Nope. The 'scam' in the flash boys (from the interviews - as per usual Slashdot expectations I haven't read the book) is that someone places a very large stock order for X at the current ask price that is routed to multiple markets. Let's call them A, B and C. From your trader the delays to those markets are 10ms, 100ms and 200ms respectively (which are ridiculously high numbers for this game). Your HFT trader has collocated servers at markets A,B,C, and minimal-latency links between their servers. So when the order arrives on market A and fills, they think 'Hmm, someone is looking for a shedload of X. They then place instructions to buy on the other markets, followed by orders to sell at a slightly increased price. They have 90ms (- the time for exchange A to match, fill, post market data etc., and the time for orders to be placed on other exchanges). It's like some slow moving person walking from stall to stall in a (physical) market buying all the oranges, and announcing loudly that they are doing so. Is it illegal to run to the next stalls, buying all the oranges and then offering them back to the slow moving guy?
All the information is public. The market data feeds are available to anyone. You pay for more up to date market data (which includes details of fills, not orders placed) - you don't pay for the 20-minute delayed stuff on google/yahoo, but you do if you want it faster. You pay for collocation. You pay for those low-latency connections. You used to pay for a trading desk on the stock exchange floor, guys in coloured blazers who could calculate and make decisions faster. The system has never been 'fair', but HFT doesn't necessarily make it worse.
And your 401k is managed by people so naive as to allow that? They don't themselves adopt similar technologies and strategies to mitigate that? I'd move my investments if I were you.
If you are a super-duper-uber driver. Otherwise it's better at keeping the wheels turning at just below the point of locking up, which is where the maximum traction is. Hence why they used to teach cadence braking. Once you've locked the wheels the coefficient of friction is lower - and therefore braking distance longer - than when the wheels are just about to lock up. The big advantage is that with ABS every driver can mash pedal and stop in very nearly the same optimal distance as the 'best' drivers, and still be able to steer as well.
Happened when I was at primary school. Something shorted out and started a fire and burnt several cars.
IIRC E36 BMW coupes caught fire occasionally because the cabling into the boot(trunk) lid would get brittle over time and split when it was flexed. In mine it manifested itself as the central locking failing because some wires shorted out. I'd imagine there's a sensor for the alarm, or supply for the central locking that's live even with the ignition off, so it's not a big leap of faith to see that a parked car could catch fire due to something shorting out.
Cars have quite a lot of 'live' when off electrical equipment - cooling fans for example - that can be on at any time, so faulty or ill designed wiring could cause problems in stationary cars.
Yeah... try talking to someone who worked in a camera shop sometime. A guy I knew had some regular customers building some pretty interesting photo collections.
I thought the WHO did a study that discovered the effects of mass evacuations were far more damaging than simply staying in the affected areas were - that the stress caused by panic and hysteria over doses of radiation that aren't particularly high is more damaging than the radiation could be. This is the first google result I could find but I'm sure there was a proper report from the WHO.
I also notice that the rhetoric has changed from 'all radiation is deadly evil' to 'ok it's probably not that bad... but you might eat some!!! Horrible death!!!!'. Pro-nuclear types always get moaned at for changing their story, but I notice the anti-nuclear brigade have changed their concerns as well. Is that because of the complete lack of even illness, let alone deaths, due to radiation from Fukushima, including amongst those who are cleaning up the mess? Is it because when people looked at it they realised the linear-no-threshold approach to determining the risk of radiation is pretty weak scientifically (disproven in some studies that found low levels of radiation are beneficial and even necessary)? Is it because a lot of the world is more radioactive anyway and people still live there without coming to a significantly greater level of harm? Is it because although there is a - greater than any nuclear disaster in some cases - risk of floods, volcanoes, hurricanes, typhoons, earthquakes and - not wanting to be insensitive - tsunamis people still seem to live pretty happy lives in those areas?
Perhaps because a couple of years ago the world was ending, Tokyo was going to be evacuated, the US was going to be irradiated as it all drifted across the ocean, dogs and cats would be living together etc. but actually what we have is a lot of water (a lot of which could have just been dispersed in the ocean), a huge decontamination operation (which probably could have been targeted at areas with actual significant radiation rather than just removing inches of topsoil from entire prefectures), no deaths (other than those caused by the evacuation - oh and the 10's of thousands caused by the Tsunami itself). It must be a big disappointment to the serious doomsday-scenario junkies that none of their predictions have come true so far. It is a disaster, but those happen pretty regularly. More people died from car bombs in the middle east last week than due to Fukushima. 300,000 people were evacuated - Syria has over a million registered refugees already and a death toll of 115,000. The Rwandan genocide has a death toll estimated at 500,000-1,000,000. Fukushima was a big accident, will cost a fortune to clean up, but is nowhere near the scale of disaster it's made out to be.
Oh, and they get between 5 and 9.5 weeks holiday, lot's of employment rights and protections and tasty cheese. The last time I saw an American commenting on France's productivity and employment laws it was the head of a tyre company - I think the French pointed out that Michelin is 20 times larger and 35 times more profitable than the US company. Also if you think the French are more concerned about money than quality of life then you have no idea what they are about. At least remember to thank them for scaring the British out...
Might be why, despite 30% of the population smoking (2005), they live 3 years longer than people in the US (20% smokers in 2006). It's not all about money...
That's not really how it works... Propagation delay is not related to clocks (at least not in the way you seem to imply). With a stable and monotonous clock then you can easily propagate the clock to every point of the chip with a controllable delay (see comments from an actual designer of this stuff above). A simple clock tree, for example, can be implemented using a fractal like structure. Basically imagine a capital H, you have an equal length to each of the 'ends' of the lines from the centre. To each end you attach another 'H', half the size, with middle of the bar on the end of the previous larger H.
Your comments are referring more to the actual time between clocks that the chip has to perform some function. This is the propagation delay of actual logic. Again, the simple way to look at it is that each logic gate between the flops takes a fixed amount of time to respond to an input change, thus there is a limit to the number of logic gates in a chain that can sit between registers. There are other factors - e.g. setup and hold times for flops, clock skew (due to the distance between the register and an end point of the clock tree), fanout - but that's the basics of it. By adding pipeline stages you reduce the amount of logic gates that need to update between clock cycles, and thus you can run them faster. The downside is using more registers, more complexity, and in the case of processors particularly, the need to stall or flush pipelines, or predict behaviour of branches etc. so that the pipeline can be kept full.
"All I know is that I will NEVER trust people to run fission power stations as people cut corners and lie. They do so when government owned, and they do so when owned by a company."
You'd best get back in your cave then... Seeing as your more than likely trust people every day to provide food, clean water, medicines, transport, power and many other essential necessities. In every country there are people with the power to do you huge amounts of harm in ways that are far more subtle than by running a Nuclear power plant, and yet every day they tend not to. You trust a whole chain of people, from waiters not spitting in your soup to sheikhs not shutting of your oil supplies. You trust drivers not to run you down and you trust pilots not to mistake your house for a runway. Every so often that trust is breached, I agree, and that is a bad thing, but unless you live in complete isolation (which you don't - I've seen you posting on the internet!) you are ignoring all the other ways people could harm you every day and worrying about one that, statistically at least, is very unlikely to do you any harm at all.
I'm glad someone else pointed this out. The OP's complaint makes it sound like we're going to run out of silicon if we use too much. Also the 'juggling between cores not being ARM's problem' bit doesn't make a lot of sense: ARM have an awful lot of interest in producing something that provides a real-world performance/power envelope that is attractive to their customers. If, due to the complexity of operating the chip or some other factors, this is not practical or possible, the chip won't sell. It's suggesting that the likes of Samsung, Apple, HTC, Nokia, LG, Sony and many others just pick the chip because it had good marketing rather than a detailed analysis of the performance against the expected uses of the device.
So basically a KERS type system? e.g. a small ICE for range and a much more powerful electric/flywheel motor for acceleration? Depending on usage that may make more sense than the current system. An ICE that can sustain say 90mph whilst still providing some left over energy to be stored for acceleration. Worst case is you do lots of high speed starts and stops (e.g. driving like a tool in traffic or urban areas) which leaves you with only the ICE power - which would likely still be enough for most purposes.
Yeah, that stuck in my craw as well. Obviously everyone is supposed to be all "Wooo 6 million transactions a day... they must be teh über-programmers! That system could never go wrong".
On second thoughts I should apply for a job. On a modern CPU I'd expect to be able to do 6 million things and find time to get a few of them wrong in less than a second! Must be able to get a cushy contract with performance figures like that.
Good luck with that. I had to phone my car insurance company three times, the first two times I was promised a manager would call me back immediately. My tone went from polite but frustrated to full on shouting down the phone and threatening to come round and burn the place down angry at which point I was magically transferred to a manager. The first thing I said to the manager was that he was an arsehole for making it so difficult to reach him that I had to deliver such a torrent of abuse to a member of his staff in order for them to be allowed to put me through to him directly. I've also told several people I've spoken to at Spark Energy that they should look for somewhere else to work because having seen the level of incompetence they show to their customers I wouldn't want to be dependent on them for my pay.
And yet the French seem to do fine with some pretty rare burgers. Ideally you'd cook steak rare to kill off any surface bacteria, then grind it with clean equipment, form it into patties and cook again. I doubt that's what actually happens though.
Also depending on how your hot water system operates it may have been sitting in a tank for a few days. If you've got an old indirect gravity fed system then there's probably a header tank in the loft that is possibly open topped and can have all sorts of gunk floating around in it.
If the head's been ground against the platter? They have shiny bits of glass and a lot of ferrous dust.
Surely there are already secured drives with an internal encryption key in some kind of fuse setup, whereby all data is encrypted with this key and on access some form of authentication occurs to release the key. Get it wrong a few times and the drive electronics blow the fuses erasing the key and preventing any access to the data ever. For the truly paranoid skip the authentication and have it coded to some undocumented command or specific pattern of initial access. If the pattern is unexpected, blow the fuses.
Because if it's actually changed the atoms making up whatever fuel is involved, it's a nuclear process. A chemical reaction will leave you with the same atoms you started with, they'll just be hanging around in different groups. In a nuclear reaction the energy output comes directly from mass e.g. some parts of those atoms are converted to energy. So the simplest answer is to measure the mass of the fuel+waste products before and after the reaction.
BSE being a disease effectively created by feeding cows on other cows. One that seems to be transferrable to humans. It's unfortunate but it's not really a counter-argument in this context. It's just an argument that you shouldn't eat beef as well as humans.
Slashdot Mirror
Nope. The 'scam' in the flash boys (from the interviews - as per usual Slashdot expectations I haven't read the book) is that someone places a very large stock order for X at the current ask price that is routed to multiple markets. Let's call them A, B and C. From your trader the delays to those markets are 10ms, 100ms and 200ms respectively (which are ridiculously high numbers for this game). Your HFT trader has collocated servers at markets A,B,C, and minimal-latency links between their servers. So when the order arrives on market A and fills, they think 'Hmm, someone is looking for a shedload of X. They then place instructions to buy on the other markets, followed by orders to sell at a slightly increased price. They have 90ms (- the time for exchange A to match, fill, post market data etc., and the time for orders to be placed on other exchanges). It's like some slow moving person walking from stall to stall in a (physical) market buying all the oranges, and announcing loudly that they are doing so. Is it illegal to run to the next stalls, buying all the oranges and then offering them back to the slow moving guy?
All the information is public. The market data feeds are available to anyone. You pay for more up to date market data (which includes details of fills, not orders placed) - you don't pay for the 20-minute delayed stuff on google/yahoo, but you do if you want it faster. You pay for collocation. You pay for those low-latency connections. You used to pay for a trading desk on the stock exchange floor, guys in coloured blazers who could calculate and make decisions faster. The system has never been 'fair', but HFT doesn't necessarily make it worse.
And your 401k is managed by people so naive as to allow that? They don't themselves adopt similar technologies and strategies to mitigate that? I'd move my investments if I were you.
I thought all HID systems (at least in the EU) were required to be self levelling...
If you are a super-duper-uber driver. Otherwise it's better at keeping the wheels turning at just below the point of locking up, which is where the maximum traction is. Hence why they used to teach cadence braking. Once you've locked the wheels the coefficient of friction is lower - and therefore braking distance longer - than when the wheels are just about to lock up. The big advantage is that with ABS every driver can mash pedal and stop in very nearly the same optimal distance as the 'best' drivers, and still be able to steer as well.
Happened when I was at primary school. Something shorted out and started a fire and burnt several cars.
IIRC E36 BMW coupes caught fire occasionally because the cabling into the boot(trunk) lid would get brittle over time and split when it was flexed. In mine it manifested itself as the central locking failing because some wires shorted out. I'd imagine there's a sensor for the alarm, or supply for the central locking that's live even with the ignition off, so it's not a big leap of faith to see that a parked car could catch fire due to something shorting out.
Cars have quite a lot of 'live' when off electrical equipment - cooling fans for example - that can be on at any time, so faulty or ill designed wiring could cause problems in stationary cars.
The conflict between your comment and your signature is making my brain hurt... is it intentional?
Yeah... try talking to someone who worked in a camera shop sometime. A guy I knew had some regular customers building some pretty interesting photo collections.
I thought the WHO did a study that discovered the effects of mass evacuations were far more damaging than simply staying in the affected areas were - that the stress caused by panic and hysteria over doses of radiation that aren't particularly high is more damaging than the radiation could be. This is the first google result I could find but I'm sure there was a proper report from the WHO.
I also notice that the rhetoric has changed from 'all radiation is deadly evil' to 'ok it's probably not that bad... but you might eat some!!! Horrible death!!!!'. Pro-nuclear types always get moaned at for changing their story, but I notice the anti-nuclear brigade have changed their concerns as well. Is that because of the complete lack of even illness, let alone deaths, due to radiation from Fukushima, including amongst those who are cleaning up the mess? Is it because when people looked at it they realised the linear-no-threshold approach to determining the risk of radiation is pretty weak scientifically (disproven in some studies that found low levels of radiation are beneficial and even necessary)? Is it because a lot of the world is more radioactive anyway and people still live there without coming to a significantly greater level of harm? Is it because although there is a - greater than any nuclear disaster in some cases - risk of floods, volcanoes, hurricanes, typhoons, earthquakes and - not wanting to be insensitive - tsunamis people still seem to live pretty happy lives in those areas?
Perhaps because a couple of years ago the world was ending, Tokyo was going to be evacuated, the US was going to be irradiated as it all drifted across the ocean, dogs and cats would be living together etc. but actually what we have is a lot of water (a lot of which could have just been dispersed in the ocean), a huge decontamination operation (which probably could have been targeted at areas with actual significant radiation rather than just removing inches of topsoil from entire prefectures), no deaths (other than those caused by the evacuation - oh and the 10's of thousands caused by the Tsunami itself). It must be a big disappointment to the serious doomsday-scenario junkies that none of their predictions have come true so far. It is a disaster, but those happen pretty regularly. More people died from car bombs in the middle east last week than due to Fukushima. 300,000 people were evacuated - Syria has over a million registered refugees already and a death toll of 115,000. The Rwandan genocide has a death toll estimated at 500,000-1,000,000. Fukushima was a big accident, will cost a fortune to clean up, but is nowhere near the scale of disaster it's made out to be.
Oh, and they get between 5 and 9.5 weeks holiday, lot's of employment rights and protections and tasty cheese. The last time I saw an American commenting on France's productivity and employment laws it was the head of a tyre company - I think the French pointed out that Michelin is 20 times larger and 35 times more profitable than the US company. Also if you think the French are more concerned about money than quality of life then you have no idea what they are about. At least remember to thank them for scaring the British out...
Might be why, despite 30% of the population smoking (2005), they live 3 years longer than people in the US (20% smokers in 2006). It's not all about money...
un-useful != incorrectly-usable
That's not really how it works... Propagation delay is not related to clocks (at least not in the way you seem to imply). With a stable and monotonous clock then you can easily propagate the clock to every point of the chip with a controllable delay (see comments from an actual designer of this stuff above). A simple clock tree, for example, can be implemented using a fractal like structure. Basically imagine a capital H, you have an equal length to each of the 'ends' of the lines from the centre. To each end you attach another 'H', half the size, with middle of the bar on the end of the previous larger H.
Your comments are referring more to the actual time between clocks that the chip has to perform some function. This is the propagation delay of actual logic. Again, the simple way to look at it is that each logic gate between the flops takes a fixed amount of time to respond to an input change, thus there is a limit to the number of logic gates in a chain that can sit between registers. There are other factors - e.g. setup and hold times for flops, clock skew (due to the distance between the register and an end point of the clock tree), fanout - but that's the basics of it. By adding pipeline stages you reduce the amount of logic gates that need to update between clock cycles, and thus you can run them faster. The downside is using more registers, more complexity, and in the case of processors particularly, the need to stall or flush pipelines, or predict behaviour of branches etc. so that the pipeline can be kept full.
"All I know is that I will NEVER trust people to run fission power stations as people cut corners and lie. They do so when government owned, and they do so when owned by a company."
You'd best get back in your cave then... Seeing as your more than likely trust people every day to provide food, clean water, medicines, transport, power and many other essential necessities. In every country there are people with the power to do you huge amounts of harm in ways that are far more subtle than by running a Nuclear power plant, and yet every day they tend not to. You trust a whole chain of people, from waiters not spitting in your soup to sheikhs not shutting of your oil supplies. You trust drivers not to run you down and you trust pilots not to mistake your house for a runway. Every so often that trust is breached, I agree, and that is a bad thing, but unless you live in complete isolation (which you don't - I've seen you posting on the internet!) you are ignoring all the other ways people could harm you every day and worrying about one that, statistically at least, is very unlikely to do you any harm at all.
Is there?
I'm glad someone else pointed this out. The OP's complaint makes it sound like we're going to run out of silicon if we use too much. Also the 'juggling between cores not being ARM's problem' bit doesn't make a lot of sense: ARM have an awful lot of interest in producing something that provides a real-world performance/power envelope that is attractive to their customers. If, due to the complexity of operating the chip or some other factors, this is not practical or possible, the chip won't sell. It's suggesting that the likes of Samsung, Apple, HTC, Nokia, LG, Sony and many others just pick the chip because it had good marketing rather than a detailed analysis of the performance against the expected uses of the device.
So basically a KERS type system? e.g. a small ICE for range and a much more powerful electric/flywheel motor for acceleration? Depending on usage that may make more sense than the current system. An ICE that can sustain say 90mph whilst still providing some left over energy to be stored for acceleration. Worst case is you do lots of high speed starts and stops (e.g. driving like a tool in traffic or urban areas) which leaves you with only the ICE power - which would likely still be enough for most purposes.
Yeah, that stuck in my craw as well. Obviously everyone is supposed to be all "Wooo 6 million transactions a day... they must be teh über-programmers! That system could never go wrong".
On second thoughts I should apply for a job. On a modern CPU I'd expect to be able to do 6 million things and find time to get a few of them wrong in less than a second! Must be able to get a cushy contract with performance figures like that.
In English I'd guess that would be an inventory. Which would work if they don't assume you've stolen the missing items.
Good luck with that. I had to phone my car insurance company three times, the first two times I was promised a manager would call me back immediately. My tone went from polite but frustrated to full on shouting down the phone and threatening to come round and burn the place down angry at which point I was magically transferred to a manager. The first thing I said to the manager was that he was an arsehole for making it so difficult to reach him that I had to deliver such a torrent of abuse to a member of his staff in order for them to be allowed to put me through to him directly. I've also told several people I've spoken to at Spark Energy that they should look for somewhere else to work because having seen the level of incompetence they show to their customers I wouldn't want to be dependent on them for my pay.
And yet the French seem to do fine with some pretty rare burgers. Ideally you'd cook steak rare to kill off any surface bacteria, then grind it with clean equipment, form it into patties and cook again. I doubt that's what actually happens though.
Also depending on how your hot water system operates it may have been sitting in a tank for a few days. If you've got an old indirect gravity fed system then there's probably a header tank in the loft that is possibly open topped and can have all sorts of gunk floating around in it.
If the head's been ground against the platter? They have shiny bits of glass and a lot of ferrous dust.
Surely there are already secured drives with an internal encryption key in some kind of fuse setup, whereby all data is encrypted with this key and on access some form of authentication occurs to release the key. Get it wrong a few times and the drive electronics blow the fuses erasing the key and preventing any access to the data ever. For the truly paranoid skip the authentication and have it coded to some undocumented command or specific pattern of initial access. If the pattern is unexpected, blow the fuses.
Because if it's actually changed the atoms making up whatever fuel is involved, it's a nuclear process. A chemical reaction will leave you with the same atoms you started with, they'll just be hanging around in different groups. In a nuclear reaction the energy output comes directly from mass e.g. some parts of those atoms are converted to energy. So the simplest answer is to measure the mass of the fuel+waste products before and after the reaction.
BSE being a disease effectively created by feeding cows on other cows. One that seems to be transferrable to humans. It's unfortunate but it's not really a counter-argument in this context. It's just an argument that you shouldn't eat beef as well as humans.
So you're arguing that society should provide medical care to those who can't afford it? I think I know of a way of implementing that...