safer. Don't you agree?
It's easier to get the funding through Congress that way.
Huh? Source? Basis? Clue? I think Congress usually wants a reason (and in the case of Iraq, apparently got a good enough one, since Repubnocrats and Demolicans both approved the action.) It's hard to give a reason to invade a complete stranger. Uh oh. IHBT?
Of course, if South America were overflowing with oil, well I'm sure someone would come up with a nice excuse. It's not like South America doesn't have its share of dictators and civil rights violations we couldn't use as justifcation.
Here's where you kill your own argument in-utero. The US gets more of its oil from Venezuela (13%) than it does from Iraq (8%). This is true now, and was true before the war. Check out that site for more surprising facts about the US and oil (some of which provide much better ammo for assaults like the one you attempted and failed at so miserably.)
My favorite (doesn't help you at all, and is quite OT): In 1996, the USA was the world's leading [oil] producer, with about 7.5 million barrels per day
Whether or not your intentions are good (I think so) doesn't aid your incredibly weak attempt at an argument.
[the US has] never invaded a neighbor for oil
*cough*
Watch any news lately?
*cough* USA invades Iraq TWICE in 12 years *cough*
No, we'd never do so for oil. No, never!
[/sarcasm]
Iraq? Neighbor?
Seen any maps lately?
And you really should see someoune about that cough^H^H^H^H^H liberal bias.