Being at 45-nm while AMD is at 65-nm means Intel's costs per die are lower. Even if the high-K step is more expensive than the corresponding gate-building step at AMD, the ability to get more transistors (and chips) from a single wafer spreads that cost over a much larger number of marketable units.
And you're right about AMD being way behind getting to the 45-nm process; but you're missing that it costs them even more, because they have to buy it from IBM, or rent space in TSMC or Chartered's contract fabs. Intel has no middle-man to pay.
AMD's original Fusion concept was much broader than what they've settled on, which is a low-end CPU and a midrange GPU on one chip.
The Nehalem solution from Intel will be a high-end CPU and a GPU of uncertain quality on one chip, maybe a range of GPUs will be available.
AMD won't be the superior product in any line against Intel in any foreseeable future.
Re:Intel mistakes: CPU development is VERY difficu
on
Is AMD Dead Yet?
·
· Score: 1
The past is the past and the future is now.
AMD's Barcelona design for the Opteron and Phenom has the thermal problems and can't be sold above 2.6 GHz and is very hard to overclock successfully.
Meanwhile, Intel's 45-nm quad-core designs start at 3 GHz and overclock to 4 GHz easily with only fan/heatsink cooling.
Intel is ahead of AMD, by a ton, and AMD is rapidly going broke (lost $2 billion last year on operations and another $2 billion due to writeoffs; has less than #2 billion in the bank and can get more only through junk bonds and massive stock dilution). Intel can always be ahead of AMD when AMD goes out of business.
They sell something worth $11.38 billion for 10.75 billion, but never booked it at 11.38 billion in the first place.
Sony shows $10.75 billion in cashflow, no appreciable decrease in assets, and covers it with profits from its new hi-def disc monopoly.
90% of its shareholders are fund, anyway, whose managers won't care as long as their funds still sell, and since SNE is only going to be 0.8% of any one fund, the effect of the graft is a tiny splash buried in the roaring surf of the market.
In general, ionic bonds (i.e., bonds with high ionic character) are stronger than covalent bonds (i.e., bonds with low or no ionic character), simply because they are in fact the same thing, except that a difference in electron affinity causes a dipole moment to be generated, which adds electrical potential to quantum potential of the bond.
The larger the difference in electronegativity (or electron affinity, however you want to measure it), and the shorter the internuclear distance, the stronger the ionic force of the bond.
It just so happens that the quantum effect in a carbon-carbon bond is pretty strong in the first place, so there aren't many bonds, even those high in ionic character, that are stronger (although N-N is nearly twice as strong, iirc)
Putting ions in a position to have to share their ionic attraction among more atoms than they have valence electrons weakens their bonds, so there aren't many ionic substances that have nearly as strong a structure in crystal lattices.
The strength of diamong is due to the fact that (1) C-C is a fairly strong bond and (2) of the valence-4 atoms, which allow for the least disruptive crystal structure, C-C has the strongest bonds. It's that combination of no bending and strong bonds that makes diamond hard. Though there are far harder substances.
This being chemistry, someone will of course find counterexamples. It's pretty amazing how so few rules for atoms can produce so many intricate variations in behavior once you get atoms close together.
Did Fosset have more money than Gray? Does Gray's wife have full control of Gray's assets if he's not dead? Fosset had a complicated set of assets, and probably didn't have all the backups in place for managing them. Fosset's wife was evidently not able to take control of certain things until Fosset was declared dead. No need to look for ulterior motives here. It's as simple as not wanting those things to decay to nothing due to neglect while you watch, helpless.
30-50 years, by my calculation; backed up by others I've seen since.
We know how much oil there is, and we have statistical models of what's left to be discovered, and we know how much we use, and how fast that grows. Simple extrapolation.
And when it runs out it's not gradual. One second you're sucking on the straw, the next second it's gurgling in the bottom of the cavern. Since we're accelerating our usage all the time, we're going to hit that wall fast and sudden.
The alternative is to start replacing our oil usage with electrical usage. There's no other portable solution that's nearly as inexpensive or manageable.
Your apartment building can install outlets in the parking lot for a few hundred dollars per unit. He's got a decade or so to get on it.
Firstly, this maximizes the relative momentum, and the force of the impact. More bang for the buck spent on the launch.
Secondly, it produces the least mean resultant momentum from both objects in the collision.
Third, anything thrown up will come down at a steeper angle.
It is possible for portions of the debris to end up with greater orbital momentum than they had before the collision, but it would be a smaller portion. And it would be in a more elliptical orbit, which would increase its incursion into the atmosphere.
Basically, the chances any of this stays up even as long as the whole satellite was supposed to stay up are very low.
But, we're decreasing the momentum of many of the pieces to the point they will probably not burn up on reentry. We may be reducing the footprint for the bulk of the debris, but there will be an expanding ball of debris that will have a larger footprint. I hope someone's figured that out.
Oh, and because there's no air up there, all the shock has to come from the collision or explosion. That changes what "explosion" means in a big way. Don't expect this thing to be blown to tiny bits. Broken into small chunks is more like it.
Small chunks that will be more likely to survive re-entry, and may have a wider and longer debris field.
Batteries are cheap and less inefficient than the Otto cycle engine in your car.
As for charging times, you can charge it when you're sleeping.
Long-distance travel will take a major hit when the oil runs out. There's nothing to use as jet fuel that's as good as jet fuel. That's why it's jet fuel.
And where do they get the electricity to reprocess the Mg02?
From an oil- or coal-burning power plant, of course.
Or a nuke plant.
These ideas of using renewable chemical fuels is all pretty silly, because they all use electricity to renew the fuel. But electric vehicles are efficient, viable, can be made attractive and fast, and they cut out the middle-man by allowing you to plug into a supply of electricity you already access. No infrastructure cost = lowest economic barrier to entry. And it's infrastructure that we have 150+ years of experience maintaining and improving.
Eventually all of our energy will be delivered from electrical utilities, generated from coal (the oil will run out soon but we have several hundred years' worth of coal left), nuclear processes (about a thousand years' worth), and the sun (several billion years, but it's terribly inefficient so far).
my upload bandwidth is a small fraction of my download bandwidth
i will not accept becoming a server for Microsoft's customers so that Microsoft can save a dollar a month on its IT bills
there are far better and smarter ways to spread out the update downloads, like actually using the scheduled-execution capabilities built into the operating system
No, it's evidence that people pay good money for Vista. The reason there are no "Suse Linux Annoyances" isn't that Linux isn't annoying, it's that there's no money in evaluating and documenting its annoyances for a user base that's 1/1000th that of Windows and who got their OS for free anyway.
A better way is to get a new computer. People are clearly installing this operating system that is sized for the next 4-5 years of computing on machines they've had for the past 4-5 years, or bought cheap recently (because they contain subsystems that go back several years).
This is a fairly expensive rig now (cost me $2400 or so in parts, a day in labor, over the past month), but it runs Vista clean and fast, and within the year will probably be in the $1000-1500 retail segment:
Intel QX9650 3 GHz quad-core CPU x38-chipset motherboard with 1333-MHz FSB 2 GB of 1333-MHz DDR-3 DRAM 1 TB of HD in 2 500-GB, RAID-0 paired, SATA-300 (i.e., SATA-II) drives.
To get the cost down, don't use a tweaker's motherboard like the ASUS Maximus Extreme; get something that still supports 1333-MHz DDR-3 DRAM and has a RAID controller for its SATA ports.
A Q6600 at 2.4 GHz ought to nearly as the QX9650 at a quarter the price. You're still blowing 9.6 billion cycles per second counting all four cores, and yes, they all get used when there's stuff to be done. A dual-core at 2.4 or even 3 GHz couldn't keep up. That's $700 off right there, and there are a couple of choices between here and the top model.
The disk system is very key. This is still the bottleneck for almost anything in the box. You rarely interact with the machine without it realizing it needs to read something new or write something important. One disk with a 4-ms nominal seek time, or a pair of disks with 10-ms seek times set up in RAID-0, pretty much appear the same.
With half-decent video gear (mine is a $60 nVidia 8400-based card) you can get Aero running without a hitch, too.
So there's the thing. People have gotten spoiled being able to buy and run $300 computers with operating systems that come stapled into magazines. If you want something new and cool, it will cost you. But you want it. You know you do. And it's not that much, really, for something that will do you a good service for several years.
Because software can pop up a box on your screen saying "go look for the sticker on the box and type the letters and numbers (and maybe the dashes or maybe not, your guess is as good as ours) you see there into this box here then click the button that says 'OK'".
Hardware says "blink"..."blink"..."blink"... and user calls customer support, adding $10 to the cost of every sale.
The ACPI settings in new installs everywhere I go generally go into low-power modes by default. I have to turn them to full-on to keep my grid computing apps up 24/7.
Because I think curing cancer and finding quasars is a bit more important than saving 20 cents' worth of electricity every day...
Which, by the way, is a waste of the amortized cost and energy used in installing the wiring, validating it's to code, developing the code, building the power plant, running the wires from that, etc., etc.
I mean, even if all my electricity were solar, maximum efficiency would require draining the batteries every night and filling them full every day.
Think holistic. Don't cut off your screensaver to spite your infrastructure.
Slashdot Mirror
Being at 45-nm while AMD is at 65-nm means Intel's costs per die are lower. Even if the high-K step is more expensive than the corresponding gate-building step at AMD, the ability to get more transistors (and chips) from a single wafer spreads that cost over a much larger number of marketable units.
And you're right about AMD being way behind getting to the 45-nm process; but you're missing that it costs them even more, because they have to buy it from IBM, or rent space in TSMC or Chartered's contract fabs. Intel has no middle-man to pay.
AMD's original Fusion concept was much broader than what they've settled on, which is a low-end CPU and a midrange GPU on one chip.
The Nehalem solution from Intel will be a high-end CPU and a GPU of uncertain quality on one chip, maybe a range of GPUs will be available.
AMD won't be the superior product in any line against Intel in any foreseeable future.
The past is the past and the future is now.
AMD's Barcelona design for the Opteron and Phenom has the thermal problems and can't be sold above 2.6 GHz and is very hard to overclock successfully.
Meanwhile, Intel's 45-nm quad-core designs start at 3 GHz and overclock to 4 GHz easily with only fan/heatsink cooling.
Intel is ahead of AMD, by a ton, and AMD is rapidly going broke (lost $2 billion last year on operations and another $2 billion due to writeoffs; has less than #2 billion in the bank and can get more only through junk bonds and massive stock dilution). Intel can always be ahead of AMD when AMD goes out of business.
IBM was right, but a year or two overzealous.
They sell something worth $11.38 billion for 10.75 billion, but never booked it at 11.38 billion in the first place.
Sony shows $10.75 billion in cashflow, no appreciable decrease in assets, and covers it with profits from its new hi-def disc monopoly.
90% of its shareholders are fund, anyway, whose managers won't care as long as their funds still sell, and since SNE is only going to be 0.8% of any one fund, the effect of the graft is a tiny splash buried in the roaring surf of the market.
Sony bought your future. Get the fuck over it.
No primordial soup for you.
In general, ionic bonds (i.e., bonds with high ionic character) are stronger than covalent bonds (i.e., bonds with low or no ionic character), simply because they are in fact the same thing, except that a difference in electron affinity causes a dipole moment to be generated, which adds electrical potential to quantum potential of the bond.
The larger the difference in electronegativity (or electron affinity, however you want to measure it), and the shorter the internuclear distance, the stronger the ionic force of the bond.
It just so happens that the quantum effect in a carbon-carbon bond is pretty strong in the first place, so there aren't many bonds, even those high in ionic character, that are stronger (although N-N is nearly twice as strong, iirc)
Putting ions in a position to have to share their ionic attraction among more atoms than they have valence electrons weakens their bonds, so there aren't many ionic substances that have nearly as strong a structure in crystal lattices.
The strength of diamong is due to the fact that (1) C-C is a fairly strong bond and (2) of the valence-4 atoms, which allow for the least disruptive crystal structure, C-C has the strongest bonds. It's that combination of no bending and strong bonds that makes diamond hard. Though there are far harder substances.
This being chemistry, someone will of course find counterexamples. It's pretty amazing how so few rules for atoms can produce so many intricate variations in behavior once you get atoms close together.
Carl Malamud is the best thing since scroll wheels.
3 down, 1497 to go.
set 'em up again!
(oh, and, damn is that a nice house; i wish i'd been modding xboxes the past few years)
Hotblack Desiato didn't fake it. "He's spending a year dead, for tax purposes," is how it was described.
Did Fosset have more money than Gray? Does Gray's wife have full control of Gray's assets if he's not dead? Fosset had a complicated set of assets, and probably didn't have all the backups in place for managing them. Fosset's wife was evidently not able to take control of certain things until Fosset was declared dead. No need to look for ulterior motives here. It's as simple as not wanting those things to decay to nothing due to neglect while you watch, helpless.
> certainly not in the next few decades
30-50 years, by my calculation; backed up by others I've seen since.
We know how much oil there is, and we have statistical models of what's left to be discovered, and we know how much we use, and how fast that grows. Simple extrapolation.
And when it runs out it's not gradual. One second you're sucking on the straw, the next second it's gurgling in the bottom of the cavern. Since we're accelerating our usage all the time, we're going to hit that wall fast and sudden.
The alternative is to start replacing our oil usage with electrical usage. There's no other portable solution that's nearly as inexpensive or manageable.
Your apartment building can install outlets in the parking lot for a few hundred dollars per unit. He's got a decade or so to get on it.
Hit it from the front.
Firstly, this maximizes the relative momentum, and the force of the impact. More bang for the buck spent on the launch.
Secondly, it produces the least mean resultant momentum from both objects in the collision.
Third, anything thrown up will come down at a steeper angle.
It is possible for portions of the debris to end up with greater orbital momentum than they had before the collision, but it would be a smaller portion. And it would be in a more elliptical orbit, which would increase its incursion into the atmosphere.
Basically, the chances any of this stays up even as long as the whole satellite was supposed to stay up are very low.
But, we're decreasing the momentum of many of the pieces to the point they will probably not burn up on reentry. We may be reducing the footprint for the bulk of the debris, but there will be an expanding ball of debris that will have a larger footprint. I hope someone's figured that out.
Oh, and because there's no air up there, all the shock has to come from the collision or explosion. That changes what "explosion" means in a big way. Don't expect this thing to be blown to tiny bits. Broken into small chunks is more like it.
Small chunks that will be more likely to survive re-entry, and may have a wider and longer debris field.
What the fuck are they thinking?
Batteries are cheap and less inefficient than the Otto cycle engine in your car.
As for charging times, you can charge it when you're sleeping.
Long-distance travel will take a major hit when the oil runs out. There's nothing to use as jet fuel that's as good as jet fuel. That's why it's jet fuel.
And where do they get the electricity to reprocess the Mg02?
From an oil- or coal-burning power plant, of course.
Or a nuke plant.
These ideas of using renewable chemical fuels is all pretty silly, because they all use electricity to renew the fuel. But electric vehicles are efficient, viable, can be made attractive and fast, and they cut out the middle-man by allowing you to plug into a supply of electricity you already access. No infrastructure cost = lowest economic barrier to entry. And it's infrastructure that we have 150+ years of experience maintaining and improving.
Eventually all of our energy will be delivered from electrical utilities, generated from coal (the oil will run out soon but we have several hundred years' worth of coal left), nuclear processes (about a thousand years' worth), and the sun (several billion years, but it's terribly inefficient so far).
my upload bandwidth is a small fraction of my download bandwidth
i will not accept becoming a server for Microsoft's customers so that Microsoft can save a dollar a month on its IT bills
there are far better and smarter ways to spread out the update downloads, like actually using the scheduled-execution capabilities built into the operating system
They can certainly afford it, but it might be three degrees of separation away from their current technical knowledge set.
33% less, you mean. Futurism was better in the past. I mean, they had flying cars in their future. We don't even try, any more.
That was probably him. Small, but 100% Apple.
I've never seen a non-packed Apple store, except the one right across the street from Apple in Cupertino...
No, it's evidence that people pay good money for Vista. The reason there are no "Suse Linux Annoyances" isn't that Linux isn't annoying, it's that there's no money in evaluating and documenting its annoyances for a user base that's 1/1000th that of Windows and who got their OS for free anyway.
A better way is to get a new computer. People are clearly installing this operating system that is sized for the next 4-5 years of computing on machines they've had for the past 4-5 years, or bought cheap recently (because they contain subsystems that go back several years).
This is a fairly expensive rig now (cost me $2400 or so in parts, a day in labor, over the past month), but it runs Vista clean and fast, and within the year will probably be in the $1000-1500 retail segment:
Intel QX9650 3 GHz quad-core CPU
x38-chipset motherboard with 1333-MHz FSB
2 GB of 1333-MHz DDR-3 DRAM
1 TB of HD in 2 500-GB, RAID-0 paired, SATA-300 (i.e., SATA-II) drives.
To get the cost down, don't use a tweaker's motherboard like the ASUS Maximus Extreme; get something that still supports 1333-MHz DDR-3 DRAM and has a RAID controller for its SATA ports.
A Q6600 at 2.4 GHz ought to nearly as the QX9650 at a quarter the price. You're still blowing 9.6 billion cycles per second counting all four cores, and yes, they all get used when there's stuff to be done. A dual-core at 2.4 or even 3 GHz couldn't keep up. That's $700 off right there, and there are a couple of choices between here and the top model.
The disk system is very key. This is still the bottleneck for almost anything in the box. You rarely interact with the machine without it realizing it needs to read something new or write something important. One disk with a 4-ms nominal seek time, or a pair of disks with 10-ms seek times set up in RAID-0, pretty much appear the same.
With half-decent video gear (mine is a $60 nVidia 8400-based card) you can get Aero running without a hitch, too.
So there's the thing. People have gotten spoiled being able to buy and run $300 computers with operating systems that come stapled into magazines. If you want something new and cool, it will cost you. But you want it. You know you do. And it's not that much, really, for something that will do you a good service for several years.
yeah, if you wanted your robot's 3-d system to suck
much cheaper and faster to use stereoscopic vision
place that becomes Nigeria's Nigeria...
And there's a reason that every mall in America now has a cell-phone store or kiosk every 100 feet.
BTW, this is one of those "McJobs" that dominate the future prospects of a post-outsourced world.
Because software can pop up a box on your screen saying "go look for the sticker on the box and type the letters and numbers (and maybe the dashes or maybe not, your guess is as good as ours) you see there into this box here then click the button that says 'OK'".
Hardware says "blink"..."blink"..."blink"... and user calls customer support, adding $10 to the cost of every sale.
The ACPI settings in new installs everywhere I go generally go into low-power modes by default. I have to turn them to full-on to keep my grid computing apps up 24/7.
Because I think curing cancer and finding quasars is a bit more important than saving 20 cents' worth of electricity every day...
Which, by the way, is a waste of the amortized cost and energy used in installing the wiring, validating it's to code, developing the code, building the power plant, running the wires from that, etc., etc.
I mean, even if all my electricity were solar, maximum efficiency would require draining the batteries every night and filling them full every day.
Think holistic. Don't cut off your screensaver to spite your infrastructure.