Intel's going to win anyway for the forseeable future, as AMD doesn't have anywhere near the fab capacity to meet the needs of Intel's market.
AMD would either have to outsource, which sucks because the process (but necessarily process size) *will* be different if they have TSMC or some other fab-for-hire cranking chips for them, or they need to blow a heck of a lot of cash building their own fabs, which is pretty risky since it takes a long time to build a fab and unused fab space is a huge money sink.
Everybody knows that communications engineers do it the right way, mixing big-endian and little-endian seemingly at random in order take best advantage of both systems.
After all, it makes perfect sense to byte-cross your data, send it to the ATM scrambler, bit-cross the result, feed it to your physical layer scrambler, and then bit-cross that result in order to calculate your CRC.
I now expect you to hang your head in shame for questioning the superior wisdom of communication engineers.
Blacklight Power may or may not be a hoax. I tend to think they may be onto something, as they have published some peer-reviewed papers (see their web-site, lots of stuff, some reviewed), but then again, it may not pan out or be completely make-believe. Either way, their tech doesn't violate thermodynamics in any fundamental way.
Essentially, with their tech, you'd take your fuel (in this case, water), and input some energy to start their reaction. Then, using thethe excess energy from the reaction would be used to start further reactions in the fuel supply, while still leaving some energy left over for useful work. This isn't perpetual motion, though, since, when you're out of fuel, its game over. They are very upfront about the fact that it would take more energy to to make more water (fuel) than what was generated by the process.
The novel compounds appear to be a useful byproduct, but cannot be re-used as fuel.
Hardware design skills are good for a programmer to have, especially if this guy wants to work on embedded stuff as a career.
I program a custom DSP, and let me tell you, the most useful classes I took in college (not GT)were my digital design classes. Programming comes pretty natural to me (as it does, I'm sure, for a lot of CS majors), but taking those classes widened my skillset quite a bit.
If I'm getting weird ass behavior out of my programs, I can grab the chip design and figure out WTF is going on (not HDL of any sort, our chip is schematically laid out in an ancient tool). If I need a new instruction in next chip, I have some idea of what is and isn't possible and can suggest ways of making it work. If I need to change the behavior on my dev board, I can grab the VHDL for the FPGA and change it. Very, very useful, allows me to get my programming done rather than waiting for other, extremely busy people to help me out.
So really, there's no reason for him to convert to the darkside. A CS with a decent digital hardware background is a good fit in the embedded world.
I had to remove a flash chip from a socket, but couldn't find the right tool, so I grabbed an 8 inch soldering pick (the kind that look like dentists tools) and proceeded to lift the flash. Unbeknownst to me, I bridged the tool to the unrectified (110V AC) power supply exposed at the switch of my dev board.
Zap! Flash! Bang!
Tossed my sorry ass about three feet backwards while the flash chip burst into flames and jumped about a foot into the air:) After recovering, I took a look at the chip and the packaging was completely gutted in the middle where (presumably) the (ex-)die was.
uh, dsl goes to about 18Kft at 384K. 10 Mb/s, while not standardized really, is easily done to about 7 kft.
Tim
Re:How does devaluing happen now??
on
The Euro
·
· Score: 1
Yup, that's my worry too.
Unifying the currency before unifying the individual economies and the government debt loads seems like a bad idea. It may not matter so much for the bigger/more-stable players, but for Italy or Spain I can forsee problems as they will want to devalue to ease debt burdens and increase competitiveness while the more stable economies will block any effort at devaluation in order to avoid inflation.
Despite the cries of Europeans to the contrary, this *is* going to create a winner-take-all situation, with Germany (likely) being the winner. There is an inequality between the economies and being locked into the same currency as Germany and France will cause rampant deflation in the less developed economies as they struggle to compete, much like how Argentina screwed themselves by locking their currency to the US dollar and keeping it that way (and refusing to devalue despite the urgings, of, oh, everybody else).
As someone else mentioned above, there is also the problem of job competition. WHile in theory, anybody from any Euro country can move to any other Euro country, in practice its not going to happen that way. It doesn't even happen that way in the US, where everybody speaks the same language. Try telling one of the laid off Lucent engineers living in the posh suburb of Holmdel NJ that you've got the perfect job for them in some backwater town in Mississippi (or vice-versa). The response, I assure you, will be negative for a number of reasons, namely culture difference and the fact that they will have to take a salary cut. And while the lower salary may represent greater purchasing power in Mississippi than their high one did in the shadow of NYC, not everything gets locally adjusted (cars, for instance, I grew up in rural PA and while $40K/yr there was theoretically equivalent to $60K/yr here in NJ, the lower salary doesn't buy a Mercedes, though it might get a bigger house than the $60K salary in NJ) and people, in general don't like pay cuts.
Now take that and apply it to even greater economic and cultural inequality in Europe, and I can see some real problems ahead.
I went through the same thing a year earlier and in a better economy. I decide to drop out of college, my fiance (at the time) calls me all sorts of nasty words for loser, we break up, I get an incredible job and drive over to her apartment in my new Beemer:)
Anyway, in this economy, the likelihood of that last part is somewhat in doubt, but just keep pushing and trying, and you'll make it. I had to pretty much bully my way into interviews, puching an 80x86 assembly version of space invaders over my resume, but I eventually got hired by a great, profitable company. And once I was there, I managed to prove myself by being the the most hardworking mofo around.
Now, I get to settle down a bit (though the stress does occassionally build up, like now), and no one really cares whether I've got a Masters in EE studying signal processing, or that I can do the work just as well as someone who does have the paper.
Dude, that's probably the most clueless thing I've ever read.
First of all lots of real americans do benefit from defense spending. I grew up in a blue-collar, rural area and this area, and many, many more just like it were very dependent on defense spending. Of the 10000 people in my hometown, around 1500 worked at the local foundries, forges, and steel mills. They had good paying union jobs and unemployment hovered around 0 in the area.
However, these companies did 60-90% of their business for the defense industry (US firms just aren't competitive for commercial products, but are necessary for defense products). When the defense build-up of the 80's ended, the area went into a severe depression. *All* of the foundries, forges, mills, and machine shops went bankrupt. Everybody, from the owners to the guys grinding castings, was out of a job. By 1994, unemployment was around 25% and the population had been reduced to less than 4000.
And people wonder why rural areas vote republican? I consider myself and my family pretty average americans that benefitted immensely from military spending. Later, while the coasts were enjoying the boom of the late 90's, rural areas were struggling to keep their unemployment under 15%.
Secondly, this isn't just make-work. There are tangible benfits from having a strong military and the industrial capacity to bolster it in case of, you know, an actual major war. With the US foundry and forge industries completely decimated, it'll be extremely tough if there actually is a need to crank out new planes on a daily basis to replace ones getting shot down.
Also, the reason the rest of the US friendly world has the luxury of not spending money on their militaries is that the US does it for them. Do you really think the Soviets would have felt constrained from taking the rest of Europe if the NATO didn't have massive amounts of US armor to make it a risky propostition? Do you think that the French could hold stop anyone given the current state of their military?
According to Goedel's Incompleteness Theorem, there are two equally valid mathematical systems, one that is consistent but incomplete (2 equations with 3 variables cannot be solved), and one that is complete but inconsistent (the above system can be solved, but it means that essentially true==false under certain conditions).
Men use the consistent system, women use the complete system. When a woman asks "Honey, does my ass look big in these?", according to her complete system there is a valid answer, despite the fact that it may not be consistent with other questions (ex. "So, how much do you think I weigh?") . Men, however are confined to a consistent system and unable to answer these questions correctly, and are left sitting there unable to speak for fear of answering wrong.
Dude, this guy isn't just doing processing with an asynchronous chip, he's doing freaking tone detection with it.
The actual logic could be asynchronous for all I care, but as a DSP programmer I will confidently state that doing tone detection without at least one known, fixed clock in the circuit is just plain deranged.
I'm in the same boat: An embedded programmer without a college degree, and I get cold calls from my company's competitors all the time.
Somehow, I think the embedded world may be different because it takes a unique skillset to be able to work on embedded processors with barely enough enough memory to do the job, write DSP code (I don't do it all, but I learned enough to do timing recovery and most time domain tasks), program FPGA's, and be able to hold an intelligent conversation with your LSI guys as they are developing your next-gen chip. And, quite frankly, a college degree just doesn't assist this skillset all that much.
Now, granted, I work with a bunch of PhD's who are working on fabulous new echo-cancellers and the like, but for the most part, they are the research types. The tech that they work on is definitely worth it, but they seem to (in general, I've met one exception) be unable to actually put their fabulous theories into practice with the exceptionally tight memory and cycle constraints of our (or really, anybody's) DSP.
On the otherhand, if you don't have a degree, you have to get official recognition in another fashion. In my case, I've got two patents (pending, but most likely will go through). One on ATM compression, another on a software "phase-lock" (not really, but it acts like one, with the theoretical minimum buffer) of USB frames and DMT symbol rate. The places I've gotten cold-calls from have explicitly mentioned these as the reason they are calling me.
Another thing that helps (alot) is, as mentioned in other posts, is knwoing people who work at other places. That's how I initially got my job where I am. It kinda feels like cheating, but remember that they probably wouldn't want you if they didn't think you were absolutely fantabulous: bringing in mediocre candidates is a big no-no, and would refelct poorly on the guy who brought you in. The vast majority of emplyee recommendations that we get are gods. People don't like to show off their mediocre friends:)
"Hmm, but now that I think of it, the raw throughput of a DS1 is 1.544megabits per second but that is 1544000 bits per second..."
Hehe. And its not even standard throughout the communications industry. Most synchronous protocols I've seen (SHDSL, DS1, T1, E1) use 1000 bits/KBit while most asynchronous protocols I've seen (ADSL, Ethernet, ATM-25), under most circumstances, use 1024 bits/Kbit, though will occassionally revert back to 1000 bits depending on implementation. Getting DMT ADSL to play nice with DS1 resulted in some really nasty, horrible things added into the framing section of the DMT standard:)
I'm guessing that this adaptation is similar to bit-swapping that is done in DMT ADSL.
Essentially (and this is pure speculation), the frequency spectrum will be divided into a number of frequency bins of a fixed width (in DMT, its 4.3125 KHz). Then, during training, a certain number of bits will be allocated to each frequency bin, depending on the Signal-to-Noise ratio of that bin. Later, if the line conditions change, the two sides will negotiate a reallocation of bits away from any frequencies that have been trashed. In DMT, these bits have to be re-allocated to another bin, but in this case, since the line conditions are likely change rather drastically when you plug in a 1500 Watt hair-dryer with an oscillating electric motor, I'm guessing that they would simply remove the bits entirely, alter the framing, and add the bits back in when conditions improve.
nvidia has stated repeatedly that they cannot OS the drivers because their drivers contain a ton of code licensed from other companies. In otherwords, they can't OS the code for the same reason MS can't put GPL code a proprietary license: it would be a copywrite violation.
I was a Physics major, and while I don't haven't dealt with a QED in a while (I've been a DSP programmer since I moved to the real world), from what I remember it was definitely possible for electrons to move freely through time.
However, this aspect of quantum physics could not be used to accomplish anything useful. Any time you actually wanted an electron to move backward through time, it would invariably move forward as an act of observation or use would change its behavior as it would have to be moving forward through time in order to have any bearing on your expirment. The same applies to the second law of thermodynamics on the quantum scale: as Feynman (I believe, could be wrong) pointed out, "Mother nature is always having an energy sale but everytime you go to the store, she's out of stock".
Think of it this way: You have a procesor in which the 'add' instruction is broken because it causes a condition in the chip in which the output will always be wrong. If you executed a 'no-op' instruction, this condition doesn't occur. However, this doesn't allow you to do additions, so the fact that the 'no-op' instruction doesn't break things isn't useful because a no-op isn't useful (in the sense of doing addition).
In the same way, the laws of physics we hold so dear may be constantly violated at the quantum level, but its basically a no-op: it doesn't accomplish anything. Anytime we actually want to interact with these paticles to do something with them, they're back to obeying the boring old laws of physics,
Tim
Re:same reason we still run gasoline engines.....
on
Clockless Chips
·
· Score: 1
The 1-bit dac is probably the coolest thin to arise from the analog LSI part of the world in a long time. I'm not an analog LSI guy, more of a DSP guy, but here's how (in a primitive sense) 1-bit DACs work and why they're so cool:
1: They are generally referred to as delta-sigma modulators. They work how they sound - the difference between two input bits (delta) is accumulated (sigma). Thus, a 1 is 1 and a 0 is -1. Sending alternating 0's and 1's is results in a '0' analog output. Modulating to digital input signal is done by biasing the input to either 0 or one to the degree that you wish to bias the current time-domain sample.
2: Why they're cool. A delta-sigma modulator has reduced board footprint, can be run faster because its serial, and is cheaper to develop since mod/demod of the input signal is absurdly easy given a large enough over-sample ratio.
It wasn't just marketing hype, it was a real development.
First-off, i'm assuming you mean after taxes. In the US, among techie types and engineers and scientists (in other words, your median out-of-college US slashdotter), I'd guess the average salary is around $60K (mostly depending on where they live, in NY or SF its higher, in Kansas its probably more like 2/3 that), or about $3150/month after taxes.
Secondly, I want to move to Spain:) I spent $36K on my BMW, and its a low-end model (323i). Apparently we're getting ripped off cuz I make way more than the your bosses:)
According to your sources, 6-8 civillians were killed (revised down from 20).
When you detonate 300-400 tons of explosives, and only kill 6-8 people that you didn't want to kill, I'd say that's doing a pretty freaking good job. Would you prefer attacks like the Allied offensives from WW2 where 5000-10000 tons of explosives may have been deployed to destroy a single factory?
1) The purpose of the airstrikes weren't intended to conquer the country by themselves. They were meant to disable airfields, destroy radar and SAM installations, and eliminate command and control centers. The strikes has done this, and the pentagon has more or less admitted that it is out of targets.
2) First of all, Tomahawk missiles have been very accurate (80% or better) in every situation they've been employed.
Secondly, smart bombs are a heck of a lot smarter than they were 10 years ago. Or do you think the US just sat on its hands for ten years while the world moved from 33MHz 386 processors to 1.2 GHz Athlons?
Thirdly, even a smart bomb that hits 50% of the time is a damn siht better than an iron dumb bomb. At 50%, it only takes 6 bombs to get a 99% chance of hitting with one of them. Would you prefer US forces deploy 13,000 tons of explosives to destroy a single target (like Ploesti)?
3) American GI's may not be needed. The Northern Alliance seems more than willin to fulfill that role for us. With US air support bombing the shit out of everything in front of them, its unlikely they will face more than token resistance.
There could be a couple of very good reasons that the US is not so forthcoming with the evidence:
1) The West has mole(s) within these terrorist organizations and we do not want to tip our hand as to who they are, resulting in either the scuttling of these operations or the deaths of those involved.
2) The West has found and deciphered the primary form of communication that the terrorists use and we'd like them to keep using it because we know about. Indicating that we've cracked their comms would result in them changing their comms system.
The US forces in Vietnam slaughtered the VC and the NVA. Ranger and SEAL units were damn near invincible, killing any enemy they came across. Everytime the NVA grouped up to assault the South, they were promptly carpet-bombed into oblivion.
After operation Phoenix, the joke was that Presidant Thieu was more free to move in his country than President Nixon was to move in his. When the US finally left Vietnam, the South Vietnamese were holding their own against the brutally weakend Northern forces. It was until a massive morale breakdown caused the South's armies to retreat without reason (much like the French in WW2, there was no reason for them to retreat other than systemic morale problems) that the North finally took Saigon.
Slashdot Mirror
Intel's going to win anyway for the forseeable future, as AMD doesn't have anywhere near the fab capacity to meet the needs of Intel's market.
AMD would either have to outsource, which sucks because the process (but necessarily process size) *will* be different if they have TSMC or some other fab-for-hire cranking chips for them, or they need to blow a heck of a lot of cash building their own fabs, which is pretty risky since it takes a long time to build a fab and unused fab space is a huge money sink.
Tim
Blashphemy!
Everybody knows that communications engineers do it the right way, mixing big-endian and little-endian seemingly at random in order take best advantage of both systems.
After all, it makes perfect sense to byte-cross your data, send it to the ATM scrambler, bit-cross the result, feed it to your physical layer scrambler, and then bit-cross that result in order to calculate your CRC.
I now expect you to hang your head in shame for questioning the superior wisdom of communication engineers.
Oh, and crack your eggs in the middle too:)
Tim
Blacklight Power may or may not be a hoax. I tend to think they may be onto something, as they have published some peer-reviewed papers (see their web-site, lots of stuff, some reviewed), but then again, it may not pan out or be completely make-believe. Either way, their tech doesn't violate thermodynamics in any fundamental way.
Essentially, with their tech, you'd take your fuel (in this case, water), and input some energy to start their reaction. Then, using thethe excess energy from the reaction would be used to start further reactions in the fuel supply, while still leaving some energy left over for useful work. This isn't perpetual motion, though, since, when you're out of fuel, its game over. They are very upfront about the fact that it would take more energy to to make more water (fuel) than what was generated by the process.
The novel compounds appear to be a useful byproduct, but cannot be re-used as fuel.
Tim
Hardware design skills are good for a programmer to have, especially if this guy wants to work on embedded stuff as a career.
I program a custom DSP, and let me tell you, the most useful classes I took in college (not GT)were my digital design classes. Programming comes pretty natural to me (as it does, I'm sure, for a lot of CS majors), but taking those classes widened my skillset quite a bit.
If I'm getting weird ass behavior out of my programs, I can grab the chip design and figure out WTF is going on (not HDL of any sort, our chip is schematically laid out in an ancient tool). If I need a new instruction in next chip, I have some idea of what is and isn't possible and can suggest ways of making it work. If I need to change the behavior on my dev board, I can grab the VHDL for the FPGA and change it. Very, very useful, allows me to get my programming done rather than waiting for other, extremely busy people to help me out.
So really, there's no reason for him to convert to the darkside. A CS with a decent digital hardware background is a good fit in the embedded world.
Tim
You do realize that British Telecom patented hyperlinks?
And now is the time on Sprockets when we dance!
Heh. Done that before.
:) After recovering, I took a look at the chip and the packaging was completely gutted in the middle where (presumably) the (ex-)die was.
:)
I had to remove a flash chip from a socket, but couldn't find the right tool, so I grabbed an 8 inch soldering pick (the kind that look like dentists tools) and proceeded to lift the flash. Unbeknownst to me, I bridged the tool to the unrectified (110V AC) power supply exposed at the switch of my dev board.
Zap! Flash! Bang!
Tossed my sorry ass about three feet backwards while the flash chip burst into flames and jumped about a foot into the air
Oops
Tim
uh, dsl goes to about 18Kft at 384K. 10 Mb/s, while not standardized really, is easily done to about 7 kft.
Tim
Yup, that's my worry too.
Unifying the currency before unifying the individual economies and the government debt loads seems like a bad idea. It may not matter so much for the bigger/more-stable players, but for Italy or Spain I can forsee problems as they will want to devalue to ease debt burdens and increase competitiveness while the more stable economies will block any effort at devaluation in order to avoid inflation.
Despite the cries of Europeans to the contrary, this *is* going to create a winner-take-all situation, with Germany (likely) being the winner. There is an inequality between the economies and being locked into the same currency as Germany and France will cause rampant deflation in the less developed economies as they struggle to compete, much like how Argentina screwed themselves by locking their currency to the US dollar and keeping it that way (and refusing to devalue despite the urgings, of, oh, everybody else).
As someone else mentioned above, there is also the problem of job competition. WHile in theory, anybody from any Euro country can move to any other Euro country, in practice its not going to happen that way. It doesn't even happen that way in the US, where everybody speaks the same language. Try telling one of the laid off Lucent engineers living in the posh suburb of Holmdel NJ that you've got the perfect job for them in some backwater town in Mississippi (or vice-versa). The response, I assure you, will be negative for a number of reasons, namely culture difference and the fact that they will have to take a salary cut. And while the lower salary may represent greater purchasing power in Mississippi than their high one did in the shadow of NYC, not everything gets locally adjusted (cars, for instance, I grew up in rural PA and while $40K/yr there was theoretically equivalent to $60K/yr here in NJ, the lower salary doesn't buy a Mercedes, though it might get a bigger house than the $60K salary in NJ) and people, in general don't like pay cuts.
Now take that and apply it to even greater economic and cultural inequality in Europe, and I can see some real problems ahead.
Tim
Whoa dude, my regrets.
:)
I went through the same thing a year earlier and in a better economy. I decide to drop out of college, my fiance (at the time) calls me all sorts of nasty words for loser, we break up, I get an incredible job and drive over to her apartment in my new Beemer
Anyway, in this economy, the likelihood of that last part is somewhat in doubt, but just keep pushing and trying, and you'll make it. I had to pretty much bully my way into interviews, puching an 80x86 assembly version of space invaders over my resume, but I eventually got hired by a great, profitable company. And once I was there, I managed to prove myself by being the the most hardworking mofo around.
Now, I get to settle down a bit (though the stress does occassionally build up, like now), and no one really cares whether I've got a Masters in EE studying signal processing, or that I can do the work just as well as someone who does have the paper.
Tim
First of all lots of real americans do benefit from defense spending. I grew up in a blue-collar, rural area and this area, and many, many more just like it were very dependent on defense spending. Of the 10000 people in my hometown, around 1500 worked at the local foundries, forges, and steel mills. They had good paying union jobs and unemployment hovered around 0 in the area.
However, these companies did 60-90% of their business for the defense industry (US firms just aren't competitive for commercial products, but are necessary for defense products). When the defense build-up of the 80's ended, the area went into a severe depression. *All* of the foundries, forges, mills, and machine shops went bankrupt. Everybody, from the owners to the guys grinding castings, was out of a job. By 1994, unemployment was around 25% and the population had been reduced to less than 4000.
And people wonder why rural areas vote republican? I consider myself and my family pretty average americans that benefitted immensely from military spending. Later, while the coasts were enjoying the boom of the late 90's, rural areas were struggling to keep their unemployment under 15%.
Secondly, this isn't just make-work. There are tangible benfits from having a strong military and the industrial capacity to bolster it in case of, you know, an actual major war. With the US foundry and forge industries completely decimated, it'll be extremely tough if there actually is a need to crank out new planes on a daily basis to replace ones getting shot down.
Also, the reason the rest of the US friendly world has the luxury of not spending money on their militaries is that the US does it for them. Do you really think the Soviets would have felt constrained from taking the rest of Europe if the NATO didn't have massive amounts of US armor to make it a risky propostition? Do you think that the French could hold stop anyone given the current state of their military?
Tim
No no no.
:)
According to Goedel's Incompleteness Theorem, there are two equally valid mathematical systems, one that is consistent but incomplete (2 equations with 3 variables cannot be solved), and one that is complete but inconsistent (the above system can be solved, but it means that essentially true==false under certain conditions).
Men use the consistent system, women use the complete system. When a woman asks "Honey, does my ass look big in these?", according to her complete system there is a valid answer, despite the fact that it may not be consistent with other questions (ex. "So, how much do you think I weigh?") . Men, however are confined to a consistent system and unable to answer these questions correctly, and are left sitting there unable to speak for fear of answering wrong.
I blame Goedel
Tim
Dude, this guy isn't just doing processing with an asynchronous chip, he's doing freaking tone detection with it.
The actual logic could be asynchronous for all I care, but as a DSP programmer I will confidently state that doing tone detection without at least one known, fixed clock in the circuit is just plain deranged.
Tim
I'm in the same boat: An embedded programmer without a college degree, and I get cold calls from my company's competitors all the time.
:)
Somehow, I think the embedded world may be different because it takes a unique skillset to be able to work on embedded processors with barely enough enough memory to do the job, write DSP code (I don't do it all, but I learned enough to do timing recovery and most time domain tasks), program FPGA's, and be able to hold an intelligent conversation with your LSI guys as they are developing your next-gen chip. And, quite frankly, a college degree just doesn't assist this skillset all that much.
Now, granted, I work with a bunch of PhD's who are working on fabulous new echo-cancellers and the like, but for the most part, they are the research types. The tech that they work on is definitely worth it, but they seem to (in general, I've met one exception) be unable to actually put their fabulous theories into practice with the exceptionally tight memory and cycle constraints of our (or really, anybody's) DSP.
On the otherhand, if you don't have a degree, you have to get official recognition in another fashion. In my case, I've got two patents (pending, but most likely will go through). One on ATM compression, another on a software "phase-lock" (not really, but it acts like one, with the theoretical minimum buffer) of USB frames and DMT symbol rate. The places I've gotten cold-calls from have explicitly mentioned these as the reason they are calling me.
Another thing that helps (alot) is, as mentioned in other posts, is knwoing people who work at other places. That's how I initially got my job where I am. It kinda feels like cheating, but remember that they probably wouldn't want you if they didn't think you were absolutely fantabulous: bringing in mediocre candidates is a big no-no, and would refelct poorly on the guy who brought you in. The vast majority of emplyee recommendations that we get are gods. People don't like to show off their mediocre friends
Tim
"Hmm, but now that I think of it, the raw throughput of a DS1 is 1.544megabits per second but that is 1544000 bits per second..."
:)
Hehe. And its not even standard throughout the communications industry. Most synchronous protocols I've seen (SHDSL, DS1, T1, E1) use 1000 bits/KBit while most asynchronous protocols I've seen (ADSL, Ethernet, ATM-25), under most circumstances, use 1024 bits/Kbit, though will occassionally revert back to 1000 bits depending on implementation. Getting DMT ADSL to play nice with DS1 resulted in some really nasty, horrible things added into the framing section of the DMT standard
Tim
I'm guessing that this adaptation is similar to bit-swapping that is done in DMT ADSL.
:)
Essentially (and this is pure speculation), the frequency spectrum will be divided into a number of frequency bins of a fixed width (in DMT, its 4.3125 KHz). Then, during training, a certain number of bits will be allocated to each frequency bin, depending on the Signal-to-Noise ratio of that bin. Later, if the line conditions change, the two sides will negotiate a reallocation of bits away from any frequencies that have been trashed. In DMT, these bits have to be re-allocated to another bin, but in this case, since the line conditions are likely change rather drastically when you plug in a 1500 Watt hair-dryer with an oscillating electric motor, I'm guessing that they would simply remove the bits entirely, alter the framing, and add the bits back in when conditions improve.
But that's just a guess
Tim
nvidia has stated repeatedly that they cannot OS the drivers because their drivers contain a ton of code licensed from other companies. In otherwords, they can't OS the code for the same reason MS can't put GPL code a proprietary license: it would be a copywrite violation.
I was a Physics major, and while I don't haven't dealt with a QED in a while (I've been a DSP programmer since I moved to the real world), from what I remember it was definitely possible for electrons to move freely through time.
However, this aspect of quantum physics could not be used to accomplish anything useful. Any time you actually wanted an electron to move backward through time, it would invariably move forward as an act of observation or use would change its behavior as it would have to be moving forward through time in order to have any bearing on your expirment. The same applies to the second law of thermodynamics on the quantum scale: as Feynman (I believe, could be wrong) pointed out, "Mother nature is always having an energy sale but everytime you go to the store, she's out of stock".
Think of it this way: You have a procesor in which the 'add' instruction is broken because it causes a condition in the chip in which the output will always be wrong. If you executed a 'no-op' instruction, this condition doesn't occur. However, this doesn't allow you to do additions, so the fact that the 'no-op' instruction doesn't break things isn't useful because a no-op isn't useful (in the sense of doing addition).
In the same way, the laws of physics we hold so dear may be constantly violated at the quantum level, but its basically a no-op: it doesn't accomplish anything. Anytime we actually want to interact with these paticles to do something with them, they're back to obeying the boring old laws of physics,
Tim
The 1-bit dac is probably the coolest thin to arise from the analog LSI part of the world in a long time. I'm not an analog LSI guy, more of a DSP guy, but here's how (in a primitive sense) 1-bit DACs work and why they're so cool:
1: They are generally referred to as delta-sigma modulators. They work how they sound - the difference between two input bits (delta) is accumulated (sigma). Thus, a 1 is 1 and a 0 is -1. Sending alternating 0's and 1's is results in a '0' analog output. Modulating to digital input signal is done by biasing the input to either 0 or one to the degree that you wish to bias the current time-domain sample.
2: Why they're cool. A delta-sigma modulator has reduced board footprint, can be run faster because its serial, and is cheaper to develop since mod/demod of the input signal is absurdly easy given a large enough over-sample ratio.
It wasn't just marketing hype, it was a real development.
Tim
First-off, i'm assuming you mean after taxes. In the US, among techie types and engineers and scientists (in other words, your median out-of-college US slashdotter), I'd guess the average salary is around $60K (mostly depending on where they live, in NY or SF its higher, in Kansas its probably more like 2/3 that), or about $3150/month after taxes.
:) I spent $36K on my BMW, and its a low-end model (323i). Apparently we're getting ripped off cuz I make way more than the your bosses :)
Secondly, I want to move to Spain
Tim
Even a single electron will act like a wave in the dual-slit experiment. Individual or not has nothing to do with it.
For a good explanation of this phenomena, I sugest feynman's QED.
Tim
According to your sources, 6-8 civillians were killed (revised down from 20).
When you detonate 300-400 tons of explosives, and only kill 6-8 people that you didn't want to kill, I'd say that's doing a pretty freaking good job. Would you prefer attacks like the Allied offensives from WW2 where 5000-10000 tons of explosives may have been deployed to destroy a single factory?
Tim
1) The purpose of the airstrikes weren't intended to conquer the country by themselves. They were meant to disable airfields, destroy radar and SAM installations, and eliminate command and control centers. The strikes has done this, and the pentagon has more or less admitted that it is out of targets.
2) First of all, Tomahawk missiles have been very accurate (80% or better) in every situation they've been employed.
Secondly, smart bombs are a heck of a lot smarter than they were 10 years ago. Or do you think the US just sat on its hands for ten years while the world moved from 33MHz 386 processors to 1.2 GHz Athlons?
Thirdly, even a smart bomb that hits 50% of the time is a damn siht better than an iron dumb bomb. At 50%, it only takes 6 bombs to get a 99% chance of hitting with one of them. Would you prefer US forces deploy 13,000 tons of explosives to destroy a single target (like Ploesti)?
3) American GI's may not be needed. The Northern Alliance seems more than willin to fulfill that role for us. With US air support bombing the shit out of everything in front of them, its unlikely they will face more than token resistance.
Tim
There could be a couple of very good reasons that the US is not so forthcoming with the evidence:
1) The West has mole(s) within these terrorist organizations and we do not want to tip our hand as to who they are, resulting in either the scuttling of these operations or the deaths of those involved.
2) The West has found and deciphered the primary form of communication that the terrorists use and we'd like them to keep using it because we know about. Indicating that we've cracked their comms would result in them changing their comms system.
Tim
Huh?
The US forces in Vietnam slaughtered the VC and the NVA. Ranger and SEAL units were damn near invincible, killing any enemy they came across. Everytime the NVA grouped up to assault the South, they were promptly carpet-bombed into oblivion.
After operation Phoenix, the joke was that Presidant Thieu was more free to move in his country than President Nixon was to move in his. When the US finally left Vietnam, the South Vietnamese were holding their own against the brutally weakend Northern forces. It was until a massive morale breakdown caused the South's armies to retreat without reason (much like the French in WW2, there was no reason for them to retreat other than systemic morale problems) that the North finally took Saigon.
Tim
I believe that treason was removed from the list of capital crimes in the US.
However, durin wartime when lives are at stake, the rules might change. Anyone have more info on this?
Tim