You definitely have a point. 1 GHz is probably already beyond the point where the higher clock rate is actually often useless because the CPU wastes its cycles waiting for data to be fetched from the memory. Of course, the 1:1 clocked cache that AMD is planning to include in the new Athlons is going to help, but it's not a fundamental cure.
What we need more than faster CPUS are indeed faster memories and faster interfaces. The problem is that it's probably not possible to get those without a fundamental architectural redesign, i.e. we'd have to abandon downwards compatibility of the hardware almost completely, and that not something people will like...
It will probably work like Unix hardlinks, so the links can't transcend partitions. And a copy of a file on the same partition is, by no stretch of imagination, a "backup", since a HD failure usually frags the whole disk.
...typically Microsoft. It may seem a joke to intelligent people, but it's the Truth to Microsoft advocats.
BTW, that's not symbolic links they re-invented, it's more similar to hardlinks. Though there's actually something interesting in there: the "new technology" includes a program that automatically finds identical files and creates links instead. Not that you couldn't do that with a shell script on Unix, and there probably are such scripts available somewhere, but it's not standard procedure to use such a thing.
This USB case is a prime example. Once again, the cheapskates in the Linux community begrudge someone making a profit. Again, they seem to expect a "free lunch".
What they want is not the result of anyone's productive work, but specifications, which are something needed for productive work.
So in a nutshell, open source is exactly like closed source (from the developers point of view), except that the developer works for nothing.
He works because he likes doing it. Besides he gets to use the result of his work as well as that of other open source developers. BTW, some open source developers are actually being paid by companies who have realized that one can make money with open source.
In my book, working for nothing is exactly the same as slavery. For this reason alone open source must be stopped, by Federal Law if need be.
OK, I think you've just stooped into the cesspits of trolldom. Someone moderate this moron down.
You're confusing a particular problem (signal analysis) with the machine being used to solve it. The loss of speed depends highly on the problem. What Seti@home does, as well as brute force encryption cracking and raytracing, to mention some examples, happens to be a problem that allows the nodes to do their work with nearly no communication.
This is not the case with other problems. With those, you need high-bandwidth, low-latency communication between the nodes, and that is something you simply don't get out of Ethernet, which, actually does have exponential speed losses when the network comes close to saturation.
This is what we call a "Milchmädchenrechnung" (milk maids computation) in German.
The point is this: with distributed computing, it is totally ridiculous to just add up the clock speeds of individual processors and expect the result to be in any way meaningful. The reason: When attempting to solve the kind of problems the $30,000,000 system is designed to solve your 170THz would disappear down the idle loop because each processor would spend 99,9% of its time waiting for data to be fetched over the woefully-underpowered ethernet LAN.
First of all, learn to turn bold off when posting in html once you're done with the boldfaced segment;)
No, I should learn to use that preview feature.
Secondly, is the server only solving 1 differential equation? If so, what is this equation, "the universe works like.. ? If there are 10,000 equations being solved, they can be panned out to 10,000 nodes ("computers" if you're going to complain about my arbitrary use of the term "node"). Third, there is nothing about a differential equation that cannot be paralellized.
To be honest, I myself know very little about differential equations. It's definitely not one equation, but interdependent systems of equations. Anyway, someone whos writing his master thesis on the stuff told me that hes doing simulations (of some sort of circuity) that can't be parallelized.
P.S. : "no real compiler support" - what are you talking about?
Ideally, you have a compiler that takes care of using all the nodes and distributing the code. If you have to hand-code all that, it just takes too long and is error-prone (debugging distributed code is a really/i> ugly task). Something like High Performance Fortran.
P.P.S. : Is your "big box" costing more than 20% more than an equivalent beowulf? I'd be willing to bet it costs more than 10 times more. The point is that there is no "equivalent" Beowulf! It's designed to solve problems that a cluster of workstations simply isn't fit for.
Thus, your 20% number (which seems kind of odd to me in the first place, I suspect poor code that wasn't designed to be distributed in the first place)
No, it's code that simply cannot parallelized with 100% efficiency.
And your question about the cost seem like you didn't understand what I wanted to say. These supercomputers spend 80% of their CPU cycles doing nothing because they work on problems which can't be parallelized perfectly. With a Beowulf cluster, the percentage would be higher, and it would keep increasing with the size of the cluster, so that it simply is not possible to build a Beowulf cluster that would solve the same problem faster.
Sure, there are applications like that. But my point was: there are also many problems for which a Beowulf is not suited. If there werent, people wouldnt shell out millions of $$ for real supercomputers (ours is going to cost about $30,000,000). But they do.
Name 1 - 1 task - that requires a supercomputer that can't be broken down into nodes well.
Did you even read my posting? I said that supercomputers are also just a collection of nodes. The difference is in the infrastructure.
I, scanning my mind, cannot come up with a single task that cannot be implemented in a way that will lose almost no speed when made as a beowulf.
The you have no knowledge whatsoever of the kind of work being done on real supercomputers. Which is mostly simulations, i.e. solving differential equations. When I talked to the guy at the supercomputing center who showed us around, he said that on average, they managed to use twenty percent of their computers' potential computing power because the problems don't scale well to many nodes. This would be considerably lower on a Beowulf due to its slower network and no real compiler support.
There are alos tasks (again simulations) that cannot be parallelized at all.
There is a big difference between a cluster of workstations and real supercomputers. Sure, they're both theoretically just nodes connected by a network, but the details are very different. Especially the networking of the nodes in supercomputers is about 2 orders of magnitude faster than even Gigabit Ethernet. Plus, you have custom compilers for that particular machine and its topology.
I'm with the Technical University Munich, and the Leibniz Supercomputing Center next door is getting a new Big Box in March, which will then be the most powerful computer in Europe. The peak transfer rate between its nosed is 10 GIGABytes per second, IIRC. At the moment, thay're still installing the cooling units (the thing will consume about 600 Kilowatts!).
He did not make one single point at all, he just harped about how great microsoft supposedly is. Thats an opinion, not a point.
I moderated it as flamebait because if he really didn't care or know about Linux, he wouldn't post in this forum and this particular thread in the first place.
To conclude. I know (from market research) that most people who use so-called alternative browser platforms also have access to the IE platform,
Maybe, but do they use IE? Almost 50% of the WWW populace uses Netscape for their daily browsing, and those people are not going to be happy about you telling them that they have to switch browsers to watch your page.
It may be somewhat sound business logic to only really care about IE and Netscape, but only IE is plain stupid.
Slashdot Mirror
The point is that some games practically require you to acquire blisters to win, even when playing only for a short time.
What we need more than faster CPUS are indeed faster memories and faster interfaces. The problem is that it's probably not possible to get those without a fundamental architectural redesign, i.e. we'd have to abandon downwards compatibility of the hardware almost completely, and that not something people will like...
As I established in an earlier discussion here: marketing is the science of lying and cheating...
It will probably work like Unix hardlinks, so the links can't transcend partitions. And a copy of a file on the same partition is, by no stretch of imagination, a "backup", since a HD failure usually frags the whole disk.
BTW, that's not symbolic links they re-invented, it's more similar to hardlinks. Though there's actually something interesting in there: the "new technology" includes a program that automatically finds identical files and creates links instead. Not that you couldn't do that with a shell script on Unix, and there probably are such scripts available somewhere, but it's not standard procedure to use such a thing.
Where's the big news in this? We've seen almost identical products before.
What they want is not the result of anyone's productive work, but specifications, which are something needed for productive work.
So in a nutshell, open source is exactly like closed source (from the developers point of view), except that the developer works for nothing.
He works because he likes doing it. Besides he gets to use the result of his work as well as that of other open source developers. BTW, some open source developers are actually being paid by companies who have realized that one can make money with open source.
In my book, working for nothing is exactly the same as slavery. For this reason alone open source must be stopped, by Federal Law if need be.
OK, I think you've just stooped into the cesspits of trolldom. Someone moderate this moron down.
This is not the case with other problems. With those, you need high-bandwidth, low-latency communication between the nodes, and that is something you simply don't get out of Ethernet, which, actually does have exponential speed losses when the network comes close to saturation.
The point is this: with distributed computing, it is totally ridiculous to just add up the clock speeds of individual processors and expect the result to be in any way meaningful. The reason: When attempting to solve the kind of problems the $30,000,000 system is designed to solve your 170THz would disappear down the idle loop because each processor would spend 99,9% of its time waiting for data to be fetched over the woefully-underpowered ethernet LAN.
No, I should learn to use that preview feature.
Secondly, is the server only solving 1 differential equation? If so, what is this equation, "the universe works like.. ? If there are 10,000 equations being solved, they can be panned out to 10,000 nodes ("computers" if you're going to complain about my arbitrary use of the term "node"). Third, there is nothing about a differential equation that cannot be paralellized.
To be honest, I myself know very little about differential equations. It's definitely not one equation, but interdependent systems of equations. Anyway, someone whos writing his master thesis on the stuff told me that hes doing simulations (of some sort of circuity) that can't be parallelized.
P.S. : "no real compiler support" - what are you talking about?
Ideally, you have a compiler that takes care of using all the nodes and distributing the code. If you have to hand-code all that, it just takes too long and is error-prone (debugging distributed code is a really/i> ugly task). Something like High Performance Fortran.
P.P.S. : Is your "big box" costing more than 20% more than an equivalent beowulf? I'd be willing to bet it costs more than 10 times more. The point is that there is no "equivalent" Beowulf! It's designed to solve problems that a cluster of workstations simply isn't fit for.
Thus, your 20% number (which seems kind of odd to me in the first place, I suspect poor code that wasn't designed to be distributed in the first place)
No, it's code that simply cannot parallelized with 100% efficiency.
And your question about the cost seem like you didn't understand what I wanted to say. These supercomputers spend 80% of their CPU cycles doing nothing because they work on problems which can't be parallelized perfectly. With a Beowulf cluster, the percentage would be higher, and it would keep increasing with the size of the cluster, so that it simply is not possible to build a Beowulf cluster that would solve the same problem faster.
Sure, there are applications like that. But my point was: there are also many problems for which a Beowulf is not suited. If there werent, people wouldnt shell out millions of $$ for real supercomputers (ours is going to cost about $30,000,000). But they do.
Did you even read my posting? I said that supercomputers are also just a collection of nodes. The difference is in the infrastructure.
I, scanning my mind, cannot come up with a single task that cannot be implemented in a way that will lose almost no speed when made as a beowulf.
The you have no knowledge whatsoever of the kind of work being done on real supercomputers. Which is mostly simulations, i.e. solving differential equations. When I talked to the guy at the supercomputing center who showed us around, he said that on average, they managed to use twenty percent of their computers' potential computing power because the problems don't scale well to many nodes. This would be considerably lower on a Beowulf due to its slower network and no real compiler support.
There are alos tasks (again simulations) that cannot be parallelized at all.
I'm with the Technical University Munich, and the Leibniz Supercomputing Center next door is getting a new Big Box in March, which will then be the most powerful computer in Europe. The peak transfer rate between its nosed is 10 GIGABytes per second, IIRC. At the moment, thay're still installing the cooling units (the thing will consume about 600 Kilowatts!).
I moderated it as flamebait because if he really didn't care or know about Linux, he wouldn't post in this forum and this particular thread in the first place.
But what's the point of making it a "standard"?
Sadly, these two are in fact one and the same. Check it out at the /. Hall of Fame.
My heartfelt congratulations! May the "imminent death" prophecies never come true...
So whe define "marketing" as the science of lying and cheating... yes, makes a lot of sense.
The other postings in this thread are effectively disproving your claim.
They represent the pinnacle of Marketing, they have done everything right.
That's certainly the case. Unfortunately, it's also the only thing they've really done right
Admittedly they have written some great software
Probably true, but the majority of their products is crap.
Well done Microsoft, you are like a shining beacon to Marketers everywhere, of the potential that the science of Marketing has to improve our lives.
You've got to be kidding... Microsoft is a shameful example of how lying and cheating can get people to choose inferior products.
And how is that improving people's lives??
Maybe, but do they use IE? Almost 50% of the WWW populace uses Netscape for their daily browsing, and those people are not going to be happy about you telling them that they have to switch browsers to watch your page.
It may be somewhat sound business logic to only really care about IE and Netscape, but only IE is plain stupid.
Mark my words, we will see another such project in the near future, claiming to build a desktop that "your grandmother can use"!
Why am I not surprised. Those iMac and iBook designs are practically made for people to whom style is everything ans substance nothing.
The price difference between an 750 MHz Athlon and a 800 MHz one is bigger than $200...
http://www.apple.com/powerboo k/pdf/PowerBook_DS-a.pdf
http://www.apple.com/hotnews/