'' Can one of the cryptographers on slashdot comment on weather this is useful to them or not? ''
One useful addition (copied from Altivec) is the vector permute instruction. What is clever about it in terms of cryptography is that you can translate a vector using a 256 byte translation table _without doing any memory access_ by using the vector permute instruction in a clever way. Now the execution time is completely data-independent, so one important attack vector is closed.
For example, the Advanced Encryption Standard (AES) algorithm gets a factor of 5 performance improvement by using the new SSE5 extension If I get one of these CPUs, I'll almost certainly be encrypting my hard drives. It was already fast enough, but now...
They copied two important features from the PowerPC instruction set: Fused multiply-add (calculate +/- x*y +/- z in one instruction), and the Altivec vector permute instruction, which can among other things rearrange 16 bytes in an arbitrary way. The latter should be really nice for AES, because it does a lot of rearranging 4x4 byte matrices (if I remember correctly).
I don't think having two dual-cores in a package instead of four cores combined is necessarily a disadvantage. To compare these properly, you would have to assume same quality of implementation. So Intel could have gone for one unified 12MB L2 cache with four access paths instead of two 6MB L2 caches with two access paths each. With same quality of implementation, the four access paths will be slower because you have to cope with four processors accessing it at the same time instead of two. So each access will be slower. Larger caches are always slower anyway, so 12 MB is slower than 6 MB again. On the other hand, 12 MB unified is better if you have only one thread that is cache intensive, and the unified cache is better if more than two threads are communicating, and you don't need to worry about allocating threads to the right processors in pairs. Communicating between L2 cache and main memory might be faster for 2x2 cores because you have two access paths, on the other hand making them fast is harder.
So all in all, the speed advantage could go either way, and will depend on the code that you run. For example, if you run four video encoders simultaneously, all the advantages of 1x4 cores don't help, but the higher speed of 2x2 does. Other tasks will be different.
Re:I wish mainstream CPUs / GPUs would focus on po
on
Via Unveils 1-Watt x86 CPU
·
· Score: 2, Interesting
'' 1.6L is considered big? I have a 2L 4cyl Focus in Canada, and that's considered "small" by our standards. Not that I really push my car, but I am curious as to how a 1.3L accelerates [to say hwy speeds]. Because even in my car I have to really floor it [re: 5000 RPM] to hit highway speeds before I exit the ramp, well that's exaggerating a bit. usually I hit speed before the dotted lines (that let you get out of the merging lane). So I probably could accelerate at like 3-4K RPM just fine. ''
Get a Diesel engine. Massive torque = massive acceleration. Not that much horse power, but that only matters at high speeds (100mph+) where you lose your driving license anyway.
In the UK, tax goes by carbon dioxide emission per km, engine size doesn't matter. There is a small number of cars that pay £35 per year, others pay between £115 and >£200 tax per year. But there are other differences: At the moment, you pay a £8 charge every time you drive into London. In the future, that will be free for cars with very low emissions, and up to £25 for very high emissions.
But the thing that really hits is company car tax. If you have a company car, you have to pay income tax on X percent of the value of the new car every year. X ranges from 15% to 35%, depending on carbon dioxide emissions. For a £20,000 car, you pay tax on £3000 to £7000, depending on emissions. At 40% tax rate, that is £1200 to £2800 tax, in other words up to £1600 punishment every year for high carbon dioxide emissions for a £20,000 car.
So they think refusing Firefox users has only little financial impact? What, like five percent? What kind of idiot would you have to be to give up five percent of your revenue voluntarily? Imagine being a webmaster trying to explain to the marketing guys that it is a good idea to reduce their revenues by five percent.
This reminds me of past stupidities, where web designers didn't design for Mac users, because Mac users are idiots who spend much too much money on shiny computers where a $400 Dell would do just fine. And marketing says "idiots who spend much too much money on shiny toys are _exactly_ who we want to visit our website"!
'' That's no excuse! A modern computer is one of the most complex machines known to Man, yet everyone expects to be able to use one without danger. ''
Of course we do. A car weighs usually between 1.5 and 2 tons and often moves at a speed of 70 miles per hour, that is a lot of kinetic energy. A bullet has much less weight, but travels at much higher speed, and if it impacts, all its kinetic energy goes into a small area. A modern computer however weighs maybe 5 pound if it is a laptop, hopefully not more than 30 pound otherwise, and is sitting still. Hard disks are inside a case that should protect me. I hope that the CD drive is carefully designed so that it doesn't crack CDs and shoots out parts of them. Yes, I expect to be able to use a computer without any danger.
>> "They would give software access to information about cache misses..." Yeah that ought to help significantly with side-channel attacks against crypto software.
I think you didn't read the spec. All that information is only available to the thread that is profiled; everything is context-switched so it can't leak out to other threads and definitely not to other processes.
'' While most obese individuals just need to start walking and lay off McDonald... ''
Considering how much money is made with products that are supposed to help fat people to get slimmer, and that all these products don't work, I think things might be a bit more difficult than you think.
'' An analysis of 40 studies involving 250,000 people, heart patients with normal BMIs were at higher risk of death from cardiovascular disease than people whose BMIs put them in the "overweight" range (BMI 25-29.9). Patients who were underweight or severely overweight had an increased risk of death from cardiovascular disease. The implications of this finding can be confounded by the fact that many chronic diseases, such as diabetes, cause weight loss before the eventual death. In light of this, higher death rates among thinner people would be the expected result. ''
This is about health insurance, and health insurance is about cost. Dead people don't cost much. Sick people cost lots.
'' Apple should be rejoicing, since they claim that the patent system is running perfectly. ''
Just shows that you didn't read properly. In that recent discussion, Google said that the patent system itself is broken. Apple said the patent system is fine, the related litigation system is broken. And clearly it is.
The other example mentioned was Microsoft being ordered to pay $1.5bn over two MP3 related patents. We all know that MP3 is covered by a few hundred patents, and Microsoft paid a few million for a license for all those patents, so one or two patents they missed could never be worth $1.5bn.
'' Simple solution: Have someone else, who has no ties to Lyons take over the blog. ''
Could you imagine what Apple could do by creating a "Fake Steve Jobs" blog now? Can you imagine Mr. Lyons trying to sue Apple for using the name "Steve Jobs" in a fake blog?
'' An iPhone will work, but really it could be any item that indicates to the woman that you're willing to spend hundreds of dollars on something pretty. ''
You still have to pull it off. And somehow I think that someone who posts on Slashdot that an iPhone "got him laid" is not the kind of person who _can_ pull it off.
Your car is not safer, it is heavier. In a collision, most of the deforming energy goes into the lighter vehicle. Direct consequence of conservation of impulse and conservation of energy.
I am still waiting for lawsuits where an owner of a small car sues the owner of a larger car for damages, because it is _your heavy car_ that caused the damage. If you hadn't acted selfishly, that other car would have sustained much less damage.
I don't know if you realise this, but they are not showing screenshots from the defendants computer.
The RIAA is showing screenshots from the computers of that gang of hackers that they hire to spy on their victims. So it is _totally_ under their control what is in these screenshots.
'' As others have pointed out, ignorance is generally not an excuse. ''
Stop right there. Ignorance of the law is no excuse; if you commit a crime then it doesn't matter whether you knew or not that it was illegal.
Ignorance _is_ often an excuse because it makes the difference between actually having committed a crime or not. Lets say at a car park you put your coat on the back of my car while you tie your shoelaces. I drive away in my car - with your coat on it. If I knew that your coat was there, it is theft. If I didn't know it was there, it is no theft.
In this case, ignorance that someone used your computer to infringe someone's copyright is most definitely an excuse.
I don't know, but since this is Slashdot, and Slashdot readers would buy a phone for its features, you would surely realise that at least two really important features of the iPhone (use as a video iPod, and use with WiFi) depend on a computer hanging around somewhere. So people without a computer would be very unlikely to buy one.
'' Not everyone who will need to be "in the bush" for a few days will know this much in advance. What happens if you learn you'll need that capability one year into the two-year contract? ''
Must be horrible. A week in the bush, and when you take your iPhone to the bush-battery-shop, they tell you they have no replacement batteries for the iPhone.
'' He said, "Can we have another option to fly? We'll call it Fly At Your Own Risk Airlines. We won't screen for anything and you can pay for your tickets five minutes before your flight just like in the old days-1997." ''
Can you imagine the hilarity when you find out that the other 199 passengers are carrying bombs as well?
'' "If the TSO throws your liquids in the trash, they don't find you a threat."
If they didn't find (you) a threat, then WHY THROW THE FREAKIN' LIQUIDS IN THE TRASH?!?!?
Jeez Louise... ''
Do you people have a brain at all? What he is saying is: Football mum goes in the queue with a bottle of water. They take away the bottle, nobody checked whether she was a threat or not. Terrorist goes in the queue with a bottle of clear liquid that will blow up an aeroplane. They take away the bottle, nobody checked whether he was a threat or not.
By throwing _any_ bottle of sufficient size in the trash, dangerous explosions are prevented without a costly determination whether someone was a threat or not. On one hand, the danger is avoided. On the other hand, terrorists will go undetected and they can try again. That's what he said, and it sounds very reasonable to me.
Many years ago when I took my obligatory first aid course, the MD teaching told us how he distinguishes between civilised and uncivilised country: In a civilised country, if you notice a traffic accident, you stop and help. In an uncivilised country, you drive on.
Some countries in Africa, and the USA, fall under the category "uncivilised".
German rules: You can't get a driver's license without passing a first aid course. You can go to jail for not helping a person in danger. You cannot be prosecuted for any damage that you do while trying to help. Your car insurance will voluntarily pay you for any damages to you that happened while helping in a traffic accident (for good reason, because you might save them lots of money).
'' If someone just went up to the machine and did a withdrawal and they got 4x as much money as they asked for, I would think that's not grounds for prosecution--though they should be expected to return the money. If someone re-plugged the unplugged machine and/or went back for a second withdrawal after observing the problem, I think that's something I'd consider prosecuting. ''
If I ask for say $50, and the machine spits out $200, what should you do? I think first it is obvious that $150 is not _your_ money, it is the banks money. Actually, it is not quite clear, because the bank might have deducted $200 from your account. It is clear that the bank made a mistake, it is not at all obvious which mistake, so you would have to take the money with you and figure out what happened. If you leave $150 there, what would happen? It is either your money, so you should take it, or it is the banks money, in which case you should take the money into your care so nobody takes it away. To be legally on the safe side, you might go to a lost&found office. There is likely a legal difference if the machine is inside the bank's building and you take the banks money off their premises.
Going back for more I think would be considered fraud.
One of the three RSA guys (can't remember which one) came up with a system that is very hard to break: First, instead of giving 1 vote to the person you want and 0 to the others, you give 2 votes to the one you want elected and 1 to everyone else. A voting machine enforces this, so if we count all votes, then subtract the number of voters from each count, we get the real count.
You put your votes on three voting strips. On each voting strip you can have one vote for each candidate (but each candidate must have one vote on one of the strips, except one must have two). A unique random number is added to each strip. All three strips are recorded, votes and number, you pick one that is printed out and you take that one home. Nobody (including the counting machine) knows which one you pick.
All voting strips are published on a website. You can check if the one strip you hold is recorded correctly. If anyone forges the recording, they can get away with one or two changes or maybe four if they are lucky, but any significant changes will be detected, _and_ you would have proof of the forgery. On the other hand, nobody can find out how you voted, because you can put any combination of votes on _one_ strip and still vote for whoever you like. If you are paranoid, you can swap your printout with a random other voter.
' For quite some time now (10.3 Panther I think) there has been a case-sensitive variant of HFS+. '
You won't be too happy with it if you run MacOS X. There is too much code out there that doesn't expect case sensitivity (and too many users as well), and very very few applications are ever tested on a case sensitive file system.
Slashdot Mirror
'' Can one of the cryptographers on slashdot comment on weather this is useful to them or not? ''
One useful addition (copied from Altivec) is the vector permute instruction. What is clever about it in terms of cryptography is that you can translate a vector using a 256 byte translation table _without doing any memory access_ by using the vector permute instruction in a clever way. Now the execution time is completely data-independent, so one important attack vector is closed.
>> And then we have the "holy shit" moment:
For example, the Advanced Encryption Standard (AES) algorithm gets a factor of 5 performance improvement by using the new SSE5 extension
If I get one of these CPUs, I'll almost certainly be encrypting my hard drives. It was already fast enough, but now...
They copied two important features from the PowerPC instruction set: Fused multiply-add (calculate +/- x*y +/- z in one instruction), and the Altivec vector permute instruction, which can among other things rearrange 16 bytes in an arbitrary way. The latter should be really nice for AES, because it does a lot of rearranging 4x4 byte matrices (if I remember correctly).
I don't think having two dual-cores in a package instead of four cores combined is necessarily a disadvantage. To compare these properly, you would have to assume same quality of implementation. So Intel could have gone for one unified 12MB L2 cache with four access paths instead of two 6MB L2 caches with two access paths each. With same quality of implementation, the four access paths will be slower because you have to cope with four processors accessing it at the same time instead of two. So each access will be slower. Larger caches are always slower anyway, so 12 MB is slower than 6 MB again. On the other hand, 12 MB unified is better if you have only one thread that is cache intensive, and the unified cache is better if more than two threads are communicating, and you don't need to worry about allocating threads to the right processors in pairs. Communicating between L2 cache and main memory might be faster for 2x2 cores because you have two access paths, on the other hand making them fast is harder.
So all in all, the speed advantage could go either way, and will depend on the code that you run. For example, if you run four video encoders simultaneously, all the advantages of 1x4 cores don't help, but the higher speed of 2x2 does. Other tasks will be different.
'' 1.6L is considered big? I have a 2L 4cyl Focus in Canada, and that's considered "small" by our standards. Not that I really push my car, but I am curious as to how a 1.3L accelerates [to say hwy speeds]. Because even in my car I have to really floor it [re: 5000 RPM] to hit highway speeds before I exit the ramp, well that's exaggerating a bit. usually I hit speed before the dotted lines (that let you get out of the merging lane). So I probably could accelerate at like 3-4K RPM just fine. ''
Get a Diesel engine. Massive torque = massive acceleration. Not that much horse power, but that only matters at high speeds (100mph+) where you lose your driving license anyway.
In the UK, tax goes by carbon dioxide emission per km, engine size doesn't matter. There is a small number of cars that pay £35 per year, others pay between £115 and >£200 tax per year. But there are other differences: At the moment, you pay a £8 charge every time you drive into London. In the future, that will be free for cars with very low emissions, and up to £25 for very high emissions.
But the thing that really hits is company car tax. If you have a company car, you have to pay income tax on X percent of the value of the new car every year. X ranges from 15% to 35%, depending on carbon dioxide emissions. For a £20,000 car, you pay tax on £3000 to £7000, depending on emissions. At 40% tax rate, that is £1200 to £2800 tax, in other words up to £1600 punishment every year for high carbon dioxide emissions for a £20,000 car.
So they think refusing Firefox users has only little financial impact? What, like five percent? What kind of idiot would you have to be to give up five percent of your revenue voluntarily? Imagine being a webmaster trying to explain to the marketing guys that it is a good idea to reduce their revenues by five percent.
This reminds me of past stupidities, where web designers didn't design for Mac users, because Mac users are idiots who spend much too much money on shiny computers where a $400 Dell would do just fine. And marketing says "idiots who spend much too much money on shiny toys are _exactly_ who we want to visit our website"!
'' That's no excuse! A modern computer is one of the most complex machines known to Man, yet everyone expects to be able to use one without danger. ''
Of course we do. A car weighs usually between 1.5 and 2 tons and often moves at a speed of 70 miles per hour, that is a lot of kinetic energy. A bullet has much less weight, but travels at much higher speed, and if it impacts, all its kinetic energy goes into a small area. A modern computer however weighs maybe 5 pound if it is a laptop, hopefully not more than 30 pound otherwise, and is sitting still. Hard disks are inside a case that should protect me. I hope that the CD drive is carefully designed so that it doesn't crack CDs and shoots out parts of them. Yes, I expect to be able to use a computer without any danger.
>> "They would give software access to information about cache misses..." Yeah that ought to help significantly with side-channel attacks against crypto software.
I think you didn't read the spec. All that information is only available to the thread that is profiled; everything is context-switched so it can't leak out to other threads and definitely not to other processes.
'' It's not like math is an important part of everyday life anyway! Who needs it... ''
I'll give you a real life example, exactly the way that I encountered it:
A person is offered a loan of £8000 at 10% interest p.A. The loan is to be repaid over five years at £200 per month.
Discuss, especially the importance of maths.
'' While most obese individuals just need to start walking and lay off McDonald... ''
Considering how much money is made with products that are supposed to help fat people to get slimmer, and that all these products don't work, I think things might be a bit more difficult than you think.
'' An analysis of 40 studies involving 250,000 people, heart patients with normal BMIs were at higher risk of death from cardiovascular disease than people whose BMIs put them in the "overweight" range (BMI 25-29.9). Patients who were underweight or severely overweight had an increased risk of death from cardiovascular disease. The implications of this finding can be confounded by the fact that many chronic diseases, such as diabetes, cause weight loss before the eventual death. In light of this, higher death rates among thinner people would be the expected result. ''
This is about health insurance, and health insurance is about cost. Dead people don't cost much. Sick people cost lots.
'' Apple should be rejoicing, since they claim that the patent system is running perfectly. ''
Just shows that you didn't read properly. In that recent discussion, Google said that the patent system itself is broken. Apple said the patent system is fine, the related litigation system is broken. And clearly it is.
The other example mentioned was Microsoft being ordered to pay $1.5bn over two MP3 related patents. We all know that MP3 is covered by a few hundred patents, and Microsoft paid a few million for a license for all those patents, so one or two patents they missed could never be worth $1.5bn.
'' Simple solution: Have someone else, who has no ties to Lyons take over the blog. ''
Could you imagine what Apple could do by creating a "Fake Steve Jobs" blog now?
Can you imagine Mr. Lyons trying to sue Apple for using the name "Steve Jobs" in a fake blog?
'' And this is what confuses me. There is no 10.5 for ICBMs ... ''
And I hope there will never be. Ok, probably better than WfW (Windows for Warships).
'' An iPhone will work, but really it could be any item that indicates to the woman that you're willing to spend hundreds of dollars on something pretty. ''
You still have to pull it off. And somehow I think that someone who posts on Slashdot that an iPhone "got him laid" is not the kind of person who _can_ pull it off.
That is basic physics.
Your car is not safer, it is heavier. In a collision, most of the deforming energy goes into the lighter vehicle. Direct consequence of conservation of impulse and conservation of energy.
I am still waiting for lawsuits where an owner of a small car sues the owner of a larger car for damages, because it is _your heavy car_ that caused the damage. If you hadn't acted selfishly, that other car would have sustained much less damage.
I don't know if you realise this, but they are not showing screenshots from the defendants computer.
The RIAA is showing screenshots from the computers of that gang of hackers that they hire to spy on their victims. So it is _totally_ under their control what is in these screenshots.
'' As others have pointed out, ignorance is generally not an excuse. ''
Stop right there. Ignorance of the law is no excuse; if you commit a crime then it doesn't matter whether you knew or not that it was illegal.
Ignorance _is_ often an excuse because it makes the difference between actually having committed a crime or not. Lets say at a car park you put your coat on the back of my car while you tie your shoelaces. I drive away in my car - with your coat on it. If I knew that your coat was there, it is theft. If I didn't know it was there, it is no theft.
In this case, ignorance that someone used your computer to infringe someone's copyright is most definitely an excuse.
'' Do other smart phones require computers? ''
I don't know, but since this is Slashdot, and Slashdot readers would buy a phone for its features, you would surely realise that at least two really important features of the iPhone (use as a video iPod, and use with WiFi) depend on a computer hanging around somewhere. So people without a computer would be very unlikely to buy one.
'' Not everyone who will need to be "in the bush" for a few days will know this much in advance. What happens if you learn you'll need that capability one year into the two-year contract? ''
Must be horrible. A week in the bush, and when you take your iPhone to the bush-battery-shop, they tell you they have no replacement batteries for the iPhone.
'' He said, "Can we have another option to fly? We'll call it Fly At Your Own Risk Airlines. We won't screen for anything and you can pay for your tickets five minutes before your flight just like in the old days-1997." ''
Can you imagine the hilarity when you find out that the other 199 passengers are carrying bombs as well?
'' "If the TSO throws your liquids in the trash, they don't find you a threat."
If they didn't find (you) a threat, then WHY THROW THE FREAKIN' LIQUIDS IN THE TRASH?!?!?
Jeez Louise... ''
Do you people have a brain at all? What he is saying is: Football mum goes in the queue with a bottle of water. They take away the bottle, nobody checked whether she was a threat or not. Terrorist goes in the queue with a bottle of clear liquid that will blow up an aeroplane. They take away the bottle, nobody checked whether he was a threat or not.
By throwing _any_ bottle of sufficient size in the trash, dangerous explosions are prevented without a costly determination whether someone was a threat or not. On one hand, the danger is avoided. On the other hand, terrorists will go undetected and they can try again. That's what he said, and it sounds very reasonable to me.
Many years ago when I took my obligatory first aid course, the MD teaching told us how he distinguishes between civilised and uncivilised country: In a civilised country, if you notice a traffic accident, you stop and help. In an uncivilised country, you drive on.
Some countries in Africa, and the USA, fall under the category "uncivilised".
German rules: You can't get a driver's license without passing a first aid course. You can go to jail for not helping a person in danger. You cannot be prosecuted for any damage that you do while trying to help. Your car insurance will voluntarily pay you for any damages to you that happened while helping in a traffic accident (for good reason, because you might save them lots of money).
'' If someone just went up to the machine and did a withdrawal and they got 4x as much money as they asked for, I would think that's not grounds for prosecution--though they should be expected to return the money. If someone re-plugged the unplugged machine and/or went back for a second withdrawal after observing the problem, I think that's something I'd consider prosecuting. ''
If I ask for say $50, and the machine spits out $200, what should you do? I think first it is obvious that $150 is not _your_ money, it is the banks money. Actually, it is not quite clear, because the bank might have deducted $200 from your account. It is clear that the bank made a mistake, it is not at all obvious which mistake, so you would have to take the money with you and figure out what happened. If you leave $150 there, what would happen? It is either your money, so you should take it, or it is the banks money, in which case you should take the money into your care so nobody takes it away. To be legally on the safe side, you might go to a lost&found office. There is likely a legal difference if the machine is inside the bank's building and you take the banks money off their premises.
Going back for more I think would be considered fraud.
One of the three RSA guys (can't remember which one) came up with a system that is very hard to break: First, instead of giving 1 vote to the person you want and 0 to the others, you give 2 votes to the one you want elected and 1 to everyone else. A voting machine enforces this, so if we count all votes, then subtract the number of voters from each count, we get the real count.
You put your votes on three voting strips. On each voting strip you can have one vote for each candidate (but each candidate must have one vote on one of the strips, except one must have two). A unique random number is added to each strip. All three strips are recorded, votes and number, you pick one that is printed out and you take that one home. Nobody (including the counting machine) knows which one you pick.
All voting strips are published on a website. You can check if the one strip you hold is recorded correctly. If anyone forges the recording, they can get away with one or two changes or maybe four if they are lucky, but any significant changes will be detected, _and_ you would have proof of the forgery. On the other hand, nobody can find out how you voted, because you can put any combination of votes on _one_ strip and still vote for whoever you like. If you are paranoid, you can swap your printout with a random other voter.
' For quite some time now (10.3 Panther I think) there has been a case-sensitive variant of HFS+. '
You won't be too happy with it if you run MacOS X. There is too much code out there that doesn't expect case sensitivity (and too many users as well), and very very few applications are ever tested on a case sensitive file system.