- You think that a corporation's public benefit is the only way for it to survive other than government subsidy, but you're busy preventing corporations from exploiting and damaging people - the other way they survive, and thrive. - You think that corporations don't buy assets they let their privileged elites use rather than let those elites buy them after taxes. - You think those people pay the same taxes on their capital gains with which they're paid as their labor pays on their cash. - You think corporations spend their money in local or even the American economy at the rate that their labor does, rather than spending it in foreign economies or just exclusively inside the old-boy network that has little to no multiplier effect in America. - You think that government actions that "distort the marketplace" are always bad, ignoring the ways in which the abuses of people an unregulated market always perpetrates are the primary government distortions. - You think that government subsidies that force innovations and lead private investment are damaging interference, though they're time and again the necessary leadership that put and kept this country on top, even as corporations usually tried stopping, and eventually just leeched off of, the resulting American success. - You think that the Constitution's statement that Congress has the right to raise taxes doesn't mean Congress has the right to raise taxes. - You think that reporting income for taxation is "self incrimination", though "incrimination" means "implicated in a crime". - You think taxes are "seizures", though only when tax evasion crimes are proven in a court can a court issue instructions for a seizure. - You think that "forceful coercion" is always wrong, even when people create public expenses but evade the bills in violation of the law. - You think chanting "Ron Paul 2012" does anything but discredit any good ideas that happened to stick to you.
And you think calling you nuts is "worthless". Yes, it's an ad hominem, but I have amply and repeatedly backed it up. You're totally nuts, and there's no reason to keep offering you helpful insights into how you're insane. That's worth a lot. Not to you, maybe, but to others who have to read your nutso assertions and think "is this just another crazy libertarian?" The answer is yes: you're nuts, and these are the reasons why. Just because someone calls you a name doesn't mean you haven't earned it. You have.
Profits are the money left from revenues after expenses are paid. Those expenses include taxes. Corporations own vast assets that deliver ample benefits to corporate executives and their associates, paid out of profits. When taxes rise, the corporation can choose to cut dividends to pay them. If it chooses to raise prices, that increased income would pay more taxes, so raising prices is the last way a reasonable corporation would try to make a tax neutral in effect. And raising prices means its competitors which lower profits instead have an advantage. You should also learn about price elasticity, since those competitors with lower prices will sell more units, and make a larger total profit than the high-priced one refusing to give up its high profit margins.
The market forces corporations to lower prices closer to the cost of production. That is what you free marketeers are always saying about the market finding efficiency. But when the same economics means taxing corporations, suddenly the market is powerless. Or sometimes it's the government, that all-powerful enemy of all business and all consumers, that is suddenly powerless, when it suits the cause of corporate tax evasion.
No, you've got it all mixed together like a standard-issue "libertarian".
Corporations are indeed government creations of limited liability by the people who control the corporation. But that doesn't make it a "person". What made a corporation a person was a fraudperpetrated by Southern Pacific Railroad corporation after it sued Santa Clara County to evade taxes. The judge did not rule at all on SPR's claims of corporate personhood, but the law clerk (who also worked for SPR) created a fraudulent headline reporting (in an SPR monopoly newspaper) the case's decision, claiming the court recognized SPR's corporation to be a person. A century of legal frauds have been perpetrated by corporations and the legal system they've bought referencing that headline. Corporations have only recently become bold enough to claim "rights" like "free speech", only after a century in which they perpetrated the accompanying fraud that spending money is speech.
Corporations don't have 1st Amendment rights, any more than a corporation formed in 1996 has the right to vote for Romney in Delaware next year, or one formed in 1977 has the right to challenge him and Bachmann for the Republican nomination next year. Otherwise Romney would have formed dozens of millions of corps across the country to vote for him in 2012.
The solution is not to remove government's power to protect people from corporations (or from other people who use corporations against them). That kind of extreme nonsense is the kind of binary thinking that only libertarians and corporate boards can tolerate. The solution is to properly regulate businesses, just as all government powers and actions must be proportionate and reasonable. "No regulations" means corporations have all the power and none of the liabilities, the makings of corporate anarchy. The rights of people to be protected from corporations, and basic sanity, makes this perfectly clear. That you even think corporations should be immune to government regulation means that you're not going to understand that obvious fact. But anyone else reading this might benefit from seeing it simply put.
The market is a rigged game, not an "instrument of measurement". Even a stopped clock is precisely correct twice a day. The market never stops - except when it crashes, and the public bails out the losers - and it's never correct.
I didn't complain that the market isn't a realtime measuring instrument. That's some kind of finance strawman. The amount of asset destruction by the endless "short term shocks" in the financial markets is what I'm complaining about. The markets are not useful in measuring anything but the monopoly money we allow its privateer rulers to manipulate it with by giving it to them.
I worked in the early 1990s at a startup that interconnected multiple parallel DSPs (AT&T DSP32C) by (Plessey, then Xilinx) FPGA. A lead DSP preprocessed the stream to dice it into parallel tasks, which it distributed through the FPGA to the parallel DSPs. Mainly it decomposed a single color image into independent color channels, transforming color spaces in the master thread for faster processing in the slave threads. The FPGA gradually absorbed more CLU logic from the original app that was prototyped on a 386 and ported to a single DSP, reducing the DSPs to more totally ALU logic. If the company had survived Bush Sr's 1991 recession, we probably would have had a multichip module by 1994 that worked essentially the way the Zynq does now in 2011.
I'm not sure how the AMBA bus scales across chips. But since one of the geniuses at my old company is one of the geniuses at Xilinx now, I expect that one way or another their Zynqs will scale using the onchip tech.
As for power efficiency, the Atom design targets mobile, while Zynq is reported to dissipate under 2W ("typical", maybe not with full FPGA execution), at under $15 a chip in quantity. If Zynq's FPGA executes logic more efficiently (BitsOutput:WattConsumed) than general purpose ARMs over the course of a whole day of diverse smartphone use, then mobile devices might swing over to them in a rush. Indeed Google's Dalvik JVM differs from Sun's JVM in being register oriented rather than stack oriented. If so, then FPGA matches Dalvik more closely than even ARM, and the performance:power ratio could be the best of any Android device.
The Zynq-7000 CPUs from Xilinx put a 500-1000MHz multicore ARM-C9 on a die highly interconnected by an AMBA bus to a large, fast FPGA fabric. For something like $20-50. The $:performance ratio just dropped. I saw this happen with DSPs in 1989-90, and it took quite a lot of people by surprise.
One of the top scientists at Xilinx was one of the founders of a company harnessing the DSP price drop where I worked. We also used the earliest FPGAs, but it's taken decades for the industry to get to the watershed it's just crossed with these new Intel and Xilinx CPUs. Largely because parallel processing is hard, but also because there's been so much lower hanging fruit for FPGA to consume first. But now the "FPGA peripheral" is here.
Linux on Zynq will include some IP cores for the FPGA right from the start. But very quickly developers will carve functions out of the Linux SW and instantiate them in FPGA, then call the FPGA instead. Likewise for Linux apps. Gradually versions of Linux will be distributed that have a lot of the performance bottlenecks ported from sequential instructions on the Atom/ARM to libraries implemented in the FPGA. As more SW is ported to HW, other developers will find optimizations in common configurations (eg. Linux + Chrome + LibreOffice + Google Apps/Docs) that keep the FPGA pipelines full more of the available time. The benefits will come in parallel media processing, upgradeable high-quality codecs, and higher bandwidth IPC.
Yes, conventional CPUs and GPUs will also continue to improve. But their techniques will cross pollinate CPU/FPGA hybrids, which can adopt all the CPU and GPU techniques into their more flexible HW. Likewise FPGA will augment "straight" CPUs, like onchip interconnects between memory and functional blocks that are better routed different ways for different simultaneous applications. And GPUs will benefit even more from some FPGA, with instructions customized for specific profiles of pixels, textures or vertex lists data. All processors will become more reconfigurable. They probably won't converge into an all FPGA platform, or even a single or small group of "universal CPU/FPGA/DSP" chips. There will be diversity of permanent HW instances to match the increasing diversity of tasks to which they're applied.
But FPGA embedded on mainstream chips like Atom and ARM will make FPGA techniques part of mainstream programming. It will no longer be a specialty of "HW engineers". While these hybrids will be programmable in sequential OOP like Java and C++, their features will also influence OOP into more topological statements of programs. Likely the rise of RDBMS programming, like the integration of SQLite into Android and its apps, will further influence programming into an object/relational/graph programming paradigm. The HW always leads, and developers find new (usually recombined) styles of telling the HW what to do. With FPGA easily available to regular developers, and an evolutionary path for porting existing SW to it for further improvement in the FPGA domain, that innovation is now unblocked. The rest will flow as quickly as new generations of computing always have.
No, you should fix the problem that those people have too much power over you. Your fear of public assembly and association, and preference to hide it rather than protect it, is positively un-American.
The reality is that recording public acts is never going to go away. The reality is that it is possible to protect your freedoms of public assembly and association. When you stop making excuses for your fear, you'll get more security from what threatens you.
No, corporations are privileged groups that own assets themselves. And spend on benefits for privileged people, like their executives, that are consumed without taxation. Groups that add many additional expenses to the public, like policing, investigations, courts, wars, trade diplomacy and bailouts. They are taxed at under 1% the total rate that humans are taxed in this country, but consume much more than half the expensive operations.
Money they don't pay in taxes they spend in foreign economies more than they spend here, and on Americans who are much less productive (bankers, lawyers, consultants, media spinners) than people to whom they must pay for their core operations.
The current corporate profits are higher than ever before in American history. Corporations are spending on neither current nor future employees (as we've seen for going on 3-5 years now). They aren't even investing in other corporate equity. With the exception of pure financial speculation, which keeps crashing the stock market and destroying real assets with every cycle. And the new exception of unlimited, secret and anonymous bribes, er, donations to politicians and their henchmen, er, henchcorps.
Somehow when getting government benefits corporations are people, but when facing liabilities like convictions and taxes corporations are not people. That is an obvious scam, that has brought this country just past the brink of catastrophe several times in just the past decade.
No they are not. They just responded sarcastically to your libertarian assertions. Libertarians like you are corporate anarchists, and the Greater Somalia they described is what your ideas would get us. As they always have everywhere they've been practiced.
Corporations cost the public a lot of money. They should pay taxes. The people who profit from them also cost the public a lot of money, which is why they should pay taxes.
Corporations pay lots of their income after operating expenses on lots of material benefits to their people, especially their executives and those people's associated staff.
Government business subsidies sometimes have necessary strategic benefits, though usually the subsidies are just corruption. The government regulating businesses is one of the most important jobs of government. Unregulated business has a perfectly unbroken record of destroying economies and people. You "libertarians" are really totally out of your minds.
Now income tax is a somewhat nonsensical basis for taxation. It's exceeded in nonsense only by residential property taxes, which are a legacy of centuries ago when land owned was roughly proportional to potential income - a way to pressure landowners to produce from their land, and to collect taxes without poormouth excuses - but that has long been false for the large majority of landowners. The sensible tax is a consumption tax, a sales tax that excludes necessities (cost of 20%ile residence in a zipcode, heat/power/telecom of that 20%ile mark, public education and transit, minimum healthcare expenses). People should pay proportional to the material benefit they get from the government that protects and enables that benefit, beyond the mere subsistence that government should take no credit for leaving people at. A 20% sales tax on non-necessities, plus a 0.1% tax on all equity sales that don't transfer control of the asset (at which 50%+1 point all cumulative sales tax is paid at 20%), would pay the US Federal budget, with a surplus paying down the debt left from the days of nonsensical taxation. Of course the military/intel waste would have to be cut from $1.5T+ to under $300B, but there's plenty of nonsense to eliminate.
If you "Libertarians" would get a grip on reality, your vocal energy would be well used getting us off income and property taxes and onto a pure sales tax. But when you say government has no business regulating business, you just discredit any good ideas you touched in your random walk outside the mainstream.
Government auditors of apparent ponzi schemes and their enabling accountants (on either side of the cash flow, including defrauded investors) are just one of the major expenses that corporate taxes should be raised to cover. Defrayed by fines actually recovered from the people who extracted the money from the corporate vehicles.
The corps that move overseas lose the benefits of being US corps. Or at least they should - that's the other part of the rigged game corps have set up in the USA in the past couple of generations. Those of us who are actually Americans should protect ourselves from those foreign corps that just exploit us.
But the threat you cite is not a real one. The corps don't move overseas now to avoid taxes, though there are places where they could. Because those places have other expenses and threats to corporations. Reduced transparency, corruption, productivity caps of suppressed labor, environmental exhaustion, etc. And the people who own and run these corps don't want to live places where society isn't protected at expenses fairly carried by those cause it.
The entire flow is that corporations have floated these threats, but cannot exercise them. Meanwhile they've inhibited the American people from discussing the counterthreats, and therefore from exercising them. Charging the corps their carrying cost is long overdue, and far more practical than the reverse that the corps have pushed us into.
You and I aren't the only ones discussing the real effects of taxing corporations (and their owners) more. Many of us are discussing it. The more of us posting it to the Web and each other the sooner we"ll do the reasonable thing and fix it>
Then don't do those things in public. Privacy is for private places.
And if those public acts enable some people to unfairly affect you, then the problem is those people's unfair power over you. That's the place to resist. Not in pretending that public acts have privacy expectations. At the very least because those people's power will be used to exploit what's available to the public to the maximum extent, unless you limit their power.
That's why Xilinx added a 500MHz FPGA tightly integrated as a bus peripheral of a multicore ARM-C9. The Zynq-7000 will cost somewhere between $20 and $50, depending on the model (and amount of FPGA).
Both the Atom and Zynq FPGA versions are about to make FPGA as mass market (embedded, in cars and industrial control, then in multimedia workstations) as DSPs have become. Every PC has multiple DSPs, at least in soundcards and often in video systems and sometimes in network interfaces; hard drives often have them, too.
I described how FPGA will evolve. I believe the Atom and Zynq chips are the first flow across the watershed. Within a matter of years FPGA will be far cheaper and more powerful than purely sequential CPUs, though as available to mainstream developers.
Do you think the Xilinx Zynq will be (relatively) easily scalable by adding more FPGA chips across an extended AMBA bus? That architecture has FPGA highly integrated on the CPU die by high bandwidth of that highly flexible bus; the Atom is linked to its onchip FPGA by only a couple of serial lines. AMBA is really the disruptive tech at work in this space.
Maybe the multicore ARMs are able to scale distributing processing across offchip cores, with a lot of FPGA fabric and an excellent bus supporting the glue. More gates executing against more data in parallel rather than faster clocks has kept Moore's law predicting "sequential" CPU performance. It should do the same for FPGA.
The measure of cash flow as profits is fundamentally wrong. The fact that profit accounting can be rigged doesn't make cash flow a better profit measure, since cash flow also can be (and is) rigged to cook the books. Hiding debt from profits or cash flow is done all the time by corrupt corporations.
There's no accounting trick that makes any accounting technique more resistant to book cooking. Only audits by parties with overriding access, the interest in catching fake books, and the power to report them make for reliable profits reporting. Keeping "profits" defined simply and accurately as "revenue from expenses, minus the expenses" is the only way to meaningfully report profits.
I didn't say it did. I just pointed out that new FPGA hybrid CPUs will see much more downloadable upgrades to CPUs than this pioneering act that is the subject of the story.
No, Apple invented its versions of each of those, which revolutionized each of their industries. Most innovation, and nearly all the most revolutionary innovation, is not the original invention but rather an improvement of it.
It's going to happen, since it's already happening and the benefits will only increase as it happens more. "Permanent circuits" are going to become one more security assumption that cannot be made a basis of trust or reliability. Which is yet another reason that SW and data security techniques and practices must be made ever more comprehensive and practicable.
The future of infosystem security is inevitably going to be nasty. It's up to those securing it to be nastier than those attacking it.
I'll repeat my answer: multitasking is a sequential CPU simulation of the parallel processing that is actual on FPGA. Typically such rotation is not practiced on FPGA, since the actual parallelism is sufficient without simulating it.
The reconfig time depends on which reconfig technique is used. In a typical FPGA (different makes/models differ in their behavior), each gate can reconfig in one clock cycle (up to 0.5-1GHz); like everything else in FPGA, all gates can change all at once, in parallel. The existing config can modify itself in microseconds. If loading an entire new config netlist, that's limited by the IO bandwidth, which is bottlenecked by the bus to the gates and the IO port. Bus speed is often Tbps on chips like the Intel Atom and Xilinx Zynq I mentioned, which is therefore not usually the bottleneck. The IO port is part of the reconfigurable HW on the chip; it's physically limited in width by the number of assignable chip pins connected to memory in which config "netlists" are stored. A chip that relies on such reconfig speed could have dozens of pins dedicated to such reconfig. So a 512KB gates netlist delivered across say 64 pins at 500MHz could take about 131us, or about 7630 times a second.
But again, that game's various custom instructions would probably all fit within a single large FPGA without reconfiguration. Even several concurrent processes' set of custom instructions (even without considering overlap with other processes') could fit on a single large FPGA. And FPGA is largely scalable; add more FPGA chips glued to each other by their pins. Then one FPGA routes data to the right FPGA which has the appropriate circuits on it for the data at that time. As FPGA becomes ever cheaper and more dense, the need for reconfig to "task switch" becomes even less.
Slashdot Mirror
You Ron Paul libertarians are completely nuts:
- You think that a corporation's public benefit is the only way for it to survive other than government subsidy, but you're busy preventing corporations from exploiting and damaging people - the other way they survive, and thrive.
- You think that corporations don't buy assets they let their privileged elites use rather than let those elites buy them after taxes.
- You think those people pay the same taxes on their capital gains with which they're paid as their labor pays on their cash.
- You think corporations spend their money in local or even the American economy at the rate that their labor does, rather than spending it in foreign economies or just exclusively inside the old-boy network that has little to no multiplier effect in America.
- You think that government actions that "distort the marketplace" are always bad, ignoring the ways in which the abuses of people an unregulated market always perpetrates are the primary government distortions.
- You think that government subsidies that force innovations and lead private investment are damaging interference, though they're time and again the necessary leadership that put and kept this country on top, even as corporations usually tried stopping, and eventually just leeched off of, the resulting American success.
- You think that the Constitution's statement that Congress has the right to raise taxes doesn't mean Congress has the right to raise taxes.
- You think that reporting income for taxation is "self incrimination", though "incrimination" means "implicated in a crime".
- You think taxes are "seizures", though only when tax evasion crimes are proven in a court can a court issue instructions for a seizure.
- You think that "forceful coercion" is always wrong, even when people create public expenses but evade the bills in violation of the law.
- You think chanting "Ron Paul 2012" does anything but discredit any good ideas that happened to stick to you.
And you think calling you nuts is "worthless". Yes, it's an ad hominem, but I have amply and repeatedly backed it up. You're totally nuts, and there's no reason to keep offering you helpful insights into how you're insane. That's worth a lot. Not to you, maybe, but to others who have to read your nutso assertions and think "is this just another crazy libertarian?" The answer is yes: you're nuts, and these are the reasons why. Just because someone calls you a name doesn't mean you haven't earned it. You have.
Goodbye.
Profits are the money left from revenues after expenses are paid. Those expenses include taxes. Corporations own vast assets that deliver ample benefits to corporate executives and their associates, paid out of profits. When taxes rise, the corporation can choose to cut dividends to pay them. If it chooses to raise prices, that increased income would pay more taxes, so raising prices is the last way a reasonable corporation would try to make a tax neutral in effect. And raising prices means its competitors which lower profits instead have an advantage. You should also learn about price elasticity, since those competitors with lower prices will sell more units, and make a larger total profit than the high-priced one refusing to give up its high profit margins.
The market forces corporations to lower prices closer to the cost of production. That is what you free marketeers are always saying about the market finding efficiency. But when the same economics means taxing corporations, suddenly the market is powerless. Or sometimes it's the government, that all-powerful enemy of all business and all consumers, that is suddenly powerless, when it suits the cause of corporate tax evasion.
No, you've got it all mixed together like a standard-issue "libertarian".
Corporations are indeed government creations of limited liability by the people who control the corporation. But that doesn't make it a "person". What made a corporation a person was a fraud perpetrated by Southern Pacific Railroad corporation after it sued Santa Clara County to evade taxes. The judge did not rule at all on SPR's claims of corporate personhood, but the law clerk (who also worked for SPR) created a fraudulent headline reporting (in an SPR monopoly newspaper) the case's decision, claiming the court recognized SPR's corporation to be a person. A century of legal frauds have been perpetrated by corporations and the legal system they've bought referencing that headline. Corporations have only recently become bold enough to claim "rights" like "free speech", only after a century in which they perpetrated the accompanying fraud that spending money is speech.
Corporations don't have 1st Amendment rights, any more than a corporation formed in 1996 has the right to vote for Romney in Delaware next year, or one formed in 1977 has the right to challenge him and Bachmann for the Republican nomination next year. Otherwise Romney would have formed dozens of millions of corps across the country to vote for him in 2012.
The solution is not to remove government's power to protect people from corporations (or from other people who use corporations against them). That kind of extreme nonsense is the kind of binary thinking that only libertarians and corporate boards can tolerate. The solution is to properly regulate businesses, just as all government powers and actions must be proportionate and reasonable. "No regulations" means corporations have all the power and none of the liabilities, the makings of corporate anarchy. The rights of people to be protected from corporations, and basic sanity, makes this perfectly clear. That you even think corporations should be immune to government regulation means that you're not going to understand that obvious fact. But anyone else reading this might benefit from seeing it simply put.
The market is a rigged game, not an "instrument of measurement". Even a stopped clock is precisely correct twice a day. The market never stops - except when it crashes, and the public bails out the losers - and it's never correct.
I didn't complain that the market isn't a realtime measuring instrument. That's some kind of finance strawman. The amount of asset destruction by the endless "short term shocks" in the financial markets is what I'm complaining about. The markets are not useful in measuring anything but the monopoly money we allow its privateer rulers to manipulate it with by giving it to them.
I worked in the early 1990s at a startup that interconnected multiple parallel DSPs (AT&T DSP32C) by (Plessey, then Xilinx) FPGA. A lead DSP preprocessed the stream to dice it into parallel tasks, which it distributed through the FPGA to the parallel DSPs. Mainly it decomposed a single color image into independent color channels, transforming color spaces in the master thread for faster processing in the slave threads. The FPGA gradually absorbed more CLU logic from the original app that was prototyped on a 386 and ported to a single DSP, reducing the DSPs to more totally ALU logic. If the company had survived Bush Sr's 1991 recession, we probably would have had a multichip module by 1994 that worked essentially the way the Zynq does now in 2011.
I'm not sure how the AMBA bus scales across chips. But since one of the geniuses at my old company is one of the geniuses at Xilinx now, I expect that one way or another their Zynqs will scale using the onchip tech.
As for power efficiency, the Atom design targets mobile, while Zynq is reported to dissipate under 2W ("typical", maybe not with full FPGA execution), at under $15 a chip in quantity. If Zynq's FPGA executes logic more efficiently (BitsOutput:WattConsumed) than general purpose ARMs over the course of a whole day of diverse smartphone use, then mobile devices might swing over to them in a rush. Indeed Google's Dalvik JVM differs from Sun's JVM in being register oriented rather than stack oriented. If so, then FPGA matches Dalvik more closely than even ARM, and the performance:power ratio could be the best of any Android device.
The Zynq-7000 CPUs from Xilinx put a 500-1000MHz multicore ARM-C9 on a die highly interconnected by an AMBA bus to a large, fast FPGA fabric. For something like $20-50. The $:performance ratio just dropped. I saw this happen with DSPs in 1989-90, and it took quite a lot of people by surprise.
One of the top scientists at Xilinx was one of the founders of a company harnessing the DSP price drop where I worked. We also used the earliest FPGAs, but it's taken decades for the industry to get to the watershed it's just crossed with these new Intel and Xilinx CPUs. Largely because parallel processing is hard, but also because there's been so much lower hanging fruit for FPGA to consume first. But now the "FPGA peripheral" is here.
Linux on Zynq will include some IP cores for the FPGA right from the start. But very quickly developers will carve functions out of the Linux SW and instantiate them in FPGA, then call the FPGA instead. Likewise for Linux apps. Gradually versions of Linux will be distributed that have a lot of the performance bottlenecks ported from sequential instructions on the Atom/ARM to libraries implemented in the FPGA. As more SW is ported to HW, other developers will find optimizations in common configurations (eg. Linux + Chrome + LibreOffice + Google Apps/Docs) that keep the FPGA pipelines full more of the available time. The benefits will come in parallel media processing, upgradeable high-quality codecs, and higher bandwidth IPC.
Yes, conventional CPUs and GPUs will also continue to improve. But their techniques will cross pollinate CPU/FPGA hybrids, which can adopt all the CPU and GPU techniques into their more flexible HW. Likewise FPGA will augment "straight" CPUs, like onchip interconnects between memory and functional blocks that are better routed different ways for different simultaneous applications. And GPUs will benefit even more from some FPGA, with instructions customized for specific profiles of pixels, textures or vertex lists data. All processors will become more reconfigurable. They probably won't converge into an all FPGA platform, or even a single or small group of "universal CPU/FPGA/DSP" chips. There will be diversity of permanent HW instances to match the increasing diversity of tasks to which they're applied.
But FPGA embedded on mainstream chips like Atom and ARM will make FPGA techniques part of mainstream programming. It will no longer be a specialty of "HW engineers". While these hybrids will be programmable in sequential OOP like Java and C++, their features will also influence OOP into more topological statements of programs. Likely the rise of RDBMS programming, like the integration of SQLite into Android and its apps, will further influence programming into an object/relational/graph programming paradigm. The HW always leads, and developers find new (usually recombined) styles of telling the HW what to do. With FPGA easily available to regular developers, and an evolutionary path for porting existing SW to it for further improvement in the FPGA domain, that innovation is now unblocked. The rest will flow as quickly as new generations of computing always have.
No, you should fix the problem that those people have too much power over you. Your fear of public assembly and association, and preference to hide it rather than protect it, is positively un-American.
The reality is that recording public acts is never going to go away. The reality is that it is possible to protect your freedoms of public assembly and association. When you stop making excuses for your fear, you'll get more security from what threatens you.
You're right. That's why I have an accountant do my taxes :).
No, corporations are privileged groups that own assets themselves. And spend on benefits for privileged people, like their executives, that are consumed without taxation. Groups that add many additional expenses to the public, like policing, investigations, courts, wars, trade diplomacy and bailouts. They are taxed at under 1% the total rate that humans are taxed in this country, but consume much more than half the expensive operations.
Money they don't pay in taxes they spend in foreign economies more than they spend here, and on Americans who are much less productive (bankers, lawyers, consultants, media spinners) than people to whom they must pay for their core operations.
The current corporate profits are higher than ever before in American history. Corporations are spending on neither current nor future employees (as we've seen for going on 3-5 years now). They aren't even investing in other corporate equity. With the exception of pure financial speculation, which keeps crashing the stock market and destroying real assets with every cycle. And the new exception of unlimited, secret and anonymous bribes, er, donations to politicians and their henchmen, er, henchcorps.
Somehow when getting government benefits corporations are people, but when facing liabilities like convictions and taxes corporations are not people. That is an obvious scam, that has brought this country just past the brink of catastrophe several times in just the past decade.
No they are not. They just responded sarcastically to your libertarian assertions. Libertarians like you are corporate anarchists, and the Greater Somalia they described is what your ideas would get us. As they always have everywhere they've been practiced.
Corporations cost the public a lot of money. They should pay taxes. The people who profit from them also cost the public a lot of money, which is why they should pay taxes.
Corporations pay lots of their income after operating expenses on lots of material benefits to their people, especially their executives and those people's associated staff.
Government business subsidies sometimes have necessary strategic benefits, though usually the subsidies are just corruption. The government regulating businesses is one of the most important jobs of government. Unregulated business has a perfectly unbroken record of destroying economies and people. You "libertarians" are really totally out of your minds.
Now income tax is a somewhat nonsensical basis for taxation. It's exceeded in nonsense only by residential property taxes, which are a legacy of centuries ago when land owned was roughly proportional to potential income - a way to pressure landowners to produce from their land, and to collect taxes without poormouth excuses - but that has long been false for the large majority of landowners. The sensible tax is a consumption tax, a sales tax that excludes necessities (cost of 20%ile residence in a zipcode, heat/power/telecom of that 20%ile mark, public education and transit, minimum healthcare expenses). People should pay proportional to the material benefit they get from the government that protects and enables that benefit, beyond the mere subsistence that government should take no credit for leaving people at. A 20% sales tax on non-necessities, plus a 0.1% tax on all equity sales that don't transfer control of the asset (at which 50%+1 point all cumulative sales tax is paid at 20%), would pay the US Federal budget, with a surplus paying down the debt left from the days of nonsensical taxation. Of course the military/intel waste would have to be cut from $1.5T+ to under $300B, but there's plenty of nonsense to eliminate.
If you "Libertarians" would get a grip on reality, your vocal energy would be well used getting us off income and property taxes and onto a pure sales tax. But when you say government has no business regulating business, you just discredit any good ideas you touched in your random walk outside the mainstream.
Saying profits was mentioned, but called "cash flow", does say that they are the same.
This story and its discussion have wrong just oozing out of every one of their many cracks.
Government auditors of apparent ponzi schemes and their enabling accountants (on either side of the cash flow, including defrauded investors) are just one of the major expenses that corporate taxes should be raised to cover. Defrayed by fines actually recovered from the people who extracted the money from the corporate vehicles.
I think you replied to the wrong comment, since to mine yours is a non sequitur.
You should also login with a username, since "Anonymous Coward" posts weaken your argument.
But the changes is causes can be revolutionary, even if the innovation was evolutionary. That is in fact the definition of "tipping point".
Put the crackpipe down, Republican loafer.
The corps that move overseas lose the benefits of being US corps. Or at least they should - that's the other part of the rigged game corps have set up in the USA in the past couple of generations. Those of us who are actually Americans should protect ourselves from those foreign corps that just exploit us.
But the threat you cite is not a real one. The corps don't move overseas now to avoid taxes, though there are places where they could. Because those places have other expenses and threats to corporations. Reduced transparency, corruption, productivity caps of suppressed labor, environmental exhaustion, etc. And the people who own and run these corps don't want to live places where society isn't protected at expenses fairly carried by those cause it.
The entire flow is that corporations have floated these threats, but cannot exercise them. Meanwhile they've inhibited the American people from discussing the counterthreats, and therefore from exercising them. Charging the corps their carrying cost is long overdue, and far more practical than the reverse that the corps have pushed us into.
You and I aren't the only ones discussing the real effects of taxing corporations (and their owners) more. Many of us are discussing it. The more of us posting it to the Web and each other the sooner we"ll do the reasonable thing and fix it>
Then don't do those things in public. Privacy is for private places.
And if those public acts enable some people to unfairly affect you, then the problem is those people's unfair power over you. That's the place to resist. Not in pretending that public acts have privacy expectations. At the very least because those people's power will be used to exploit what's available to the public to the maximum extent, unless you limit their power.
That's why Xilinx added a 500MHz FPGA tightly integrated as a bus peripheral of a multicore ARM-C9. The Zynq-7000 will cost somewhere between $20 and $50, depending on the model (and amount of FPGA).
Both the Atom and Zynq FPGA versions are about to make FPGA as mass market (embedded, in cars and industrial control, then in multimedia workstations) as DSPs have become. Every PC has multiple DSPs, at least in soundcards and often in video systems and sometimes in network interfaces; hard drives often have them, too.
I described how FPGA will evolve. I believe the Atom and Zynq chips are the first flow across the watershed. Within a matter of years FPGA will be far cheaper and more powerful than purely sequential CPUs, though as available to mainstream developers.
Do you think the Xilinx Zynq will be (relatively) easily scalable by adding more FPGA chips across an extended AMBA bus? That architecture has FPGA highly integrated on the CPU die by high bandwidth of that highly flexible bus; the Atom is linked to its onchip FPGA by only a couple of serial lines. AMBA is really the disruptive tech at work in this space.
Maybe the multicore ARMs are able to scale distributing processing across offchip cores, with a lot of FPGA fabric and an excellent bus supporting the glue. More gates executing against more data in parallel rather than faster clocks has kept Moore's law predicting "sequential" CPU performance. It should do the same for FPGA.
The measure of cash flow as profits is fundamentally wrong. The fact that profit accounting can be rigged doesn't make cash flow a better profit measure, since cash flow also can be (and is) rigged to cook the books. Hiding debt from profits or cash flow is done all the time by corrupt corporations.
There's no accounting trick that makes any accounting technique more resistant to book cooking. Only audits by parties with overriding access, the interest in catching fake books, and the power to report them make for reliable profits reporting. Keeping "profits" defined simply and accurately as "revenue from expenses, minus the expenses" is the only way to meaningfully report profits.
I didn't say it did. I just pointed out that new FPGA hybrid CPUs will see much more downloadable upgrades to CPUs than this pioneering act that is the subject of the story.
No, Apple invented its versions of each of those, which revolutionized each of their industries. Most innovation, and nearly all the most revolutionary innovation, is not the original invention but rather an improvement of it.
It's going to happen, since it's already happening and the benefits will only increase as it happens more. "Permanent circuits" are going to become one more security assumption that cannot be made a basis of trust or reliability. Which is yet another reason that SW and data security techniques and practices must be made ever more comprehensive and practicable.
The future of infosystem security is inevitably going to be nasty. It's up to those securing it to be nastier than those attacking it.
I'll repeat my answer: multitasking is a sequential CPU simulation of the parallel processing that is actual on FPGA. Typically such rotation is not practiced on FPGA, since the actual parallelism is sufficient without simulating it.
The reconfig time depends on which reconfig technique is used. In a typical FPGA (different makes/models differ in their behavior), each gate can reconfig in one clock cycle (up to 0.5-1GHz); like everything else in FPGA, all gates can change all at once, in parallel. The existing config can modify itself in microseconds. If loading an entire new config netlist, that's limited by the IO bandwidth, which is bottlenecked by the bus to the gates and the IO port. Bus speed is often Tbps on chips like the Intel Atom and Xilinx Zynq I mentioned, which is therefore not usually the bottleneck. The IO port is part of the reconfigurable HW on the chip; it's physically limited in width by the number of assignable chip pins connected to memory in which config "netlists" are stored. A chip that relies on such reconfig speed could have dozens of pins dedicated to such reconfig. So a 512KB gates netlist delivered across say 64 pins at 500MHz could take about 131us, or about 7630 times a second.
But again, that game's various custom instructions would probably all fit within a single large FPGA without reconfiguration. Even several concurrent processes' set of custom instructions (even without considering overlap with other processes') could fit on a single large FPGA. And FPGA is largely scalable; add more FPGA chips glued to each other by their pins. Then one FPGA routes data to the right FPGA which has the appropriate circuits on it for the data at that time. As FPGA becomes ever cheaper and more dense, the need for reconfig to "task switch" becomes even less.