Analyst Concerned About Microsoft, Intel on China Weakness

[blazorthon]

[citation][nom]vkg1[/nom]You think too small. Anytime anyone like apple makes something that the sheep love, samsung will make something identical but use economies of scale to make it cheaper. So they will be more successful and always make the best products available. There will not be an oxygen anymore for the old companies. Products like the Galaxy tab will cause people to stop using their PCs. When they stop using their PCs, they will not replace them anymore. When they don't replace them, Intel, MS, Apple will not receive money anymore. Only Samsung, ARM, and asian manufacturers will make money. In the meantime, they will start developing their alternatives to Android and ARM if the free software movement doesn't simply already have something for them. Intel will run out of money. Your whole argument will become irrelevant. They will be on equal playing field. And on equal playing field these old evil companies that appeal to sheep will get destroyed by the new generation of companies that cater to smarter, savier, and less ignorant people.[/citation]

Yet another long post, but I'm game for lecturing politely if I must. I put a summary at the bottom, but if you want more detail you'll have to read the whole post.

Samsung's products are not identical to Apple's. That is a myth and furthermore, Apple has a fan base and not much of a user base. Unlike Apple, most of Samsung's customers tend to be less of a fan base and more of a user base. The difference being that Apple has more loyal customers than regular customers and the opposite is true for Samsung.

Tablets can't replace a desktop for even close to everything. They can't handle heavy tasks and their form factor is simply not suitable for a lot of work that they might even have enough performance for. Beyond that, Intel, MS, and Apple are all in the tablet market or going into it. Samsung isn't going to kill them off and ARM seems to be losing ground to Intel's Medfield quite fast. The only ARM chips in the market that can truly beat even Medfield right now in are Qualcomm's highly customized chips. Medfield is only a 32nm processor with a severely cut-down Core 2 arch. The performance increase from a die shrink and a move to a base in Ivy Bridge could be tremendous despite decreasing power usage. Heck, using the current designs and simply increasing core count would let it meet or beat the fastest ARM processors.

So, we know that Intel isn't going anywhere, especially since they are a huge presence in the business/enterprise markets as well as customer markets. You might think big, but you don't really understand the industries that you're talking about very well.

MS is also a huge presence in the business markets and is also the prime part of the consumer OS markets. Even with free architectures (which I've already explained why they won't make it as free architectures), there would need to be operating systems. This is no easier to make than a CPU architecture. Heck, it might be even more difficult and need even more money. So, who's going to make the operating system for these processors? What will it be? Will it be a port of Linux and if so, who's going to port it? If you go the free route to not pay coders to port it, then it will take a lot longer and chances are that it will be buggier. It's not because they aren't good at this sort of work, but these people would not have incentive to work.

They need money to feed themselves and if they have them, their families, among other needs for money. To make money takes work and work takes time. It would be even worse if making an OS more or less from scratch is attempted. MS also has something that no replacement would have: A huge software base that is also mostly compatible with every new version of Windows for years to come and what isn't will be replaced in due time. A new platform would have nothing and be unable to replace Windows or even OSX because of this no matter how much better it could be (which it assuredly wouldn't be if it had not money backing it anyway). This would, yet again, be a very time-consuming process on top of designing the architecture and getting an OS for it.

So, MS has this advantage and isn't going anywhere. Google's Android is already freely available to people anyway, so what's your complaint with Android about anyway? It's not free to be free either. It takes a lot of money for Google to make and update it. I think I've given enough reasons for why this isn't cheap, but if you ask, I can give more and more. The list goes on.

Then there's Apple... Apple has a huge fan base. These aren't people who will switch even if there are superior products at a fraction of the price. Apple has pretty much secured their future as a long-lasting company so long as they don't get in worse trouble than they can handle with the laws of various countries, granted this can happen and isn't even too unlikely. Apple would need to screw up extremely badly in all or most of their products and at the same time in order to really be hurt and even then, I'm not sure if this is enough to kill Apple off. Fans might just stay with Apple anyway or get back with Apple as soon as the next Apple products come out.

So, chances are that Apple isn't going anywhere either.

You also ignore the fact that many of, if not most of, the other companies (such as various OEM companies) work not only with MS, Google, and more in a variety of ways and deals, especially Asian manufacturers of many products.

There are so many other factors to consider. Guess what, many of the companies here have a huge hand in the other companies that make products or at least design products that are necessary for any computer or such gadget. If MS, Intel, AMD, and more go down, then most of the Asian manufacturers are screwed as well as their customers. Nvidia can't survive without these other companies either and that would be even more devastating for consumers and businesses alike. Any company with a big stake in GPGPU or consumer graphics usage (such as gaming) would be screwed too.

Without all of these companies and groups of companies, there is not much money left. With no money to back free movements such as free CPU architectures and many manufactures pretty much screwed, the entire computing industry could fail. Free simply doesn't work in today's economy. It never has and probably never will with the way things have been going for us as a species. There is no money to develop alternatives to all of these companies and their products. No money means it simply won't happen. The entire computing industries can't run on the free time of a few experts. You might be looking at a big picture, but don't forget... you miss out on the depth on what the picture depicts.The world has at least three dimensions and a picture can't show all of them. Even if it could, it can't do so in enough detail to see everything that goes on. When you widen your view, your view of any particular part of the picture blurs and although you see the whole picture, you don't see how it works.

Asian manufacturers, Samsung, and ARM will make software and such? Are you crazy? Where will they get the billions of dollars to pay for all of that? They can't sell products that most people aren't buying and expect to have enough money to do that. The opensource/freeware movements, granted I do support them, have only come so far. There is no open source photo editing software that can compare to some of the commercial versions nor even come close. There are no open source office suites that have every feature that MS Office has, granted they are doing better than the photo editing side of this. There are no open source audio editing tools that can compare to commercial tools.

The open source movements have come far, but not nearly far enough. Then there's the fact that they have yet to even get a tenth of MS's market. Even combining Linux and OSX's market shares, MS still has almost ten times more OS market share on X86 consumer computers. Take Apple out of that and MS is wining by almost 20-25 times over compared to Linux in the consumer market. Sure, Linux wins in super computing and that is a huge victory, but that's not nearly enough to kill off MS and isn't even related to the consumer markets. Also, the chances of anyone making an OS, although rather slim, aren't even nearly as bad as the chance of it being as good as current OSes and their software support. Heck, take a look at Windows Phone 7. It's been around for a while now and has thousands of apps. The problem? It's still a small fraction of Android and iOS's app counts and is growing slower too. Imagine how much worse that would be going if it didn't have MS pouring huge amounts of money into it and all of the paid devs left.

Also, this is the short version. There are far more problems with what you propose that I don't have the space here to write about and you probably wouldn't have the patience to read.


SUMMARY:
There is no such thing as a free architecture. It takes a huge amount of money to make a CPU architecture, especially one that can compete with current architectures. Look at Intel. This is a huge juggernaut of a company that can handle pretty much anything that you throw it's way quickly and effectively if it wants to. Intel isn't just a part of the computer industry, Intel is a backbone of the computer industry and all because they can afford to be. Most of the professional/business/enterprise markets rely on Intel's processors because they are simply faster for their work than pretty much any and all competition while being the most energy efficient.

A free architecture also needs an OS. This can be even more expensive than making the architecture itself. It can't be just a mere phone/tablet OS if it is to replace Windows and OSX. It also needs software and a lot of it. This would take even longer than the OS and architecture would simply because of the huge amount of work that it would need.

Apple has a huge fan base and more money on-hand than perhaps any other tech company in the world. Fans don't care if it isn't the best because they either are too ignorant to know that it is not nearly the best, they just don't care, they already have Apple products and that can make it more difficult to switch, or even because Apple is simply the cool company to buy from. Then there are situations where Apple can have advantages every now and then anyway.

A free architecture needs fabs and it needs other hardware. RAM, storage, graphics, and so much more. Much of the current hardware would be compatible because most of it is based on proprietary technologies and/or interfaces. Basically, you'd be reinventing the wheel in every aspect of the freely available machine type that you demand and every step is more expensive and time-consuming than the last. You can't just drop a new CPU into existing hardware. You most certainly can't expect the several hundred billion dollar computing industry to just let something like this happen if it was going to happen either.

Software and OS support. This would be non-existent on a new CPU architecture. Even porting Linux over wouldn't be an easy task and it would take a lot of time. However, it would not solve the problem of program support which is by far the most important aspect of a computing device of any kind. No matter what, if it can't run what you need it to run, then it is next to useless for that purpose. Porting Linux over won't port Linux programs over and even if it did, they aren't enough.

 

[vkg1]

You sound like a typical american who has no sense of how much the average person in the rest of the world wants to use products that are not american or from these companies. Your perspective reflects an attitude that only exists in the US.

Additionally, a couple years from now there will basically be nothing a notebook can do that a tablet can't. This is already largely the case as the Office monopoly has been broken everywhere except for in the US.

Samsung, or some group of companies if they get sick of Google Android, or China, some EU entity such as DLR, or people within the US working on free software can all make Linux just as good as Windows on any architecture. Even Google looks like they are willing to do this for us with Chrome.

ARM CPUs may only be 1/10th the power of Intel CPU today, but they need much less energy so fit in any device and are doubling in power every year. In other words, even the most fearsome Intel workstation CPU today will still be slow compared to a battery-powered ARM in a few years. But 99% of people will still only need the power of one of today's ARMs for their Facebook, Youtube, and Google Documents creation.

The fates of these companies are already sealed. Google gets the approach and tries to do it themselves. But they'll be eaten alive too.

 

[blazorthon]

[citation][nom]vkg1[/nom]You sound like a typical american who has no sense of how much the average person in the rest of the world wants to use products that are not american or from these companies. Your perspective reflects an attitude that only exists in the US.Additionally, a couple years from now there will basically be nothing a notebook can do that a tablet can't. This is already largely the case as the Office monopoly has been broken everywhere except for in the US.Samsung, or some group of companies if they get sick of Google Android, or China, some EU entity such as DLR, or people within the US working on free software can all make Linux just as good as Windows on any architecture. Even Google looks like they are willing to do this for us with Chrome.ARM CPUs may only be 1/10th the power of Intel CPU today, but they need much less energy so fit in any device and are doubling in power every year. In other words, even the most fearsome Intel workstation CPU today will still be slow compared to a battery-powered ARM in a few years. But 99% of people will still only need the power of one of today's ARMs for their Facebook, Youtube, and Google Documents creation.The fates of these companies are already sealed. Google gets the approach and tries to do it themselves. But they'll be eaten alive too.[/citation]

Just about everything that you said here is wrong. Most, if not all, of Intel's current notebook and desktop CPUs (excluding very outdated and/or extremely low end models, not actual modern CPUs) aren't just ten times faster than an ARM, but several dozen times faster or even more, notebooks will probably always be able to do a huge variety of things that a tablet can't, Android is already a derivative of Linux and porting any Linux to another architecture still leaves it without software support and isn't easy either, AMD CPUs aren't 1/10th the power of Intel CPUs and the math is never that simple in CPU technology, Far more than a mere 1% of the people whom use computers use them for more than just what ARM CPUs can handle, and these companies don't have their "fates sealed". I'm not some American idiot who doesn't understand other people's perspectives. I am telling you how it is and if you don't like that, well, there's probably nothing that you nor I can do about this.

Tablets have very small fractions of the computing power of even some of the slowest modern notebooks. Anything that is even remotely intensive can't be done on a tablet properly, at least not without several times more time, let alone compared to high end notebooks. Go ahead and try a real modern game on any tablet or smart phone. Oh wait, they can't even handle games like this from several years ago, so it can't even be done. Heck, games aren't even nearly the most intensive things that hundreds of millions of people do on computers. Are they the majority? Not really. However, they are a huge fraction and even many more average users do a lot of stuff that tablets can't do nearly as well as notebooks and especially desktops.

You don't really think that a several watt device can keep up with a two to several dozen watt notebook, let alone several hundred watt to over a thousand watt desktops? It simply can't be done with technology of the same or even similar generations. Not even a top-end Krait can beat a mere dual-core Atom processor in a decent netbook, nor even come close to the still very, very low end E-350 netbook APU and up. Most of the ARM processors can't even come close to the top-end Krait processors. Keep in mind that both the Atom and the E APUs are built on quite old technology and were the weakest processors for their time. ARM CPUs can not and never will be able to match CPUs of even older generations simply because they can't match the power consumption. A dual-core, Hyper-Threaded Sandy Bridge i3 can churn out incredibly more performance than any ARM CPU can simply because it can use more power.

Tiny little processors with their tiny little graphics and up to a few watt TDPs have no chance of beating something like an i3 that can use ten-twenty times more power and deliver similarly higher performance. ARM would need to be scaled up, but it is not an architecture that scales well at all. It actually scales very poorly if it is built up to try to compete with higher power processors such as the i3s. ARM can only go so much farther before it can't scale well enough at all at this rate, whereas X86 and such still have a ton of headroom for further scaling.

There is no chance of a battery powered ARM CPU ever beating full workstations of the same age. It can't be done, plain and simple, when both use similar process technology. A six core SB-E processor such as the i7-3930K, an excellent choice for a workstation, is yet another three to four times faster than the i3 from the previous example in integer workloads. You obviously know next to nothing about processors if what you've said is to be considered, so I'll lecture about that a little.

ARM uses a RISC instruction set architecture, aka Reduced Instruction Set Computing. This means that it does not support advanced instructions like X86 processors do. Beyond that, it has even more inferior FP instruction support. This is even more crippling for a workstation replacement. ARM is literally designed to use as little power as possible, but in doing so, it sacrifices pretty much everything that it would need in order to compete in the high-performance computing. It can't beat a workstation processor even if it has twice the power and size available to it simply because raw performance is meaningless without support for the proper code and it can't even scale up that way. Furthermore, your assumption here relies on Intel not improving either and this is simply not the case. Intel has no intention of just stopping their improvements for a decade or two just to let ARM catch up to their slowest processors, if ARM can even be scaled up that far without going highly parallel in cores and such and in that case, any software that can't be widely parallelized would still perform like crap on ARM.

Any free architecture to replace ARM would, again, need a huge monetary backing for R&D. Samsung wouldn't do this, no manufacturing company would do this, and although ARM might try, it would be a far greater undertaking than their current architectures have been and I don't think that they'd succeed. It took AMD and Intel decades to get where they are. ARM and such won't manage this any faster than they did and by the time they reach current workstations in performance, Intel and probably AMD too would be decades ahead of ARM yet again.

You can try to argue with me all that you want, but I know far more about what would be needed for what you think will happen to happen and far more about the technologies themselves. This isn't going anywhere. You simply fail to realize how the concept of freely available electronics design is not sustainable. There needs to be far too much money and no one will invest so much in something that won't pay off in return. People and companies simply don't throw out billions of dollars into R&D just so that their projects are freely available for any other company to use as they wish.

 

[vkg1]

Why can't Samsung together with about 10 other companies who are sick of Intel $500 processors simply get together and make the next Alpha or SPARC or Power or any of the many others that were created by much smaller entities? Or China, or Russia, or S Korea, or EU, etc as a research initiative that ends up commercializing? They can. (If in fact it's really the case that all current architectures other than the ancient x86 could never scale adequately). I'd go so far as to say that it's inevitable because there are many entities who are much much hungrier and are fundamentally underestimated by people like yourself. And of course, we saw Intel Itanic and PIV were huge disasters showing how fallible Intel can be. It's only a matter of time. History will repeat, and next time there will be entity less starved and incapable to taking advantage of situation than AMD was. In particular, a company or group of companies like samsung and similars.

 

[blazorthon]

[citation][nom]vkg1[/nom]Why can't Samsung together with about 10 other companies who are sick of Intel $500 processors simply get together and make the next Alpha or SPARC or Power or any of the many others that were created by much smaller entities? Or China, or Russia, or S Korea, or EU, etc as a research initiative that ends up commercializing? They can. (If in fact it's really the case that all current architectures other than the ancient x86 could never scale adequately). I'd go so far as to say that it's inevitable because there are many entities who are much much hungrier and are fundamentally underestimated by people like yourself. And of course, we saw Intel Itanic and PIV were huge disasters showing how fallible Intel can be. It's only a matter of time. History will repeat, and next time there will be entity less starved and incapable to taking advantage of situation than AMD was. In particular, a company or group of companies like samsung and similars.[/citation]

I never said that they can't... I only said that they won't and gave reasons for why they won't. The problem that you don't realize is that these are not markets that Samsung and these other companies have experience in, among other problems. Furthermore, you said a new free architecture, not an old one, and in that sense, it would be exceptionally difficult and even less likely. Furthermore, companies aren't sick of Intel or else they would have done something years ago when it was far easier. You don't realize that most of Intel's consumer processors are not $500 and there are even far cheaper than $500 professional, business, and enterprise processors from Intel. Itanium hasn't done stupendously well, but not only is it still going, but with HP, it's making Intel quite a lot of money.

Furthermore, I never said that only x86 can scale effectively. I said that ARM can't because it can't, but I didn't say anything about SPARC and PowerPC. SPARC has been more of a parallel oriented processor that isn't suited to many common tasks and it might not scale as effectively as some other architectures can, but PowerPC seems rather capable. However, you fail to realize that much smaller companies aren't making much of these. Who designs processors based on these architectures, or really any architectures? It's almost purely large companies who do this, especially who do this well.

Intel most certainly won't repeat history with Itanium. Itanium is a completely different architecture in many wys from x86 and even though it is not the best, like I said, it is still going. New Itanium processors are due to be released this or next year.

Pentium 4 wasn't a failure. It sold extremely well. It had its problems and was nowhere near as good as Athlon 64 that came out a while after P4s did, but it still sold extremely well. Intel won't do that again either. History might repeat itself in your mind, but Intel knows exactly what went wrong with P4 and isn't stupid. Intel won't try that again. Regardless, it wasn't a disaster at all considering how well it sold. AMD, like I said, had far faster CPUs at the time, so they obviously weren't incapable. You know why they couldn't get traction? Simple, really. Intel had more money and used that money to get their way. You think that a bunch of companies with zero experience with real notebook/desktop processors will manage to usurp the current companies that not only have far more money, operating system support, software support, hardware support with other companies, but also are already ingrained into the market with their huge customer bases? It's not going to happen. AMD has been trying for years and they've got far more advantages than Samsung could ever have, yet they've hardly gotten anywhere in the CPU markets.

Heck, how would Samsung even get other companies on-board considering that many of these companies rely on Intel and other such companies that you want gone for their revenue?

 

[vkg1]

All your thinking is pre-2007. When the world was Intel or AMD. ARM wasn't even a blip on anyone's radar. Now everything that uses Intel cpus is declining in popularity and nothing can stop the sales of Galaxies. You have the mentality of the greatest horse carriage manufacturer in the world 5 years after the first horseless carriage was made. Your keyboards, mice, and 100W CPUs are history. If Intel isn't being run by people who have a very different mentality than you, then the same thing that happened to Nokia will happen to Intel in darn near the same timeframe. Anyway, it will happen sooner or later and I think much much sooner than you realize. But thanks for the argument it was informative.

 

[blazorthon]

[citation][nom]vkg1[/nom]All your thinking is pre-2007. When the world was Intel or AMD. ARM wasn't even a blip on anyone's radar. Now everything that uses Intel cpus is declining in popularity and nothing can stop the sales of Galaxies. You have the mentality of the greatest horse carriage manufacturer in the world 5 years after the first horseless carriage was made. Your keyboards, mice, and 100W CPUs are history. If Intel isn't being run by people who have a very different mentality than you, then the same thing that happened to Nokia will happen to Intel in darn near the same timeframe. Anyway, it will happen sooner or later and I think much much sooner than you realize. But thanks for the argument it was informative.[/citation]

Workstations can't be replaced by tablets within even the next few decades, plain and simple. Gamers can't use tablets in order to play many games from even a few years ago well at all. So, here we have hundreds of millions of people in total who can't use tablets to replace their desktops and notebooks. This doesn't even include business usage and enthusiasts who make up another huge amount of people who can't use tablets to replace their desktops and computers. Also, many desktop CPUs aren't even 100w CPUs. None of Intel's LGA 1155 CPUs have a TDP higher than 95w and many of them are only 65w. AMD, however, has a lot of more power-hungry CPUs, but again, most of their CPUs can range from 65w to 95w with only a fairly large fraction, but still not most, going to 125w, and AMD seems intent on mostly killing off the 125w variants except for higher end systems.

You think my mentality is old? Well, no, it's not. No tablet has even the CPU power of most, if not all, of my older laptops. Keyboards and mice are still the best interface for gaming and most regular work. Touch screens can be good for some stuff, but not nearly as much, especially with things such as Office work. However, the far inferior CPU performance holds them back anyway, so it's not like they could do most heavy work even if their input interface worked better for such work.

If you want to just agree to disagree and end the argument here, then fine. However, what you've proposed is unlikely to an extreme and even worse, if it is attempted, it is even more unlikely that it would succeed.