In America (and Canada, Britain, and Australia) the law is based on an adversarial legal process. If everyone is friends, then this process doesn't really work. Theoretically, the government isn't supposed to be friends with anyone. The founding father's never trusted government, and hence they built in safeguards to protect the country from tyranny. Today's situation where the government is closely linked to large corporations is a new and different form of tyranny. Unfortunately, this was not invisaged when the founding father's wrote the constitution, and hence the courts are not set up to deal with it.
Plan 9 was Bell Labs successor to Unix. It was used to test many new concepts like Unicode and UTF-8, and has some pioneering features for clusters/networks of computers. I think most current operating systems including Windows and Linux have implemented concepts that were tested on Plan 9. For more information try plan 9 google search.
The university system was set up to preserve and expand knowledge. The tenure system works well in that regard. Most tenured professors keep doing research, and keep graduating PhDs. Witness the number of retired (Professor Emeritus) professors that are still active in their fields.
Once you realize that universities were never meant to teach large numbers of undergraduate students, then the problems start becoming obvious. What does research and tenure have to do with undergraduate teaching? If you are lucky, in the fourth year you might start to get current knowledge in most engineering programs. Everything taught before that has been known since the 1950's (and often much earlier like the 1700's). As such, current research has almost nothing to do with the undergraduate program. Even current employment trends, on the whole, have nothing to do with the curriculum of the undergraduate program at most universities. (Witness the large number of liberal-arts majors and the correspondingly small number of associated liberal-arts jobs.)
The explosion that is about to happen is that:
a) students want to pay for something that gets them a job,
b) universities were never originally structured for job training, and
c) the universities have no funding formula to pay for the practical facilities for practical job training. This means that students graduate without practical skills, and this makes them unemployable.
We are heading to a world where we have many highly-educated, unemployed, indebted and poor former university students.
rsync to the backup drive with the --backup switch. That way if the file system decides to overwrite a bunch of key files with 0 length files, your backup keeps the originals. It might take you a while to restore, but at least the originals are still there.
The backup heirarchy:
- Natural disasters occur rarely.
- Hardware failures occur infrequently.
- Software failures occur more frequently.
- User failures occur often.
- The unexpected happens all the time.
On Wallstreet, public companies must always maximize short term profits, whereas private companies can make decisions to ensure long-term profitability. Hopefully, in a big corporation, maximizing short-term profits will also maximize long-term profits. However, that does not always occur.
A good example of the difference in strategy is the American auto industry. The public companies (GM, Ford, Chrysler) routinely underperform, and often lose money. However, lots of privately held or privately controlled companies consistently make money. Magna is a good example of this. These companies keep a lid on their costs, and do not do anything to impair the long-term profits of the company.
I was at an analysts presentation on the mistakes GM, Ford, and Chrysler made. Every single mistake involved optimizing short-term profits at the expense of long-term profits. Individually, none of these decisions would have bankrupted GM. However, after a pattern of decades of short-term optimization, GM was broke.
If Dell wants to compete with HP, they only need to accept a 0.25% less per year return on investment than Wall Street. A private investor can make that decision, because he knows that if the company is well-managed, then the investment will pay off.
Having access to cheap capital in a discipllined, well-managed company is a huge advantage. The big companies engage in endlessly complicated financial manipulations to boost short-term profits. In a private company the decision is easy: focus on outcomes that maximize the long-term success of the company.
In a well-managed private company, there is no Enron-like manipulations that destroy the long-term shareholder value.
Thus, Dell can adopt a strategy where it ensures its products are competitive and sell, and then wait for HP to implode. After some of HP's recent CEOs, it is a probably a safe bet that HP will implode. That would leave Dell as the only large North American PC vendor, which would be a pretty nice place to be (for Dell).
That's not how the middle-east guys work. The CIA needs Arab spies to spy on the middle east baddies (and their is a very long list including Hezbollah, Hamas, Iran, etc.) After a while, the baddies work out who is working for the CIA, and then determine who the spies family is. Then the spy either becomes a double, or the family killed. This is a huge problem for the CIA.
This problem is also why America has made such poor progress in Afghanistan. The Taliban will wipe out your family. On the other hand, American soldiers don't go after peoples families, and mostly follow a reasonable moral code.
Thus, middle-eastern students in North America look over their shoulders, because they don't know who will go after their family back home.
Can you imagine convincing people to fight when they don't know who will threaten their families?
Collisions with bicyclists or pedestrians are the biggest source of fatalities in a modern city. Reducing the speed limit to 20 mph (30 km/h) from 30 mph (50 km/h) in the city will make a bigger impact on fatalities than setting rev-limiters to 70, 80, or 90 mph (110 - 130 km/h).
On modern superhighways, it is tough to have a high-speed collision. Thus, most of the speed related accidents occur on the slower speed city roads. After you get past the drunk/impaired/crazy driver issue, most people don't drive very fast on city roads. However, a pedestrian or a cyclist can easily be hurt or killed in a low-speed accident.
On country roads, automatic braking systems to deal with people inadvertently running stop signs, inadvertently driving in the wrong lane, or to give advance notice of approaching traffic on blind corners, would make a much bigger difference to accidents than simple rev-limiters. Cars are rated for 80 km/h head-on impacts. A head on collision involving two vehicles travelling in opposite directions is 160 km/h, or four times the energy that the safety systerms are rated to absorb. (Energy is proportional to velocity squared.) As such, you either need to prevent country road head-on collisions, or reduce the speed limit on country roads from 80 km/h to 40 km/h (which would really suck.)
Given the capabilities of modern driving technology, I can't imagine that implementing speed-governors from the 1890's is the best way of reducing road fatalities.
The concept behind TPM could work really well, if every user compiled their own operating system, and set up the unique keys such that only their code was trusted. Thus, every user would have complete control over all the source and binary software on the system. Even in a business environment, if at least the business was in complete control of all of the source and binary software, then TPM would be of some use.
The problem is that Microsoft wants to use TPM to play a bunch of DRM movies. The DRM schemes are inherently insecure, so Microsoft opens its security window accordingly. The result is that Microsoft's security model becomes "trust Microsoft, the NSA, movie companies. music companies, game companies, and etc", with no one knowing who the "etc" is. As such, from a secure systems perspective, the resulting DRM operating system has no obvious chain of accountability. Worse, any lesson in security starts with "never trust the vendors default installation." DRM assumes "never trust the customer." With the end result being that no one trusts anyone and TPM can never be secure (with commercial closed-box software.)
For TPM to truly deliver on its security promises, everyone needs to switch to open source software where everyone compiles unique binaries with custom keys. Microsoft will never do this.
The article quotes researchers delivering numbers between about 240 dpi and 477 dpi. When 300 dpi laser printers were popular, I remember being able to spot the dots. However, I had to try. Since then 600+ dpi laser printers have taken over the market, and I can't easily spot the dots with the newer high-resolution laser printers.
As such, the observations from both the print and the display researchers are consistent. Somewhere between 200 and 400 dpi the technology becomes "good enough" for many people. Somewhere between 400 and 600 dpi, the technology becomes "good enough" for almost everyone.
This often happens when the simulation results are influenced by variations in the accuracy of the built-in functions. Every floating point unit (FPU) returns an approximation of the correct result to an arbitrary level of accuracy, and the accuracy level of these results varies considerably when built-in functions like sqrt(), sin(), cos(), ln(), and exp() are considered. Normally, the accuracy of these results is pretty high. However, the initial 8087 FPU hardware from Intel was pretty old, and it necessarily made approximations.
At one point, Cyrix released an 80287 clone FPU that was faster and more accurate than Intel's 80287 equivalent. This broke many programs. Since then, Intel and AMD have been developing FPUs that are compatible with the 8087, ideally at least as accurate, and much faster. The GPU vendors have been doing something similar, however in video games, speed is more important than accuracy. For compatibility reasons (CPUs) and speed reasons (GPUs), vendors have focused on returning fast, compatible and reasonably accurate results.
In terms of accuracy, the results of the key transcendental functions, exponential functions, logarithmic functions, and the sqrt function should be viewed with suspicion. At high-accuracy levels, the least-significant bits of the results may vary considerably between processor generations, and CPU/GPU vendors. Additionally, slight differences in the results of double-precision floating point to 64-bit integer conversion functions can be detected, especially when 80-bit intermediate values are considered. Given these approximations, getting repeatable results for accuracy-sensitive simulations is tough.
It is likely that the articles weather simulations and the parent poster's simulations have differing results due to the approximations in built-in functions. Inaccuracies in the built-in functions are often much more significant that the differences due to round-off errors.
In Canada, Losec was taken off the market as Nexium launched, to ensure patients switched to the new patented drug (Nexium) before the patents on Losec expired.
Now that the Losec patents have expired, Losec is back on the market.
Apple is shipping 5.4 million iPhones and iPads per week. PC sales are at 5.9 million units per week. These numbers were pieced together from macworld and reuters. Assuming all of Apple's dreams come true, in 2015 Apple must plan for the case it is selling more processors than Intel, which will mean that Apple needs all the fab capacity it can get. Additionally, given the recent track record on new product launches at TSMC and GloFo, Apple needs a backup plan if one or more fab suppliers have problems.
Even if Apple purchased a new fab, additional reserve capacity might be needed. That may be enough to ink a deal with Samsung. Samsung is the only company with the proven ability to make enough cell phone and tablet processors to cover the majority of the world wide market, including Apple.
I think the US government has switched to an MS Windows type infrastructure for its less classified information. This effectively makes things wide open, in comparison to a well designed secure system. In particular, in Windows, if you have backup operator permissions, you have access to everything - no questions asked.
For highly-secret information, it is necessary to look at the contents of the file and previous queries before determining if an access request will be allowed. For instance, any kind of multiple download request should automatically trigger security checks. This is fundamentally different that the access locks in most commercial operating systems, because the history of previous requests affects your authorizations. Bulk downloads will trigger alarms.
Backups can be performed with secure operating systems. The backups are done with special encryption. A backup operator cannot tell which (or if) data changed by differential analysis of the encrypted backup.
I looked at this, and almost all Canadian internet traffic is monitored and goes through the US.
Firstly, all of the major Canadian ISPs peer via Chicago in one way or another. I periodically check out the connections to my website from different ISPs, and the traceroutes between ISPs. From my current location, connections to Rogers Canada (Rogers.com) and Bell Canada (Bell.ca) both route through US ISPs and Chicago. I'm pretty certain that this pattern will persist accross the country (I've tried). It would be interesting to see if a connection between Rogers' customer and Rogers.com routes through Chicago too. Somehow, I suspect it might, especially if the customer is far from Toronto.
Theoretically, you could set up conversations between two computers via the same Canadian ISP, and those would be kept in Canada. However, if I had access to well-protected information, I would discover that both Rogers and Bell have sold their souls to the Communications Security Establishment Canada (CSEC), and that organization is an active partner with the NSA. Also, Bell and Rogers both have holdings outside Canada, and as such, both probably work with the NSA directly too.
Some smaller Canadian ISPs, for example TekSavvy, are probably sufficiently small that they are not monitored actively. However, in Ontario, almost all of these smaller ISPs borrow their lines from Bell Canada. As such, if someone really wanted to monitor your conversations, they could just contact Bell Canada.
Finally, I looked at the option of encrypting all my communications between two Canadian locations with the same ISP. Bell and Rogers started throttling all encrypted communications because they assumed all of their users were bandwidth pirates and "torrenting". At the time, Rogers shut down a university research project at the University of Ottawa with this policy. Thus, in Canada, encryption is no solution.
In the end, I decided (a) everything is monitored, (b) most traffic is not monitored accurately - the ISPs are primarily interested in blocking bandwidth hogs, and (c) in Canada their is no way for a single user to stop the monitoring.
A simple sure-fire plan:
1. Outsource all of your core competencies - parts, production, everything. Keep nothing in house.
2. Profit!!!
Quietly, suppliers start selling direct to customers to make more money. 3. Find cheaper suppliers - more Profits!!!
Discover your original suppliers now sell a better product. 4. Liquidation sale! More Profits!!!
Last Step:
1. Write a business school textbook, preaching the virtues of the first 3 steps.
These problems have already happened, they caused the Blackout of 1965, and the Blackout of 2003. Essentially, whenever there is a significant mismatch between the load and the supply on the electrical grid, then a massive destabilization effect happens.
1. Excessive load will cause an excessive amount of reactive power to be drawn over certain power lines. Once this power mismatch becomes excessive, a protective device (somewhere) will trip. This safely will disconnect the power line, but will not disconnect the load from the grid, because a grid is multiply fed.
2. Other power lines on the grid will attempt to make up for the loss of the single power line. However, if the disturbance is sufficiently large, then the protective devices on these other power lines will also trip. Eventually, this will disconnect the load.
3. With the sudden loss in load, the protective devices at the power generating stations trip, because they are now supplying way to much energy to the grid, and the uncontrolled reactive energies in the grid are far in excess of what the generators can handle safely.
4. With no load, the power stations are outputting far too much energy, and must go into emergency shut-down. In the case of any thermal station (nuclear, coal, etc.), this is not good, because they are suddenly producing vast amounts of waste heat that must be quenched. The sudden changes in temperature cause the heat exchangers can go into thermal shock. Also, once cooled, any brick refractory material must be checked. This process takes days, especially in the case of nuclear generating stations where numerous rigorous safety checks must be completed. Thus, once the grid has been destabilized, it cannot restart quickly.
Larger grids deal with changing loads better than smaller grids. Specifically, if a power system has been designed to handle a 5% load mismatch, then 5% of the combined generating capacity of Ontario Hydro + the US North East grid is much larger than 5% of a single 100 MW generator. Thus, if the local steel mill suddenly spins up a 50 MW load, a big grid can handle it. A single 100MW generator will not (assuming additional loads exist). Even with a massive grid, a steel mill will call the local utility before starting massive loads, because if another problem is occurring, even a well-supplied grid might not be ready for the sudden start-up of an additional large load.
A new effect in the grid is the existence of unpredictable generators. Previous generating technologies were reliable. For instance, when hydro, coal and nuclear fired plants are started, they guarantee the supply of vast amounts of power. Some of the newer renewable technologies are unreliable. For instance, a wind farm can be at full capacity during a wind gust, and then at 20% capacity as soon as the wind gust slows. This has a huge destabilizing effect on the grid. Currently, it is estimated that unpredictable renewable power (particularly wind) can never supply more than 30% of the electrical grid. Otherwise, catastrophic destabilization cannot be avoided.
With a high-stakes career in academics, where one accusation could cause years of grief,
the rule is that you never do anything with any university-connected female that could ever be misinterpreted as sexual.
You do not ask females to go out to dinner to discuss their research. You do not invite (pester) females to visit your university, repeatedly. You do not discuss an abstruse academic point in a bar until late. You do not go to the golf course with a female co-worker (married or unmarried). You do nothing that could ever be misinterpreted, which often means you do nothing at all. This applies even if you are at a conference where the only opportunity to discuss things is late at night, or over dinner, or in a hotel room, or in a bar.
On the other hand, with a male colleague, you find a common social activity and bond.
Over the course of 15 years, subtle effects like this make a huge difference in the quality of social relationships formed between researchers in a field. Good social relationships open the doors that make good professor's famous.
It is not energy efficient to run the main rotor based on simple compressed air. Thermodynamically compressing the air and then letting it decompress results in a great deal of waste heat energy. Mechanical systems are significantly more energy inefficient.
I'm wondering if they have separated the compressor and turbine stages in a conventional jet engine in an effort to get a fuel economy or weight improvement.
The microcontrollers are not rad-hardened. The PDP with core memory and 54-series TTL logic will probably survive a small nuclear blast. There are no highly vulnerable EMI susceptable components in a PDP that I can think of. In fact, I think the military has used (does use?) this and the earlier DTL technologies in its missile computers.
For super-computing type workloads, ARM does not have a CPU fast enough to deliver the Ethernet, Infiniband, SSD, and other communications traffic to keep a Tesla fed with data.
However, Nvidia's long-term strategy must be to sell low-power and high-power ARM chips with GPU accelerators. Within 2 to 3 years, Intel will have a Xeon product that merges the existing 12-core Xeon processors with the 60-core Xeon Phi accelerators. Similarly, AMD will be building equivalent APU units with their mixed x86, ARM and GPU technologies. To be even marginally useful, Nvidia needs something to compete.
Personally, I think AMD stands a decent chance of having the fastest APUs. I think attempting to maintain cache-coherency between massive numbers of cores reduces the performance/watt advantage of the Xeon Phi. Also, if you are going to have heterogenous cores where the CPUs cannot run standard x86 code (like the Xeon Phi), then why not go fully heterogenous to maximize APU performance? Currently, AMD has the fastest merged processing units.
Intel periodically cuts patent cross-licensing deals with AMD that have the side-effect of bailing AMD out financially. This keeps AMD around as a competitor.
If Intel adopted Apple's "thermonuclear war" attitude, AMD would have been out of business from the legal fees and injunctions long ago. However, if AMD was out of business, then Intel would be a monopoly and that would be bad for Intel.
Intel manages AMD, as best it can, such that AMD gets 20% market share, and no x86 profits to speak of. With "only" 80% market share, Intel gets to keep all of the profitable market segments, with no FTC and DOJ oversight. AMD is left appealing to those who want cheap CPUs.
The distributions do periodically update to the latest versions of the software they distribute. Using 5.30 documentation on version 5.31 might work. However, that rapidly gets thin after a few years of updates.
Then again, this might be a quiet way for Oracle discontinue updates on MySQL, so that they can sell more copies of Oracle.
Slashdot Mirror
In America (and Canada, Britain, and Australia) the law is based on an adversarial legal process. If everyone is friends, then this process doesn't really work. Theoretically, the government isn't supposed to be friends with anyone. The founding father's never trusted government, and hence they built in safeguards to protect the country from tyranny. Today's situation where the government is closely linked to large corporations is a new and different form of tyranny. Unfortunately, this was not invisaged when the founding father's wrote the constitution, and hence the courts are not set up to deal with it.
Plan 9 was Bell Labs successor to Unix. It was used to test many new concepts like Unicode and UTF-8, and has some pioneering features for clusters/networks of computers. I think most current operating systems including Windows and Linux have implemented concepts that were tested on Plan 9. For more information try plan 9 google search.
The university system was set up to preserve and expand knowledge. The tenure system works well in that regard. Most tenured professors keep doing research, and keep graduating PhDs. Witness the number of retired (Professor Emeritus) professors that are still active in their fields.
Once you realize that universities were never meant to teach large numbers of undergraduate students, then the problems start becoming obvious. What does research and tenure have to do with undergraduate teaching? If you are lucky, in the fourth year you might start to get current knowledge in most engineering programs. Everything taught before that has been known since the 1950's (and often much earlier like the 1700's). As such, current research has almost nothing to do with the undergraduate program. Even current employment trends, on the whole, have nothing to do with the curriculum of the undergraduate program at most universities. (Witness the large number of liberal-arts majors and the correspondingly small number of associated liberal-arts jobs.)
The explosion that is about to happen is that:
a) students want to pay for something that gets them a job,
b) universities were never originally structured for job training, and
c) the universities have no funding formula to pay for the practical facilities for practical job training. This means that students graduate without practical skills, and this makes them unemployable.
We are heading to a world where we have many highly-educated, unemployed, indebted and poor former university students.
rsync to the backup drive with the --backup switch. That way if the file system decides to overwrite a bunch of key files with 0 length files, your backup keeps the originals. It might take you a while to restore, but at least the originals are still there.
The backup heirarchy:
- Natural disasters occur rarely.
- Hardware failures occur infrequently.
- Software failures occur more frequently.
- User failures occur often.
- The unexpected happens all the time.
On Wallstreet, public companies must always maximize short term profits, whereas private companies can make decisions to ensure long-term profitability. Hopefully, in a big corporation, maximizing short-term profits will also maximize long-term profits. However, that does not always occur.
A good example of the difference in strategy is the American auto industry. The public companies (GM, Ford, Chrysler) routinely underperform, and often lose money. However, lots of privately held or privately controlled companies consistently make money. Magna is a good example of this. These companies keep a lid on their costs, and do not do anything to impair the long-term profits of the company.
I was at an analysts presentation on the mistakes GM, Ford, and Chrysler made. Every single mistake involved optimizing short-term profits at the expense of long-term profits. Individually, none of these decisions would have bankrupted GM. However, after a pattern of decades of short-term optimization, GM was broke.
If Dell wants to compete with HP, they only need to accept a 0.25% less per year return on investment than Wall Street. A private investor can make that decision, because he knows that if the company is well-managed, then the investment will pay off.
Having access to cheap capital in a discipllined, well-managed company is a huge advantage. The big companies engage in endlessly complicated financial manipulations to boost short-term profits. In a private company the decision is easy: focus on outcomes that maximize the long-term success of the company.
In a well-managed private company, there is no Enron-like manipulations that destroy the long-term shareholder value. Thus, Dell can adopt a strategy where it ensures its products are competitive and sell, and then wait for HP to implode. After some of HP's recent CEOs, it is a probably a safe bet that HP will implode. That would leave Dell as the only large North American PC vendor, which would be a pretty nice place to be (for Dell).
That's not how the middle-east guys work. The CIA needs Arab spies to spy on the middle east baddies (and their is a very long list including Hezbollah, Hamas, Iran, etc.) After a while, the baddies work out who is working for the CIA, and then determine who the spies family is. Then the spy either becomes a double, or the family killed. This is a huge problem for the CIA.
This problem is also why America has made such poor progress in Afghanistan. The Taliban will wipe out your family. On the other hand, American soldiers don't go after peoples families, and mostly follow a reasonable moral code. Thus, middle-eastern students in North America look over their shoulders, because they don't know who will go after their family back home. Can you imagine convincing people to fight when they don't know who will threaten their families?
Collisions with bicyclists or pedestrians are the biggest source of fatalities in a modern city. Reducing the speed limit to 20 mph (30 km/h) from 30 mph (50 km/h) in the city will make a bigger impact on fatalities than setting rev-limiters to 70, 80, or 90 mph (110 - 130 km/h).
On modern superhighways, it is tough to have a high-speed collision. Thus, most of the speed related accidents occur on the slower speed city roads. After you get past the drunk/impaired/crazy driver issue, most people don't drive very fast on city roads. However, a pedestrian or a cyclist can easily be hurt or killed in a low-speed accident.
On country roads, automatic braking systems to deal with people inadvertently running stop signs, inadvertently driving in the wrong lane, or to give advance notice of approaching traffic on blind corners, would make a much bigger difference to accidents than simple rev-limiters. Cars are rated for 80 km/h head-on impacts. A head on collision involving two vehicles travelling in opposite directions is 160 km/h, or four times the energy that the safety systerms are rated to absorb. (Energy is proportional to velocity squared.) As such, you either need to prevent country road head-on collisions, or reduce the speed limit on country roads from 80 km/h to 40 km/h (which would really suck.)
Given the capabilities of modern driving technology, I can't imagine that implementing speed-governors from the 1890's is the best way of reducing road fatalities.
120 - 140 Degrees Fahrenheit
About 50 to 60 Degrees Celsius
The NSA is the new J. Edgar Hoover. They are impossible to stop. They have files on everyone.
Worse, what is their to stop the NSA from simply making something up? or saying they had something, even if they didn't?
The concept behind TPM could work really well, if every user compiled their own operating system, and set up the unique keys such that only their code was trusted. Thus, every user would have complete control over all the source and binary software on the system. Even in a business environment, if at least the business was in complete control of all of the source and binary software, then TPM would be of some use.
The problem is that Microsoft wants to use TPM to play a bunch of DRM movies. The DRM schemes are inherently insecure, so Microsoft opens its security window accordingly. The result is that Microsoft's security model becomes "trust Microsoft, the NSA, movie companies. music companies, game companies, and etc", with no one knowing who the "etc" is. As such, from a secure systems perspective, the resulting DRM operating system has no obvious chain of accountability. Worse, any lesson in security starts with "never trust the vendors default installation." DRM assumes "never trust the customer." With the end result being that no one trusts anyone and TPM can never be secure (with commercial closed-box software.)
For TPM to truly deliver on its security promises, everyone needs to switch to open source software where everyone compiles unique binaries with custom keys. Microsoft will never do this.
$1.5 million per year over 120,000 PCs works out to $12.50 per PC per year. Is anyone else getting those prices for Microsoft Office?
The article quotes researchers delivering numbers between about 240 dpi and 477 dpi. When 300 dpi laser printers were popular, I remember being able to spot the dots. However, I had to try. Since then 600+ dpi laser printers have taken over the market, and I can't easily spot the dots with the newer high-resolution laser printers.
As such, the observations from both the print and the display researchers are consistent. Somewhere between 200 and 400 dpi the technology becomes "good enough" for many people. Somewhere between 400 and 600 dpi, the technology becomes "good enough" for almost everyone.
This often happens when the simulation results are influenced by variations in the accuracy of the built-in functions. Every floating point unit (FPU) returns an approximation of the correct result to an arbitrary level of accuracy, and the accuracy level of these results varies considerably when built-in functions like sqrt(), sin(), cos(), ln(), and exp() are considered. Normally, the accuracy of these results is pretty high. However, the initial 8087 FPU hardware from Intel was pretty old, and it necessarily made approximations.
At one point, Cyrix released an 80287 clone FPU that was faster and more accurate than Intel's 80287 equivalent. This broke many programs. Since then, Intel and AMD have been developing FPUs that are compatible with the 8087, ideally at least as accurate, and much faster. The GPU vendors have been doing something similar, however in video games, speed is more important than accuracy. For compatibility reasons (CPUs) and speed reasons (GPUs), vendors have focused on returning fast, compatible and reasonably accurate results.
In terms of accuracy, the results of the key transcendental functions, exponential functions, logarithmic functions, and the sqrt function should be viewed with suspicion. At high-accuracy levels, the least-significant bits of the results may vary considerably between processor generations, and CPU/GPU vendors. Additionally, slight differences in the results of double-precision floating point to 64-bit integer conversion functions can be detected, especially when 80-bit intermediate values are considered. Given these approximations, getting repeatable results for accuracy-sensitive simulations is tough.
It is likely that the articles weather simulations and the parent poster's simulations have differing results due to the approximations in built-in functions. Inaccuracies in the built-in functions are often much more significant that the differences due to round-off errors.
In Canada, Losec was taken off the market as Nexium launched, to ensure patients switched to the new patented drug (Nexium) before the patents on Losec expired.
Now that the Losec patents have expired, Losec is back on the market.
Apple is shipping 5.4 million iPhones and iPads per week. PC sales are at 5.9 million units per week. These numbers were pieced together from macworld and reuters. Assuming all of Apple's dreams come true, in 2015 Apple must plan for the case it is selling more processors than Intel, which will mean that Apple needs all the fab capacity it can get. Additionally, given the recent track record on new product launches at TSMC and GloFo, Apple needs a backup plan if one or more fab suppliers have problems.
Even if Apple purchased a new fab, additional reserve capacity might be needed. That may be enough to ink a deal with Samsung. Samsung is the only company with the proven ability to make enough cell phone and tablet processors to cover the majority of the world wide market, including Apple.
I think the US government has switched to an MS Windows type infrastructure for its less classified information. This effectively makes things wide open, in comparison to a well designed secure system. In particular, in Windows, if you have backup operator permissions, you have access to everything - no questions asked.
For highly-secret information, it is necessary to look at the contents of the file and previous queries before determining if an access request will be allowed. For instance, any kind of multiple download request should automatically trigger security checks. This is fundamentally different that the access locks in most commercial operating systems, because the history of previous requests affects your authorizations. Bulk downloads will trigger alarms.
Backups can be performed with secure operating systems. The backups are done with special encryption. A backup operator cannot tell which (or if) data changed by differential analysis of the encrypted backup.
I looked at this, and almost all Canadian internet traffic is monitored and goes through the US.
Firstly, all of the major Canadian ISPs peer via Chicago in one way or another. I periodically check out the connections to my website from different ISPs, and the traceroutes between ISPs. From my current location, connections to Rogers Canada (Rogers.com) and Bell Canada (Bell.ca) both route through US ISPs and Chicago. I'm pretty certain that this pattern will persist accross the country (I've tried). It would be interesting to see if a connection between Rogers' customer and Rogers.com routes through Chicago too. Somehow, I suspect it might, especially if the customer is far from Toronto.
Theoretically, you could set up conversations between two computers via the same Canadian ISP, and those would be kept in Canada. However, if I had access to well-protected information, I would discover that both Rogers and Bell have sold their souls to the Communications Security Establishment Canada (CSEC), and that organization is an active partner with the NSA. Also, Bell and Rogers both have holdings outside Canada, and as such, both probably work with the NSA directly too.
Some smaller Canadian ISPs, for example TekSavvy, are probably sufficiently small that they are not monitored actively. However, in Ontario, almost all of these smaller ISPs borrow their lines from Bell Canada. As such, if someone really wanted to monitor your conversations, they could just contact Bell Canada.
Finally, I looked at the option of encrypting all my communications between two Canadian locations with the same ISP. Bell and Rogers started throttling all encrypted communications because they assumed all of their users were bandwidth pirates and "torrenting". At the time, Rogers shut down a university research project at the University of Ottawa with this policy. Thus, in Canada, encryption is no solution.
In the end, I decided (a) everything is monitored, (b) most traffic is not monitored accurately - the ISPs are primarily interested in blocking bandwidth hogs, and (c) in Canada their is no way for a single user to stop the monitoring.
A simple sure-fire plan:
1. Outsource all of your core competencies - parts, production, everything. Keep nothing in house.
2. Profit!!!
Quietly, suppliers start selling direct to customers to make more money.
3. Find cheaper suppliers - more Profits!!!
Discover your original suppliers now sell a better product.
4. Liquidation sale! More Profits!!!
Last Step:
1. Write a business school textbook, preaching the virtues of the first 3 steps.
These problems have already happened, they caused the Blackout of 1965, and the Blackout of 2003. Essentially, whenever there is a significant mismatch between the load and the supply on the electrical grid, then a massive destabilization effect happens.
1. Excessive load will cause an excessive amount of reactive power to be drawn over certain power lines. Once this power mismatch becomes excessive, a protective device (somewhere) will trip. This safely will disconnect the power line, but will not disconnect the load from the grid, because a grid is multiply fed.
2. Other power lines on the grid will attempt to make up for the loss of the single power line. However, if the disturbance is sufficiently large, then the protective devices on these other power lines will also trip. Eventually, this will disconnect the load.
3. With the sudden loss in load, the protective devices at the power generating stations trip, because they are now supplying way to much energy to the grid, and the uncontrolled reactive energies in the grid are far in excess of what the generators can handle safely.
4. With no load, the power stations are outputting far too much energy, and must go into emergency shut-down. In the case of any thermal station (nuclear, coal, etc.), this is not good, because they are suddenly producing vast amounts of waste heat that must be quenched. The sudden changes in temperature cause the heat exchangers can go into thermal shock. Also, once cooled, any brick refractory material must be checked. This process takes days, especially in the case of nuclear generating stations where numerous rigorous safety checks must be completed. Thus, once the grid has been destabilized, it cannot restart quickly.
Larger grids deal with changing loads better than smaller grids. Specifically, if a power system has been designed to handle a 5% load mismatch, then 5% of the combined generating capacity of Ontario Hydro + the US North East grid is much larger than 5% of a single 100 MW generator. Thus, if the local steel mill suddenly spins up a 50 MW load, a big grid can handle it. A single 100MW generator will not (assuming additional loads exist). Even with a massive grid, a steel mill will call the local utility before starting massive loads, because if another problem is occurring, even a well-supplied grid might not be ready for the sudden start-up of an additional large load.
A new effect in the grid is the existence of unpredictable generators. Previous generating technologies were reliable. For instance, when hydro, coal and nuclear fired plants are started, they guarantee the supply of vast amounts of power. Some of the newer renewable technologies are unreliable. For instance, a wind farm can be at full capacity during a wind gust, and then at 20% capacity as soon as the wind gust slows. This has a huge destabilizing effect on the grid. Currently, it is estimated that unpredictable renewable power (particularly wind) can never supply more than 30% of the electrical grid. Otherwise, catastrophic destabilization cannot be avoided.
With a high-stakes career in academics, where one accusation could cause years of grief, the rule is that you never do anything with any university-connected female that could ever be misinterpreted as sexual.
You do not ask females to go out to dinner to discuss their research. You do not invite (pester) females to visit your university, repeatedly. You do not discuss an abstruse academic point in a bar until late. You do not go to the golf course with a female co-worker (married or unmarried). You do nothing that could ever be misinterpreted, which often means you do nothing at all. This applies even if you are at a conference where the only opportunity to discuss things is late at night, or over dinner, or in a hotel room, or in a bar.
On the other hand, with a male colleague, you find a common social activity and bond.
Over the course of 15 years, subtle effects like this make a huge difference in the quality of social relationships formed between researchers in a field. Good social relationships open the doors that make good professor's famous.
It is not energy efficient to run the main rotor based on simple compressed air. Thermodynamically compressing the air and then letting it decompress results in a great deal of waste heat energy. Mechanical systems are significantly more energy inefficient.
I'm wondering if they have separated the compressor and turbine stages in a conventional jet engine in an effort to get a fuel economy or weight improvement.
The microcontrollers are not rad-hardened. The PDP with core memory and 54-series TTL logic will probably survive a small nuclear blast. There are no highly vulnerable EMI susceptable components in a PDP that I can think of. In fact, I think the military has used (does use?) this and the earlier DTL technologies in its missile computers.
For super-computing type workloads, ARM does not have a CPU fast enough to deliver the Ethernet, Infiniband, SSD, and other communications traffic to keep a Tesla fed with data.
However, Nvidia's long-term strategy must be to sell low-power and high-power ARM chips with GPU accelerators. Within 2 to 3 years, Intel will have a Xeon product that merges the existing 12-core Xeon processors with the 60-core Xeon Phi accelerators. Similarly, AMD will be building equivalent APU units with their mixed x86, ARM and GPU technologies. To be even marginally useful, Nvidia needs something to compete.
Personally, I think AMD stands a decent chance of having the fastest APUs. I think attempting to maintain cache-coherency between massive numbers of cores reduces the performance/watt advantage of the Xeon Phi. Also, if you are going to have heterogenous cores where the CPUs cannot run standard x86 code (like the Xeon Phi), then why not go fully heterogenous to maximize APU performance? Currently, AMD has the fastest merged processing units.
Intel periodically cuts patent cross-licensing deals with AMD that have the side-effect of bailing AMD out financially. This keeps AMD around as a competitor.
If Intel adopted Apple's "thermonuclear war" attitude, AMD would have been out of business from the legal fees and injunctions long ago. However, if AMD was out of business, then Intel would be a monopoly and that would be bad for Intel.
Intel manages AMD, as best it can, such that AMD gets 20% market share, and no x86 profits to speak of. With "only" 80% market share, Intel gets to keep all of the profitable market segments, with no FTC and DOJ oversight. AMD is left appealing to those who want cheap CPUs.
The distributions do periodically update to the latest versions of the software they distribute. Using 5.30 documentation on version 5.31 might work. However, that rapidly gets thin after a few years of updates.
Then again, this might be a quiet way for Oracle discontinue updates on MySQL, so that they can sell more copies of Oracle.