As a scientific programmer, all I know is that this will eventually be a huge benefit to all my MPI and OpenMP codes.
Unfortunately, these kinds of processors are pointless for most scientific applications (there are some exceptions, but not many). Scientific apps are limited by arithmetic units and memory bandwidth, and these processors do nothing to improve them. The Cell processor at least has multiple FPUs, this one doesn't even have that.
For your MPI codes, you are much better off using a workstation cluster, because unlike with these kinds of processors, you get a separate memory subsystem and a separate FPU with each thread that way.
Threads are actually one of the simplest form of parallelism to deal with and we have had decades of experience with them. That's why Sun loves them: it fits in well with their big-iron philosophy and hardware and makes it easy for their customers to migrate to the next generation.
But the future of high-end computing, both in business and in science, will not look like that. Networks of cheap computing nodes scale better and more cost-effectively. Many manufacturers have already gone over to that for their high-end designs. That's where the real software challenges are, but they are being addressed.
Processors with lots of thread parallelism will probably be useful in some niche applications, but they will not become a staple of high-end computing.
Well, that's quickly changing: these days, computers can break themselves, be it via automatic upgrades, spyware, or worms that come in through vendor-supplied security holes.
You have to remember that, although low, we have also had some inflation over the last 20-30 years. So, that $300 PC is more like a $150 machine of a couple of decades ago. Compare that with the VIC-20, which cost about $400 in 1981 (with 64k of memory).
It's not quite the same. Using Macs, you get both your hardware, your OS and many applications from the same vendor.
There are several vendors of Linux machines that give you the hardware, software, expansion hardware, and software upgrades all from a single source. And they work.
You know your wifi-card is going to work with your updated OS without you having to find an upgrade for the driver. You know your internal bluetooth is going to work. You know the OS will support your keyboard's backlight or your PowerBook's motion sensor out of the box.
Well, at least that's what Apple promises...
There's no custom hardware that needs special support. Every piece of hardware a Mac contains when you buy it just works without you having to find drivers, even if you format your disk and install the OS from scratch.
Yes, and you get exactly the same behavior when you buy a Linux or Windows machine from a vendor that provides updates for their systems.
So what's unreliable? A lot of USB stuff doesn't have Mac drivers because it's not needed -- the Mac has built-in support for much of it.
Same in Linux.
As for third-party USB hardware, I've not had a problem. My Macs have lots of USB accessories:
All of that hardware works with Windows and Linux as well.
Open source UNIX-alikes will never gain much market-share
Open source UNIX-alikes already have a larger market share than Macintosh.
True, Macs work best with Apple hardware... which makes sense, since that means they've been validated to work together from day one.
And the same is true for Linux and Windows: buy hardware that is supported by, and tested with, the OS, and you are going to be fine on any OS. Macintosh is no better than Linux in this regard.
Submitting papers in DOC format is generally not a good idea, for many reasons. Anybody using DOC format should also be able to use RTF format. An even better choice is PDF; OpenOffice can export to PDF.
You did, just now. Your suggesting we refuse manned exploration unless it makes economic sense. Science says we send manned probes so long as there is data that can be gathered by a human being that can not be gathered by unmanned probes.
No, that's not what "science says". Science is always about making tradeoffs under budgetary constraints.
Considering that the most valuable potential results of space exploration involve humans leaving Earth it makes good sense to continue gathering data and researching how to get us farther away.
Even if your premise were true, continuing manned space travel using known technologies doesn't yield any useful data or research results. The best way of advancing manned space travel right now would be through funding for new propulsion technologies, antimatter generation, fusion, and medicine, and those are either terrestrial or unmanned projects for now.
Only idiots expect to turn on a computer, slap in a card, run an automated driver install program and expect the thing to work.
I had to try three 802.11 USB sticks before I found one that actually worked on Windows. I have been through four Bluetooth USB devices, and none of them work correctly.
Macintosh is even worse: most of the USB hardware I have doesn't even have drivers for Macintosh, so it won't work at all. For supposedly supported hardware, the track record is not much better than on Windows. The only thing that I found works reliably on Macintosh is all-Apple hardware.
So, please stop spreading FUD: this is a big problem with all current operating systems. The only way you can avoid it is by picking the hardware and software you install very carefully to get the stuff that works. And that's true on all platforms. In fact, its true for most high-tech products we buy in general.
I personally don't really care about whether Jamie switches to Mac, Windows or a toaster. I believe the main issue here is that after oh so many years of dev being done, Linux still ain't ready for the desktop. Period.
Hey, I keep trying to install OS X on my AMD64, but audio just won't work. Obviously, OS X isn't ready for the desktop if something as simple as audio doesn't work on my PC.
After all, you are also a UNIX Hater. What I can't figure out is why you guys still bother us UNIX and X11 users. Apple and Microsoft have answered your prayers: 95% of the market is using systems that, according to your criteria, are far superior to UNIX and X11. Just go play with your Windows PC or Macintosh, and leave us to do our own thing, OK?
Zawinski's flames against the X11 window system have not helped the platform. Let him live with XCode, Objective-C, and Quartz for a while, and let the Macintosh community put up with him.
this research is a wonderful example of...
on
The Evil in E-Mail
·
· Score: 2, Insightful
Graduate students, take notice. This research is a wonderful example of... going where the wind is blowing; that gives you media coverage and funding from people who know even less than you... not doing your background research; doing your background research would just discourage you, and it takes time that isn't required for convincing people who know less than you that your sexy proposal is worth funding
I don't understand the use of the word "paradoxically" in this context. Microsoft is doing what every corporation needs to do: they are maximizing profit. If in China that involves collaborating with an undemocratic regime and restricting speech.
Corporations generally do what they can get away with legally and what governments require of them. If we don't want them to do something, we, the people, have to impose regulations and requirements on them. So, if we don't want Microsoft to monopolize the market, then we have to use laws and courts to prevent them from doing so. If we don't want Microsoft to restrict free speech in China, then we have to impose regulations on them that keep them from doing that.
Corporations aren't people. They don't have a conscience or morality. They aren't "good" or "bad", and notions of social responsibility, duty, or morality are foreign to them. They are rational agents that respond to risk and regulation. So, if you don't like this, don't blame Microsoft for this, blame US regulators for it if they continue to permit it.
Objective-C was a derivative of C that incorporated a few bits and pieces of Smalltalk.
OS X uses what Apple likes to refer to as "Display PDF". Apple was forced to change over from DPS (Display Postscript) because DPS turned out to be a complete disaster. DPS was released by NeXT in 1987; I believe NeWS was an independent development from Sun, but it may not have been. However, both NeWS and DPS were failures, and for good reason.
I think the technologies you deride were ahead of their time, and the technologies you espouse are popular due to their simplicity, not their elegance.
Objective-C and DPS were just evolutionary dead ends: poorly conceived derivative designs that would have been long dead and buried if not resurrected by Apple.
Mach innovated, but it turned out that its approach to OS design hasn't worked very well so far. I can't tell yet whether that is because Mach was a poor implementation or whether microkernel architectures themselves are a bad idea. In any case, right now, Linux is a more practical choice, even though it has a 1960's OS architecture.
X11 was a novel approach to window system design at the time, an approach that has been enormously successful and stood the test of time.
Java is a traditional OOL, something we could already have used in the 1970's, and C# has some significant technical enhancements that take it into the 1980's. Neither of them is innovative, but both are far better designed languages than Objective-C.
In any case, I don't want technologies that are "ahead of their time", I want technologies that work and grow well in the real world.
because they have chosen this cool new processor architecture that lets them take advantage of all these innovations. Xserve won't be just a zillion times faster than Windows and Linux servers, it will be a ZILLION zillion times faster and all because Jobs is so smart that he picked the x86 architecture, while Windows and Linux have to make do with... oh, wait.
Maybe in 20years time the automatic parts will improve by several magnitudes but problems like the speed of light - think comand and controll - will remain.
Sufficiently intelligent probes do not need to communicate in real time.
In any case, I think we basically agree: unmanned probes for the next couple of decades, and we can reevaluate the tradeoffs again then.
Willfully refusing to conduct experiments that would yield valuable data for ANY reason is an opposition to science.
Who said anything about "willfully refusing"? In cases where manned exploration yields the same bang-for-the-buck as unmanned probes, I'm all for manned exploration. Until then, we should stick with unmanned probes.
However, there are numerous things that they can not do and there is nothing that can do that can not be done by a human supplied with life essentials.
Most scientifically interesting missions are, in fact, only possible with unmanned probes at this point. Human missions are infeasible for many reasons: humans want to have a return trip, they need to be shielded from radiation, and they can't endure years in space, for example.
I think it's a serious problem that unmanned and manned space exploration are so intertwined.
Unmanned exploration is science and should be funded as such. It shouldn't have to compete with politically motivated "manned exploration" projects.
In fact, ideally, I think NASA should leave manned exploration to the private sector. NASA should be turned into an agency dedicated to unmanned exploration: remote sensing, robotics, and new propulsion technologies.
Remember, you said that removing an App bundle was playing "russian roulette" with the system, and that people had "abandoned" the mac because it was incapable of supporting complex applications. Both unsubstantiated claims that you have yet to provide a shred of evidence to support.
There is nothing wrong with removing app bundles. The problems are with users removing applications. Your (and Apple's) fault is to think that app bundles and applications can be equated.
If you look at applications on Linux and Windows systems, you see that many of them cannot be represented as app bundles; for example, many web applications modify global Apache configuration files, add stuff to the registry or etc. directory, create databases, and add new subdirectories to the web directory. They may also generate large log files and cache files. Linux and Windows uninstallers can undo those changes cleanly, Apple's app bundles can't. With drag-and-drop installs, like on Apple, if you just undo the installation by dragging the install into the trash, all the other stuff the application changed will still hang around.
I've backed that up with fairly simplistic but detailed enough descriptions of how bundles prevent what you're claiming they cause,
You haven't backed up anything, you simply keep repeating how you think things are intended to work. But it is pure speculation on your part that those mechanisms actually accomplish what they are intended to accomplish for the average computer user. Demonstrating that requires usability tests, and none of those have been published for the Macintosh. All the usability assertions and the "it just works" claims about Macintosh are hot air from the Apple marketing department and people like you.
The data we have is that Macintosh market share has declined greatly, from about 16% at its top to about 2-3% today, and that Macintosh has been losing market share from 2000-2004 in most market segments except for education (source: Wall Street Journal). There must be a reason for those declines. Both of us can only speculate what the reason is. My explanation (based on my experience with the platform) is that the platform is technically deficient and does not satisfy the needs of most computer users. What's your explanation?
If you really care about a system and minimal downtime due to upgrades, have two root partitions (it's only an extra 5-10G). Instead of upgrading, you make a clean install on the unused root partition. Clean installs generally work better to begin with, and you have the old install both mounted and bootable to figure out any problems and copy over configuration files.
As for complaints about this sort of thing, I still prefer Debian. I just spent several hours upgrading an OS X system from Jaguar to Tiger. A trivial file system inconsistency in HFS caused the installer to crash reproducibly and eventually required me to manually patch inodes (apparently a fairly common problem on Macintosh). And I'll have to wipe a Windows machine clean tomorrow because mysteriously hardware has stopped working and no amount of fiddling will do.
In comparison, these Debian upgrade woes seem minor. And unlike the Mac and Windows problems, the Debian upgrade problems will generally fix themselves after a few days when the package maintainers catch up.
You've completely lost it, I never talked about usability, growth, performance or ease of development.
But you did: you were claiming that Mac software installation is a simple, functional solution. Those are usability claims, whether you realize it or not.
I provided you with detailed descriptions of what I was talking about, which is package management on the Mac.
That was not necessary because I understand how software installation works on the Mac (as we speak, I'm trying to get a Mac upgraded to Tiger and the f*cking Tiger installer keeps crashing with obscure error messages).
I don't know what other data I can supply you with about that subject.
Yes, and that is the problem: you have no data to support your claims that Macintosh package management is "simple" or that it "works". If anything, the history of the Macintosh platform suggests that it has some serious problems that keep it from catching on on the desktop.
Dude, don't care.
Dude, and I do care when Apple's unsubstantiated marketing claims get parrotted uncritically and start affecting community efforts like Linux.
Slashdot Mirror
As a scientific programmer, all I know is that this will eventually be a huge benefit to all my MPI and OpenMP codes.
Unfortunately, these kinds of processors are pointless for most scientific applications (there are some exceptions, but not many). Scientific apps are limited by arithmetic units and memory bandwidth, and these processors do nothing to improve them. The Cell processor at least has multiple FPUs, this one doesn't even have that.
For your MPI codes, you are much better off using a workstation cluster, because unlike with these kinds of processors, you get a separate memory subsystem and a separate FPU with each thread that way.
Threads are actually one of the simplest form of parallelism to deal with and we have had decades of experience with them. That's why Sun loves them: it fits in well with their big-iron philosophy and hardware and makes it easy for their customers to migrate to the next generation.
But the future of high-end computing, both in business and in science, will not look like that. Networks of cheap computing nodes scale better and more cost-effectively. Many manufacturers have already gone over to that for their high-end designs. That's where the real software challenges are, but they are being addressed.
Processors with lots of thread parallelism will probably be useful in some niche applications, but they will not become a staple of high-end computing.
You don't clean your refrigerator and your microwave? That's disgusting.
Computers don't break themselves. Users break computers.
Well, that's quickly changing: these days, computers can break themselves, be it via automatic upgrades, spyware, or worms that come in through vendor-supplied security holes.
You have to remember that, although low, we have also had some inflation over the last 20-30 years. So, that $300 PC is more like a $150 machine of a couple of decades ago. Compare that with the VIC-20, which cost about $400 in 1981 (with 64k of memory).
Wouldn't it be nice if Dell preconfigured Firefox, Thunderbird, and OpenOffice, and stripped out some of the junk in Microsoft Windows?
It's not quite the same. Using Macs, you get both your hardware, your OS and many applications from the same vendor.
There are several vendors of Linux machines that give you the hardware, software, expansion hardware, and software upgrades all from a single source. And they work.
You know your wifi-card is going to work with your updated OS without you having to find an upgrade for the driver. You know your internal bluetooth is going to work. You know the OS will support your keyboard's backlight or your PowerBook's motion sensor out of the box.
Well, at least that's what Apple promises...
There's no custom hardware that needs special support. Every piece of hardware a Mac contains when you buy it just works without you having to find drivers, even if you format your disk and install the OS from scratch.
Yes, and you get exactly the same behavior when you buy a Linux or Windows machine from a vendor that provides updates for their systems.
So what's unreliable? A lot of USB stuff doesn't have Mac drivers because it's not needed -- the Mac has built-in support for much of it.
Same in Linux.
As for third-party USB hardware, I've not had a problem. My Macs have lots of USB accessories:
All of that hardware works with Windows and Linux as well.
Open source UNIX-alikes will never gain much market-share
Open source UNIX-alikes already have a larger market share than Macintosh.
True, Macs work best with Apple hardware... which makes sense, since that means they've been validated to work together from day one.
And the same is true for Linux and Windows: buy hardware that is supported by, and tested with, the OS, and you are going to be fine on any OS. Macintosh is no better than Linux in this regard.
Submitting papers in DOC format is generally not a good idea, for many reasons. Anybody using DOC format should also be able to use RTF format. An even better choice is PDF; OpenOffice can export to PDF.
You did, just now. Your suggesting we refuse manned exploration unless it makes economic sense. Science says we send manned probes so long as there is data that can be gathered by a human being that can not be gathered by unmanned probes.
No, that's not what "science says". Science is always about making tradeoffs under budgetary constraints.
Considering that the most valuable potential results of space exploration involve humans leaving Earth it makes good sense to continue gathering data and researching how to get us farther away.
Even if your premise were true, continuing manned space travel using known technologies doesn't yield any useful data or research results. The best way of advancing manned space travel right now would be through funding for new propulsion technologies, antimatter generation, fusion, and medicine, and those are either terrestrial or unmanned projects for now.
Only idiots expect to turn on a computer, slap in a card, run an automated driver install program and expect the thing to work.
I had to try three 802.11 USB sticks before I found one that actually worked on Windows. I have been through four Bluetooth USB devices, and none of them work correctly.
Macintosh is even worse: most of the USB hardware I have doesn't even have drivers for Macintosh, so it won't work at all. For supposedly supported hardware, the track record is not much better than on Windows. The only thing that I found works reliably on Macintosh is all-Apple hardware.
So, please stop spreading FUD: this is a big problem with all current operating systems. The only way you can avoid it is by picking the hardware and software you install very carefully to get the stuff that works. And that's true on all platforms. In fact, its true for most high-tech products we buy in general.
I personally don't really care about whether Jamie switches to Mac, Windows or a toaster. I believe the main issue here is that after oh so many years of dev being done, Linux still ain't ready for the desktop. Period.
Hey, I keep trying to install OS X on my AMD64, but audio just won't work. Obviously, OS X isn't ready for the desktop if something as simple as audio doesn't work on my PC.
Geez, get a clue.
After all, you are also a UNIX Hater. What I can't figure out is why you guys still bother us UNIX and X11 users. Apple and Microsoft have answered your prayers: 95% of the market is using systems that, according to your criteria, are far superior to UNIX and X11. Just go play with your Windows PC or Macintosh, and leave us to do our own thing, OK?
Zawinski's flames against the X11 window system have not helped the platform. Let him live with XCode, Objective-C, and Quartz for a while, and let the Macintosh community put up with him.
Graduate students, take notice. This research is a wonderful example of ... going where the wind is blowing; that gives you media coverage and funding from people who know even less than you ... not doing your background research; doing your background research would just discourage you, and it takes time that isn't required for convincing people who know less than you that your sexy proposal is worth funding
You can screw with the programming foundations only after you kill Agent Smith.
I don't understand the use of the word "paradoxically" in this context. Microsoft is doing what every corporation needs to do: they are maximizing profit. If in China that involves collaborating with an undemocratic regime and restricting speech.
Corporations generally do what they can get away with legally and what governments require of them. If we don't want them to do something, we, the people, have to impose regulations and requirements on them. So, if we don't want Microsoft to monopolize the market, then we have to use laws and courts to prevent them from doing so. If we don't want Microsoft to restrict free speech in China, then we have to impose regulations on them that keep them from doing that.
Corporations aren't people. They don't have a conscience or morality. They aren't "good" or "bad", and notions of social responsibility, duty, or morality are foreign to them. They are rational agents that respond to risk and regulation. So, if you don't like this, don't blame Microsoft for this, blame US regulators for it if they continue to permit it.
You got your history wrong.
Objective-C was a derivative of C that incorporated a few bits and pieces of Smalltalk.
OS X uses what Apple likes to refer to as "Display PDF". Apple was forced to change over from DPS (Display Postscript) because DPS turned out to be a complete disaster. DPS was released by NeXT in 1987; I believe NeWS was an independent development from Sun, but it may not have been. However, both NeWS and DPS were failures, and for good reason.
I think the technologies you deride were ahead of their time, and the technologies you espouse are popular due to their simplicity, not their elegance.
Objective-C and DPS were just evolutionary dead ends: poorly conceived derivative designs that would have been long dead and buried if not resurrected by Apple.
Mach innovated, but it turned out that its approach to OS design hasn't worked very well so far. I can't tell yet whether that is because Mach was a poor implementation or whether microkernel architectures themselves are a bad idea. In any case, right now, Linux is a more practical choice, even though it has a 1960's OS architecture.
X11 was a novel approach to window system design at the time, an approach that has been enormously successful and stood the test of time.
Java is a traditional OOL, something we could already have used in the 1970's, and C# has some significant technical enhancements that take it into the 1980's. Neither of them is innovative, but both are far better designed languages than Objective-C.
In any case, I don't want technologies that are "ahead of their time", I want technologies that work and grow well in the real world.
because they have chosen this cool new processor architecture that lets them take advantage of all these innovations. Xserve won't be just a zillion times faster than Windows and Linux servers, it will be a ZILLION zillion times faster and all because Jobs is so smart that he picked the x86 architecture, while Windows and Linux have to make do with ... oh, wait.
Maybe in 20years time the automatic parts will improve by several magnitudes but problems like the speed of light - think comand and controll - will remain.
Sufficiently intelligent probes do not need to communicate in real time.
In any case, I think we basically agree: unmanned probes for the next couple of decades, and we can reevaluate the tradeoffs again then.
Willfully refusing to conduct experiments that would yield valuable data for ANY reason is an opposition to science.
Who said anything about "willfully refusing"? In cases where manned exploration yields the same bang-for-the-buck as unmanned probes, I'm all for manned exploration. Until then, we should stick with unmanned probes.
However, there are numerous things that they can not do and there is nothing that can do that can not be done by a human supplied with life essentials.
Most scientifically interesting missions are, in fact, only possible with unmanned probes at this point. Human missions are infeasible for many reasons: humans want to have a return trip, they need to be shielded from radiation, and they can't endure years in space, for example.
I think it's a serious problem that unmanned and manned space exploration are so intertwined.
Unmanned exploration is science and should be funded as such. It shouldn't have to compete with politically motivated "manned exploration" projects.
In fact, ideally, I think NASA should leave manned exploration to the private sector. NASA should be turned into an agency dedicated to unmanned exploration: remote sensing, robotics, and new propulsion technologies.
Remember, you said that removing an App bundle was playing "russian roulette" with the system, and that people had "abandoned" the mac because it was incapable of supporting complex applications. Both unsubstantiated claims that you have yet to provide a shred of evidence to support.
There is nothing wrong with removing app bundles. The problems are with users removing applications. Your (and Apple's) fault is to think that app bundles and applications can be equated.
If you look at applications on Linux and Windows systems, you see that many of them cannot be represented as app bundles; for example, many web applications modify global Apache configuration files, add stuff to the registry or etc. directory, create databases, and add new subdirectories to the web directory. They may also generate large log files and cache files. Linux and Windows uninstallers can undo those changes cleanly, Apple's app bundles can't. With drag-and-drop installs, like on Apple, if you just undo the installation by dragging the install into the trash, all the other stuff the application changed will still hang around.
I've backed that up with fairly simplistic but detailed enough descriptions of how bundles prevent what you're claiming they cause,
You haven't backed up anything, you simply keep repeating how you think things are intended to work. But it is pure speculation on your part that those mechanisms actually accomplish what they are intended to accomplish for the average computer user. Demonstrating that requires usability tests, and none of those have been published for the Macintosh. All the usability assertions and the "it just works" claims about Macintosh are hot air from the Apple marketing department and people like you.
The data we have is that Macintosh market share has declined greatly, from about 16% at its top to about 2-3% today, and that Macintosh has been losing market share from 2000-2004 in most market segments except for education (source: Wall Street Journal). There must be a reason for those declines. Both of us can only speculate what the reason is. My explanation (based on my experience with the platform) is that the platform is technically deficient and does not satisfy the needs of most computer users. What's your explanation?
If you really care about a system and minimal downtime due to upgrades, have two root partitions (it's only an extra 5-10G). Instead of upgrading, you make a clean install on the unused root partition. Clean installs generally work better to begin with, and you have the old install both mounted and bootable to figure out any problems and copy over configuration files.
As for complaints about this sort of thing, I still prefer Debian. I just spent several hours upgrading an OS X system from Jaguar to Tiger. A trivial file system inconsistency in HFS caused the installer to crash reproducibly and eventually required me to manually patch inodes (apparently a fairly common problem on Macintosh). And I'll have to wipe a Windows machine clean tomorrow because mysteriously hardware has stopped working and no amount of fiddling will do.
In comparison, these Debian upgrade woes seem minor. And unlike the Mac and Windows problems, the Debian upgrade problems will generally fix themselves after a few days when the package maintainers catch up.
don't read the freaking release notes, you will have problems
That's a bug with the software, not with the users.
Also, this breakage gives us a yet another reason to bash C++ as a poor excuse for a language
It's not C++'s fault, at best it's the fault of the compiler/ABI. Even C compilers have problems like this.
You've completely lost it, I never talked about usability, growth, performance or ease of development.
But you did: you were claiming that Mac software installation is a simple, functional solution. Those are usability claims, whether you realize it or not.
I provided you with detailed descriptions of what I was talking about, which is package management on the Mac.
That was not necessary because I understand how software installation works on the Mac (as we speak, I'm trying to get a Mac upgraded to Tiger and the f*cking Tiger installer keeps crashing with obscure error messages).
I don't know what other data I can supply you with about that subject.
Yes, and that is the problem: you have no data to support your claims that Macintosh package management is "simple" or that it "works". If anything, the history of the Macintosh platform suggests that it has some serious problems that keep it from catching on on the desktop.
Dude, don't care.
Dude, and I do care when Apple's unsubstantiated marketing claims get parrotted uncritically and start affecting community efforts like Linux.