Mod parent up. There are some software patents that are actually worthy of patents. They demonstrate a deep understanding of a problem and a solution that very few people would be able to independently arrive at. RSA is a good example. There are also a lot of patents on hardware that are not. Take a look at a Dyson vacuum cleaner some time. It is covered by almost 200 patents. One, for example, is for rounding corners of tubes that air flows through to reduce turbulence. In the patent, it specifically states that this technique has been used in large industrial sites for decades, but apparently because the tubes are small it's now worth a patent.
Because now you wouldn't be doing it in your garage. That kind of R&D effort will be confined to big companies who have cross-licensing agreements and so can pretend that the patent system doesn't exist among themselves.
Not sure which BBC you were reading, but the Pirate Party got a disproportionate amount of coverage on the BBC news site that I read. I wouldn't be surprised if they had a higher ratio of words written on the BBC news site to votes than any other political party.
Neat, but not especially hard. There have been several X11 implementations. I remember one written in Java about 10 years ago (maybe this is a port of that one?). There are several C implementations, suck as kdrive. Implementing X is not hard for the core protocol. Most of the complexity in X.org is in the drivers. Beyond that, there are all of the extensions.
Or you could just fire up iTunes and have it rip the DVD for you. Oh, wait, no you can't because the consortium including the company launching this service refuses to provide a DVD license to companies producing software that does this and lobbied for laws that make it possible for them to go after companies that do.
Really? I downloaded the Mac build of Handbrake over Christmas and it worked fine out of the box. These days I just use dvdbackup and dump the VOB files on my NAS. Disk space is so cheap that there's little point in recompressing, and this way I get all of the menus and special features - vlc plays them back quite happily.
Who says it is? It's only been reproduced consistently on DragonFly BSD. There's no reason to believe that it hasn't caused stack corruption on other systems...
Nonense. It doesn't happen every time, or it would have been trivial to find. It happens under load, so it is most likely dependent on some other factor (e.g. CPU temperature, cache miss ratio, context switch timing) which may not have appeared in testing.
Meh, if Intel maintains domination over laptops, desktops, servers and supercomputers
That's a big if. Desktops, sure - but they're a shrinking market. Laptops? Well, I have one Intel laptop and one ARM laptop at the moment. The Intel one wins on raw speed, but the ARM one wins on everything else (battery life, cost, and so on). And increasingly laptops are likely to want to use the same chips as smartphones and tablets, which are rapidly growing markets. Servers? They increasingly care about power consumption. The A15 is likely to make big inroads into this market, because you can fit quite a lot of them in a 1U chassis without loading the power supply or air conditioning. A lot of the rest of this market is small business servers, where having something small and appliance-like is very useful, and these tend to use MIPS or ARM chips (historically more MIPS than ARM, but this is changing). Supercomputers? This is where ARM is currently pushing, with CPU and GPU on the same die, behind the same memory controller, so running part of a calculation on the GPU is only marginally more expensive than running it on the SIMD unit. nVidia is aiming squarely at this market with their ARMv8 offerings, due later this year. I wouldn't bet on Intel holding on to that market, although no doubt they'll remain a significant player in it for a while.
The ARM heads love to go on about that but as of yet there are no ARM chips that compete in the desktop space
In case you haven't been paying attention for the past 5 years, the desktop is a shrinking market. The big growth markets for CPUs are mobile and low power servers. If you want to know how well being the market leader in a shrinking market segment works out in the long run, just ask SGI...
Intel can pour money into process R&D because they have lots of money. It's easy to forget how much bigger Intel is than AMD. Intel spent about $6.5bn on R&D last year. This is more than AMD's total revenue. It is simply not possible for AMD to spend more than about a fifth of what Intel does on R&D. Their sales volumes are also much lower, so they can't amortise the R&D cost over a lot more chips. For every CPU AMD sells, Intel sells four. That means that if Intel spends one dollar per chip on R&D and AMD spends one dollar per chip on R&D, then Intel is still outspending AMD 4:1.
That was the entire point of spinning off GF in the first place. AMD can't outspend Intel on process R&D (including fab construction), but AMD plus other GF customers might be able to between them.
The two are not unrelated. ARM is pushing Intel on the high-volume, low margin end. ARM SoC vendors will sell you a SoC for less than Intel will sell you a CPU - and it will use a tenth the power. The high end is getting really small. Supercomputers are still around, but that entire market is well under a million CPUs per year, and it's being attacked by GPU and DSP makers. Worse, it's become a lot cheaper to design a custom ASIC over the past decade, so a lot of tasks where performance is all that matters are now being done on dedicated silicon, rather than off-the-shelf Intel / IBM / DEC CPUs as they were in the '90s. You can do a run of a thousand custom chips on a decent process for a lot less than the cost of building a supercomputer, and it takes a lot less space. The middle, where Intel is king, is also shrinking. In the 90s, a computer more than six months old was slow. Now people are using 5-10 year old computers and, aside from a few niches like gamers (who are increasingly using consoles, or playing console ports so not needing the latest and greatest desktop) finding that they're still mostly fast enough.
Not true. pop, pop, ret is a sequence that you are likely to see in any function that makes use of two or more callee-save registers - it will push them before using them and then pop them at the end. If you're lucky, the register allocator will have done some peephole optimisation and moved the pops earlier...
Re:Linux security or trust
on
GitHub Hacked
·
· Score: 1
A commit made on github by Linus is not automatically pulled into kernel.org. It will be reviewed and merged. Most likely someone would say 'Linus, why did you commit here when you are in charge of the upstream repository? That was a strange thing to do' and he would say 'I did not to that, let us inspect the diff and see if someone has done something malicious'. It's pretty easy to spot this kind of thing...
Re:Linux security or trust
on
GitHub Hacked
·
· Score: 1
In this case, however, git itself was not hacked. The web interface of github was hacked.
I don't think the second part of that is true. The problem with a benevolent monarchy is that it often degenerates into a malevolent monarchy, and then people revolt. Most benevolent dictatorships become less benevolent over the lifetime of the dictator. Surviving for two leaders is very hard.
I've not really paid attention to Santorum, but listening to Obama speak and looking at what he did at Harvard it's pretty obvious that he's smarter than average. Whether or not he is more competent as a political leader than average is not necessarily dependent on this.
Re:Linux security or trust
on
GitHub Hacked
·
· Score: 5, Informative
That's idiocy on the part of the submitter. Linux is mirrored on github, and it was the authoritative repository for a while after kernel.org was hacked, but now it is not the authoritative repository and patches from there will not be pulled into the official tree unchecked.
Slashdot Mirror
Mod parent up. There are some software patents that are actually worthy of patents. They demonstrate a deep understanding of a problem and a solution that very few people would be able to independently arrive at. RSA is a good example. There are also a lot of patents on hardware that are not. Take a look at a Dyson vacuum cleaner some time. It is covered by almost 200 patents. One, for example, is for rounding corners of tubes that air flows through to reduce turbulence. In the patent, it specifically states that this technique has been used in large industrial sites for decades, but apparently because the tubes are small it's now worth a patent.
Because now you wouldn't be doing it in your garage. That kind of R&D effort will be confined to big companies who have cross-licensing agreements and so can pretend that the patent system doesn't exist among themselves.
$15 to Microsoft and $15 to Apple for each Android device? Suddenly Android doesn't look like quite such a cheap option...
Not sure which BBC you were reading, but the Pirate Party got a disproportionate amount of coverage on the BBC news site that I read. I wouldn't be surprised if they had a higher ratio of words written on the BBC news site to votes than any other political party.
There is a 1 in 625 chance that I will be taking a long holiday and be unavailable for comment in 2040...
Do you have an actual point, or are you still 12?
Neat, but not especially hard. There have been several X11 implementations. I remember one written in Java about 10 years ago (maybe this is a port of that one?). There are several C implementations, suck as kdrive. Implementing X is not hard for the core protocol. Most of the complexity in X.org is in the drivers. Beyond that, there are all of the extensions.
Or you could just fire up iTunes and have it rip the DVD for you. Oh, wait, no you can't because the consortium including the company launching this service refuses to provide a DVD license to companies producing software that does this and lobbied for laws that make it possible for them to go after companies that do.
Really? I downloaded the Mac build of Handbrake over Christmas and it worked fine out of the box. These days I just use dvdbackup and dump the VOB files on my NAS. Disk space is so cheap that there's little point in recompressing, and this way I get all of the menus and special features - vlc plays them back quite happily.
Then why has it only been seen on DragonFly BSD?
Who says it is? It's only been reproduced consistently on DragonFly BSD. There's no reason to believe that it hasn't caused stack corruption on other systems...
Nonense. It doesn't happen every time, or it would have been trivial to find. It happens under load, so it is most likely dependent on some other factor (e.g. CPU temperature, cache miss ratio, context switch timing) which may not have appeared in testing.
Meh, if Intel maintains domination over laptops, desktops, servers and supercomputers
That's a big if. Desktops, sure - but they're a shrinking market. Laptops? Well, I have one Intel laptop and one ARM laptop at the moment. The Intel one wins on raw speed, but the ARM one wins on everything else (battery life, cost, and so on). And increasingly laptops are likely to want to use the same chips as smartphones and tablets, which are rapidly growing markets. Servers? They increasingly care about power consumption. The A15 is likely to make big inroads into this market, because you can fit quite a lot of them in a 1U chassis without loading the power supply or air conditioning. A lot of the rest of this market is small business servers, where having something small and appliance-like is very useful, and these tend to use MIPS or ARM chips (historically more MIPS than ARM, but this is changing). Supercomputers? This is where ARM is currently pushing, with CPU and GPU on the same die, behind the same memory controller, so running part of a calculation on the GPU is only marginally more expensive than running it on the SIMD unit. nVidia is aiming squarely at this market with their ARMv8 offerings, due later this year. I wouldn't bet on Intel holding on to that market, although no doubt they'll remain a significant player in it for a while.
The ARM heads love to go on about that but as of yet there are no ARM chips that compete in the desktop space
In case you haven't been paying attention for the past 5 years, the desktop is a shrinking market. The big growth markets for CPUs are mobile and low power servers. If you want to know how well being the market leader in a shrinking market segment works out in the long run, just ask SGI...
Intel can pour money into process R&D because they have lots of money. It's easy to forget how much bigger Intel is than AMD. Intel spent about $6.5bn on R&D last year. This is more than AMD's total revenue. It is simply not possible for AMD to spend more than about a fifth of what Intel does on R&D. Their sales volumes are also much lower, so they can't amortise the R&D cost over a lot more chips. For every CPU AMD sells, Intel sells four. That means that if Intel spends one dollar per chip on R&D and AMD spends one dollar per chip on R&D, then Intel is still outspending AMD 4:1.
That was the entire point of spinning off GF in the first place. AMD can't outspend Intel on process R&D (including fab construction), but AMD plus other GF customers might be able to between them.
The two are not unrelated. ARM is pushing Intel on the high-volume, low margin end. ARM SoC vendors will sell you a SoC for less than Intel will sell you a CPU - and it will use a tenth the power. The high end is getting really small. Supercomputers are still around, but that entire market is well under a million CPUs per year, and it's being attacked by GPU and DSP makers. Worse, it's become a lot cheaper to design a custom ASIC over the past decade, so a lot of tasks where performance is all that matters are now being done on dedicated silicon, rather than off-the-shelf Intel / IBM / DEC CPUs as they were in the '90s. You can do a run of a thousand custom chips on a decent process for a lot less than the cost of building a supercomputer, and it takes a lot less space. The middle, where Intel is king, is also shrinking. In the 90s, a computer more than six months old was slow. Now people are using 5-10 year old computers and, aside from a few niches like gamers (who are increasingly using consoles, or playing console ports so not needing the latest and greatest desktop) finding that they're still mostly fast enough.
Okay, you do that in your compiler first, and then we'll add more binary size and i-cache usage benchmarks to our marketing materials...
Can I come and live in your world where FreeBSD 5.x never happened please?
GCC dies with "internal error" compiling some of my software
Is this different from the expected behaviour in some way?
Not true. pop, pop, ret is a sequence that you are likely to see in any function that makes use of two or more callee-save registers - it will push them before using them and then pop them at the end. If you're lucky, the register allocator will have done some peephole optimisation and moved the pops earlier...
A commit made on github by Linus is not automatically pulled into kernel.org. It will be reviewed and merged. Most likely someone would say 'Linus, why did you commit here when you are in charge of the upstream repository? That was a strange thing to do' and he would say 'I did not to that, let us inspect the diff and see if someone has done something malicious'. It's pretty easy to spot this kind of thing...
In this case, however, git itself was not hacked. The web interface of github was hacked.
I believe that the point of the grandparent's post was to provide a concrete example of the mechanism described in TFA.
I don't think the second part of that is true. The problem with a benevolent monarchy is that it often degenerates into a malevolent monarchy, and then people revolt. Most benevolent dictatorships become less benevolent over the lifetime of the dictator. Surviving for two leaders is very hard.
I've not really paid attention to Santorum, but listening to Obama speak and looking at what he did at Harvard it's pretty obvious that he's smarter than average. Whether or not he is more competent as a political leader than average is not necessarily dependent on this.
That's idiocy on the part of the submitter. Linux is mirrored on github, and it was the authoritative repository for a while after kernel.org was hacked, but now it is not the authoritative repository and patches from there will not be pulled into the official tree unchecked.