the Athlon II X2 245 is literally at least four times faster at everything
Depending on what you do, four times faster may be imperceptible.
I have two systems with dual core Atom ITX mobos, one with the 330 chip and another with the 510 chip. The slower board I use as a file server and music player, it has 2 GB memory and 1 TB hd. It's in a small case bolted to the wall behind my desk and runs 24/7. It has no keyboard mouse, or display, I access it through NFS and ssh.
The 510 system has a 1 TB hd and 4 GB RAM. I recently did an upgrade to my desktop system and the new mobo had neither a parallel printer nor an RS-232 serial port. I still have equipment in my electronics hobby shop that uses those interfaces, like my multimeter and PIC development system, so I put together a small computer to use in the shop.
My experience has been that these two systems have all the CPU power I need for the tasks I created them for. I even tested them briefly with XP. although they only have Linux installed now. I found that it's possible to run DirectX 8 games in them but DirectX 9 is too slow. They can display decent video up to 480 lines but it becomes choppy at higher resolutions. This gives an approximate idea of the possibilities of such systems.
For the price they paid for CPU+mobo they could have got a mobo with an Atom CPU soldered in. That socket doesn't come for free and, after all, when was the last time you had a CPU upgrade? By the time you want more performance you will most likely get a whole new system.
For a 2m rig it will die even quicker. For true reliability in the mountains he should go QRP, but then he would need much more knowledge than he seems to be willing to get.
Our universe has a one-dimensional timeline, which means there's a definite beginning to it
All we know is that a significant event we call the "Big Bang" occurred and we have no visibility before that. For all we know, the universe could perfectly well have existed before the big bang, it could have existed forever without ever having been created.
This in no way contradict the laws of physics as we know them. It's like a black hole, we have no visibility beyond the event horizon, but we cannot state that nothing exists inside that horizon.
Myself I prefer to say "I don't believe in an anthropomorphic god".
It's a fact that the universe exists and *something* caused the universe to exist. This something could be the laws of physics or some hitherto unknown mathematical or logical principle. You are welcome to call that principle "god" if you wish.
But to extend that basic principle to some super-accountant being somewhere who's keeping tabs on everything we do and will intervene in our existence if we nag him enough and will punish those of us who don't praise him enough...
The problem with speed perception is that many of the things we do with computers could just as well be done with paper and pencil. CPU speed has always been enough for those tasks, if it's not wasted by bloatware.
I remember when the Pentium 4 came out, most articles about it mentioned that it would be useful mostly for people doing specialized tasks, such as video encoding, for all other users the Pentium III was good enough. What AMD did was to run jobs that didn't have any need for a Pentium 4 and claim the Athlon was just as good.
However, if you did need the extra power, the Athlon sucked in comparison with the Pentium 4 for the simple reason that the Pentium 4 had SSE2. For math intensive applications like video encoding or, in my case, digital signal processing, the Pentium 4 had twice as much math hardware built in than the Athlon. The consequence was that the Pentium 4 had a higher power dissipation without a true benefit to most users, but it had a superior performance in those tasks that actually demand top performance.
Intel was, IMHO, more honest than AMD because they used a true number, raw CPU clock, to rate their CPUs, while AMD created a synthetic number. Is it honest to claim "processing power like a 2.2 GHz chip" if the people who actually need 2.2 GHz will not be the ones who get the best performance from that chip?
I'm a Christian, and am not embarrassed to admit it. I'm embarrassed by these assholes, though. (Atheists often think that Christian == fundamentalist, which simply isn't true.)
I'm not sure it's more logical to say that the universe created itself than it was created by someone, but to each his own, I guess.
Funny how you can contradict yourself in two sentences.
It's always simpler to say the universe created itself than to say something else first created itself and later created the universe.
I consider myself a Christian in the sense that I've read the Bible and believe Jesus taught the right lessons in ethics. But I'm perfectly able to separate the Genesis from Jesus. I refuse to accept a Middle Age transcript of a Bronze Age legend as some kind of fundamental truth in the same way I accept "love thy neighbor" as fundamental truth.
I doubt that an anthropomorphic god such as postulated by the Christian churches exist. i even doubt that the man Jesus was someone who actually lived on earth. Call me an Atheist Christian if you wish.
I believe the New Testament was a compilation of teachings by some Jewish scholars somewhere in Israel two thousand years ago but, no matter where those ideas came from, there's good value in them, if you can interpret them right.
Intel has been in the news often, being accused of various unfair business practices
Yes, but AMD isn't clean either.
I've often been called a "troll" here for stating this simple fact, but AMD invented a certain "megahertz myth" that's a half truth and for a time invested massively in marketing based on that.
I once bought a notebook with an AMD CPU labeled "2200+", which was meant to imply it was faster than an Intel Pentium 4 with a 2200 MHz clock. That could be, for that specific benchmark AMD created, but it was not true for my own applications. For me, that "2200+" actually meant about "1500-".
It's one thing to state that a computer's performance does not depend on CPU clock alone. It's an entirely different thing to create a fake number and pretend that this number is an exact measure of performance.
In the end, the almighty market fixes things up, but not before innocent people waste money. AMD has stopped with that fraudulent practice of inventing fake numbers to pretend having superior performance. It's obvious now that it backfired on them, but many people, me included, bought computers with inferior performance based on those fake numbers.
Why do they mod (+3, Informative) a post with a link proving exactly the opposite of what the post says?
You say "The Amiga was every bit as expandable as the IBM PC and way more open"
Your wikipedia link says "One expansion port for add-ons (memory, SCSI adaptor, etc), electrically and physically identical to the Amiga 500 expansion port (though the Amiga 500's version is inverted)"
Excuse me, but my IBM-PC had seven expansion slots. And, much more important, I could go to any computer store and actually buy cards, both from IBM and from third party vendors, that I could plug into an IBM-PC ISA expansion slot.
It's interesting that IBM thought they had made a mistake in creating such an open architecture and tried to backtrack in the PS/2 version, where the ISA slots had been substituted by MCA slots. The result was that the PS/2 had even less market share than the Amiga. So much for the "magical three letters that sold anything" theory.
That graph completely ignores the number of cores in the CPU.
That was exactly the point. You didn't read the post, did you? I will quote:
"It's obvious that, all other parameters being equal, doubling the CPU clock will double your performance"
Changing the number of cores in the CPU changes other parameters. What I said was that doubling your CPU clock will automatically double performance, assuming the rest of the system can keep up with that increased speed. The real problem is keeping the rest of the system up to the increased CPU speed.
On the other hand, doubling your number of cores will NOT NECESSARILY double your performance. Unless you have an application that has an intrinsic benefit from multiprocessing, AND that application has been designed to use multiprocessing, you will get only a marginal improvement, if you get any improvement at all.
With the speed of todays computers, though, you shouldnt (usually) need that amount of optimization.
After about eight years stuck at the same speed, I wonder what do you mean by "today".
Yes, yes, I know all about the Megahertz Myth. But that's the marketese speech for evading the issue. The fact is that the "Megahertz Myth" is a myth. It's obvious that, all other parameters being equal, doubling the CPU clock will double your performance. It was only when some companies couldn't keep increasing the CPU clock that they invented this lie that megahertz don't matter.
The problem with increasing performance by other ways than raw CPU clock is that it will not necessarily translate in performance increases for all applications. Apple and AMD had to carefully construct their benchmarks to justify this pseudo-myth.
It's true that, with the CPUs we have had for a decade now, many tasks can be done easily with interpreted languages that weren't possible before. That may be true for many corporate desktop applications, no one will care too much if it takes one or two seconds to get a list of how many "snibbo" widgets were returned for maintenance last month.
However, there are still many applications where CPU performance still falls short of the ideal. Improving those applications demands a language like C, because with "higher level" languages it's harder to get increased performance from parallelization. It's not always obvious which operations can be done in parallel and unless interpreters become a lot more efficient at this, there will always be room for compiled languages at the top.
I must say I rarely find a comment on/. that I agree as much as I do with yours.
C and Python march hand in hand, one is for machine performance, the other is for programmer performance. If someone thinks C is too complex or too hard to learn then he shouldn't be working with programming computers, he's likely to cause great damage sooner or later.
However, there's one point where C will need a new approach: multiprocessing is coming. Since it seems like Moore's law has hit the ceiling at 3 GHz CPU speeds, all progress in performance for the foreseeable future will come from increasing the number of CPUs and cores working together.
I have done a lot of programming in multithreads using the pthread library lately and I feel that something better is needed, pthread is not close enough to the metal. I think some new fundamental elements may be needed in the language.
C is so great for programming because it mirrors the hardware closely. For instance, pointers work so well because they represent memory addresses. Before I learned C I had worked with Fortran, I still have some programs I wrote over 25 years ago. Today I look at those old Fortran programs and I wonder why I did some things the way I did. I see some convoluted loops and wonder why I did that because, with a quarter century hindsight on using pointers, I create almost instinctively the most efficient set of pointers to handle a data structure.
What programmers often don't realize is that the correct data structure may get orders of magnitude improvement in performance. To give one example, years ago, when I studied artificial neural networks, I read an article in the Doctor Dobb's magazine (January 1989, page 32, "Neural Networks and Noise Filtering" by Casey Klimasaukas). It was a good article, but the source code in C that came with it sucked. There was a struct _pe representing a processing element and each struct _pe had an array of struct _conn representing the connections to that element.
The problem is that in an artificial neural network what each neuron is doing is, basically, a convolution of two arrays. To do that efficiently in hardware you need to have the array elements contiguous in memory. When you put the connection weight in a structure together with other data you will not have that value contiguous with the weights of the other connections.
From an "object oriented" point of view that program was perfect. But if you want to use your multi-core CPU with that, the program sucks. That's the benefit you can get from programming in C that you won't get with other languages.
And don't tell me that raw performance does not matter because you can always get faster hardware. CPU clock speed has stopped at 3 GHz, we must learn to use our multicores if we want to evolve from now on.
Could be, I have no favorites between the CEO and the web designer, it's indifferent to me how they split the cash among themselves.
The important point is that in most corporations savings are distributed among several items. Part goes to executive bonuses, some of it goes to dividends, and the rest is used to lower prices.
Lower prices mean increased market share and more sales, which means more profit, and more profit means more dividends and fatter bonuses for executives, so it's natural that companies will always try to lower their prices with whatever savings they get.
Just wait until wannabe designers in low-wage nations like India, China, Brazil, etc (using cracked copies of design software) start entering into the process. $269 will seem overpriced.
Your argument about "cracked" software, is a straw man argument, an "established" designer is more likely to use "cracked" software than some designer who is cost-conscious and learned how to use free software.
As for lower prices, that's what they call progress. There was a time when if you needed a book you had to pay a clerk to create if for you. Today you can have it mass-produced for a very small fraction of the price a skilled craftsman would charge.
no American can earn a living doing piecework for rent-a-coder. Most would be better off working at McDonalds
I sense a business opportunity there. If that's true, next time I go to a McDonalds I will take a look at their installations and see if i can come up with some labor-saving invention that I can sell to them.
it boils down to the little people (designers) receiving less money from their corporate clients
Either way somebody wins. The corporate clients paying less means the corporations will have less expenses.
This means they will, pick one or more options:
lower their prices
pay more dividends
raise their employees salaries
Some people think about corporations as some magical entity that drains wealth out of the system and stashes it in some capitalist's swimming pool but that's not how things work. Every penny a corporation saves ends up in someone's hand. It might not be someone you like, but the money does not disappear, it will end somewhere.
Myself, if given the option between getting a lower price at the expense of seeing a somewhat crappier website, or paying more for the same product, only with an award winning company website, I'd rather pay the lower price.
I could stick my hand in a blender and have my fingers it chopped off. There are no excuses for this defect.
All blenders I know of have caps which should be in place before you turn them on. I have even seen one that has a safety switch making it impossible to be turned on without the cap.
None of the phones cited were designed to be stuck in your bush.
I have a friend who worked in the emergency department in a hospital and she told me there had been cases of women who came in with phones stuck in there.
Apparently the "vibra phone" feature has more uses than one imagines at first.
There have been several totally bungled operations in other countries too.
Sadly, the conclusion must be that the terrorists are winning. They aimed to destroy the western way of life and they are certainly making progress at it.
I work at a commercial communications satellite company and we have an old earth station with a 32 meter antenna that's rarely used today, but we still keep it as a backup. Actually, the cost of bringing it down is more than the scrap value, so it's mostly just standing there.
What nobody realized was that the antenna had been tracking a single geostationary satellite for decades, so it was moving very slightly around one position. Geostationary satellites aren't exactly stationary, but close, there's a slight movement around a central point.
The result was that, when they tried to point the antenna to a different satellite they found that the circular steel rail had been cold-rolled over the years so each wheel was sitting in a small valley in the rail. The azimuth motor didn't have enough torque to get off that valley and point the antenna to a different position, although there was no problem in tracking a satellite in the old position.
The solution was to jack up the whole antenna, cut off a section of rail and weld a new piece of rail beneath each wheel. The trickiest part was grinding the rail so that the new parts were perfectly aligned.
Slashdot Mirror
Depending on what you do, four times faster may be imperceptible.
I have two systems with dual core Atom ITX mobos, one with the 330 chip and another with the 510 chip. The slower board I use as a file server and music player, it has 2 GB memory and 1 TB hd. It's in a small case bolted to the wall behind my desk and runs 24/7. It has no keyboard mouse, or display, I access it through NFS and ssh.
The 510 system has a 1 TB hd and 4 GB RAM. I recently did an upgrade to my desktop system and the new mobo had neither a parallel printer nor an RS-232 serial port. I still have equipment in my electronics hobby shop that uses those interfaces, like my multimeter and PIC development system, so I put together a small computer to use in the shop.
My experience has been that these two systems have all the CPU power I need for the tasks I created them for. I even tested them briefly with XP. although they only have Linux installed now. I found that it's possible to run DirectX 8 games in them but DirectX 9 is too slow. They can display decent video up to 480 lines but it becomes choppy at higher resolutions. This gives an approximate idea of the possibilities of such systems.
Pentium 4 had SSE2, with twice the processing power.
For the price they paid for CPU+mobo they could have got a mobo with an Atom CPU soldered in. That socket doesn't come for free and, after all, when was the last time you had a CPU upgrade? By the time you want more performance you will most likely get a whole new system.
For a 2m rig it will die even quicker. For true reliability in the mountains he should go QRP, but then he would need much more knowledge than he seems to be willing to get.
All we know is that a significant event we call the "Big Bang" occurred and we have no visibility before that. For all we know, the universe could perfectly well have existed before the big bang, it could have existed forever without ever having been created.
This in no way contradict the laws of physics as we know them. It's like a black hole, we have no visibility beyond the event horizon, but we cannot state that nothing exists inside that horizon.
Myself I prefer to say "I don't believe in an anthropomorphic god".
It's a fact that the universe exists and *something* caused the universe to exist. This something could be the laws of physics or some hitherto unknown mathematical or logical principle. You are welcome to call that principle "god" if you wish.
But to extend that basic principle to some super-accountant being somewhere who's keeping tabs on everything we do and will intervene in our existence if we nag him enough and will punish those of us who don't praise him enough...
Because a system that allows people to freely express their opinions is less likely to spawn people who impose their own opinions on others.
The problem with speed perception is that many of the things we do with computers could just as well be done with paper and pencil. CPU speed has always been enough for those tasks, if it's not wasted by bloatware.
I remember when the Pentium 4 came out, most articles about it mentioned that it would be useful mostly for people doing specialized tasks, such as video encoding, for all other users the Pentium III was good enough. What AMD did was to run jobs that didn't have any need for a Pentium 4 and claim the Athlon was just as good.
However, if you did need the extra power, the Athlon sucked in comparison with the Pentium 4 for the simple reason that the Pentium 4 had SSE2. For math intensive applications like video encoding or, in my case, digital signal processing, the Pentium 4 had twice as much math hardware built in than the Athlon. The consequence was that the Pentium 4 had a higher power dissipation without a true benefit to most users, but it had a superior performance in those tasks that actually demand top performance.
Intel was, IMHO, more honest than AMD because they used a true number, raw CPU clock, to rate their CPUs, while AMD created a synthetic number. Is it honest to claim "processing power like a 2.2 GHz chip" if the people who actually need 2.2 GHz will not be the ones who get the best performance from that chip?
Funny how you can contradict yourself in two sentences.
It's always simpler to say the universe created itself than to say something else first created itself and later created the universe.
I consider myself a Christian in the sense that I've read the Bible and believe Jesus taught the right lessons in ethics. But I'm perfectly able to separate the Genesis from Jesus. I refuse to accept a Middle Age transcript of a Bronze Age legend as some kind of fundamental truth in the same way I accept "love thy neighbor" as fundamental truth.
I doubt that an anthropomorphic god such as postulated by the Christian churches exist. i even doubt that the man Jesus was someone who actually lived on earth. Call me an Atheist Christian if you wish.
I believe the New Testament was a compilation of teachings by some Jewish scholars somewhere in Israel two thousand years ago but, no matter where those ideas came from, there's good value in them, if you can interpret them right.
Yes, but AMD isn't clean either.
I've often been called a "troll" here for stating this simple fact, but AMD invented a certain "megahertz myth" that's a half truth and for a time invested massively in marketing based on that.
I once bought a notebook with an AMD CPU labeled "2200+", which was meant to imply it was faster than an Intel Pentium 4 with a 2200 MHz clock. That could be, for that specific benchmark AMD created, but it was not true for my own applications. For me, that "2200+" actually meant about "1500-".
It's one thing to state that a computer's performance does not depend on CPU clock alone. It's an entirely different thing to create a fake number and pretend that this number is an exact measure of performance.
In the end, the almighty market fixes things up, but not before innocent people waste money. AMD has stopped with that fraudulent practice of inventing fake numbers to pretend having superior performance. It's obvious now that it backfired on them, but many people, me included, bought computers with inferior performance based on those fake numbers.
Why do they mod (+3, Informative) a post with a link proving exactly the opposite of what the post says?
You say "The Amiga was every bit as expandable as the IBM PC and way more open"
Your wikipedia link says "One expansion port for add-ons (memory, SCSI adaptor, etc), electrically and physically identical to the Amiga 500 expansion port (though the Amiga 500's version is inverted)"
Excuse me, but my IBM-PC had seven expansion slots. And, much more important, I could go to any computer store and actually buy cards, both from IBM and from third party vendors, that I could plug into an IBM-PC ISA expansion slot.
It's interesting that IBM thought they had made a mistake in creating such an open architecture and tried to backtrack in the PS/2 version, where the ISA slots had been substituted by MCA slots. The result was that the PS/2 had even less market share than the Amiga. So much for the "magical three letters that sold anything" theory.
Wrong. At least not in recent decades. Increasing voltage to get higher speed was done in CMOS chips about 30 years ago, but that's long gone.
The era of 5 volts digital circuits is dead, current CPUs work at 3.3 volts and DDR memories work at 1.5 volts or less.
That was exactly the point. You didn't read the post, did you? I will quote:
"It's obvious that, all other parameters being equal, doubling the CPU clock will double your performance"
Changing the number of cores in the CPU changes other parameters. What I said was that doubling your CPU clock will automatically double performance, assuming the rest of the system can keep up with that increased speed. The real problem is keeping the rest of the system up to the increased CPU speed.
On the other hand, doubling your number of cores will NOT NECESSARILY double your performance. Unless you have an application that has an intrinsic benefit from multiprocessing, AND that application has been designed to use multiprocessing, you will get only a marginal improvement, if you get any improvement at all.
After about eight years stuck at the same speed, I wonder what do you mean by "today".
Yes, yes, I know all about the Megahertz Myth. But that's the marketese speech for evading the issue. The fact is that the "Megahertz Myth" is a myth. It's obvious that, all other parameters being equal, doubling the CPU clock will double your performance. It was only when some companies couldn't keep increasing the CPU clock that they invented this lie that megahertz don't matter.
The problem with increasing performance by other ways than raw CPU clock is that it will not necessarily translate in performance increases for all applications. Apple and AMD had to carefully construct their benchmarks to justify this pseudo-myth.
It's true that, with the CPUs we have had for a decade now, many tasks can be done easily with interpreted languages that weren't possible before. That may be true for many corporate desktop applications, no one will care too much if it takes one or two seconds to get a list of how many "snibbo" widgets were returned for maintenance last month.
However, there are still many applications where CPU performance still falls short of the ideal. Improving those applications demands a language like C, because with "higher level" languages it's harder to get increased performance from parallelization. It's not always obvious which operations can be done in parallel and unless interpreters become a lot more efficient at this, there will always be room for compiled languages at the top.
I must say I rarely find a comment on /. that I agree as much as I do with yours.
C and Python march hand in hand, one is for machine performance, the other is for programmer performance. If someone thinks C is too complex or too hard to learn then he shouldn't be working with programming computers, he's likely to cause great damage sooner or later.
However, there's one point where C will need a new approach: multiprocessing is coming. Since it seems like Moore's law has hit the ceiling at 3 GHz CPU speeds, all progress in performance for the foreseeable future will come from increasing the number of CPUs and cores working together.
I have done a lot of programming in multithreads using the pthread library lately and I feel that something better is needed, pthread is not close enough to the metal. I think some new fundamental elements may be needed in the language.
C is so great for programming because it mirrors the hardware closely. For instance, pointers work so well because they represent memory addresses. Before I learned C I had worked with Fortran, I still have some programs I wrote over 25 years ago. Today I look at those old Fortran programs and I wonder why I did some things the way I did. I see some convoluted loops and wonder why I did that because, with a quarter century hindsight on using pointers, I create almost instinctively the most efficient set of pointers to handle a data structure.
What programmers often don't realize is that the correct data structure may get orders of magnitude improvement in performance. To give one example, years ago, when I studied artificial neural networks, I read an article in the Doctor Dobb's magazine (January 1989, page 32, "Neural Networks and Noise Filtering" by Casey Klimasaukas). It was a good article, but the source code in C that came with it sucked. There was a struct _pe representing a processing element and each struct _pe had an array of struct _conn representing the connections to that element.
The problem is that in an artificial neural network what each neuron is doing is, basically, a convolution of two arrays. To do that efficiently in hardware you need to have the array elements contiguous in memory. When you put the connection weight in a structure together with other data you will not have that value contiguous with the weights of the other connections.
From an "object oriented" point of view that program was perfect. But if you want to use your multi-core CPU with that, the program sucks. That's the benefit you can get from programming in C that you won't get with other languages.
And don't tell me that raw performance does not matter because you can always get faster hardware. CPU clock speed has stopped at 3 GHz, we must learn to use our multicores if we want to evolve from now on.
Could be, I have no favorites between the CEO and the web designer, it's indifferent to me how they split the cash among themselves.
The important point is that in most corporations savings are distributed among several items. Part goes to executive bonuses, some of it goes to dividends, and the rest is used to lower prices.
Lower prices mean increased market share and more sales, which means more profit, and more profit means more dividends and fatter bonuses for executives, so it's natural that companies will always try to lower their prices with whatever savings they get.
Your argument about "cracked" software, is a straw man argument, an "established" designer is more likely to use "cracked" software than some designer who is cost-conscious and learned how to use free software.
As for lower prices, that's what they call progress. There was a time when if you needed a book you had to pay a clerk to create if for you. Today you can have it mass-produced for a very small fraction of the price a skilled craftsman would charge.
I sense a business opportunity there. If that's true, next time I go to a McDonalds I will take a look at their installations and see if i can come up with some labor-saving invention that I can sell to them.
Either way somebody wins. The corporate clients paying less means the corporations will have less expenses.
This means they will, pick one or more options:
Some people think about corporations as some magical entity that drains wealth out of the system and stashes it in some capitalist's swimming pool but that's not how things work. Every penny a corporation saves ends up in someone's hand. It might not be someone you like, but the money does not disappear, it will end somewhere.
Myself, if given the option between getting a lower price at the expense of seeing a somewhat crappier website, or paying more for the same product, only with an award winning company website, I'd rather pay the lower price.
This reminds me of that old joke about the Nazi staff meeting:
-"Our plan is to kill six million Jews and one Argentinian"
-"Why the Argentinian?"
-"You see, no one cares about six million Jews"
Moral of the story: when you add an unusual element to a set, the main part of the set loses focus.
All blenders I know of have caps which should be in place before you turn them on. I have even seen one that has a safety switch making it impossible to be turned on without the cap.
I have a friend who worked in the emergency department in a hospital and she told me there had been cases of women who came in with phones stuck in there.
Apparently the "vibra phone" feature has more uses than one imagines at first.
Just to mention another interesting detail, this earth station I mentioned has Wernher von Braun's signature in the visitor's book.
What, no more worlds to conquer?
There have been several totally bungled operations in other countries too.
Sadly, the conclusion must be that the terrorists are winning. They aimed to destroy the western way of life and they are certainly making progress at it.
I work at a commercial communications satellite company and we have an old earth station with a 32 meter antenna that's rarely used today, but we still keep it as a backup. Actually, the cost of bringing it down is more than the scrap value, so it's mostly just standing there.
What nobody realized was that the antenna had been tracking a single geostationary satellite for decades, so it was moving very slightly around one position. Geostationary satellites aren't exactly stationary, but close, there's a slight movement around a central point.
The result was that, when they tried to point the antenna to a different satellite they found that the circular steel rail had been cold-rolled over the years so each wheel was sitting in a small valley in the rail. The azimuth motor didn't have enough torque to get off that valley and point the antenna to a different position, although there was no problem in tracking a satellite in the old position.
The solution was to jack up the whole antenna, cut off a section of rail and weld a new piece of rail beneath each wheel. The trickiest part was grinding the rail so that the new parts were perfectly aligned.