word size may improve performance : you are correct there.
However, increasing word size doesn't mean you will execute more instructions per cycle or per second.
>so doubling the word size, on the bus will improve performance
Maybe - keep in mind that the size of an instruction is not neccesarily the size of the data retrieved from memory. The cache could access memory 128 bits at a time while instructions are only 32bits in length.
True, it does seem like dark matter is a open invitation for everyone to invent whatever they want when they should really be looking harder for normal stuff.
I guess I was a little hasty with my judgement that they aren't the missing dark matter. They could be part of the problem, a little chunk off of the grand total of missing mass that we have.
right.... all matter has mass and thus gravity. But maybe something as of yet unclassified also generates gravitational pull. That was all I meant.
The dark matter could be due to neutrinos, black holes, WIMPs, planets(non-light emitting matter like you suggested), other universes, or any other number of bizzare things that scientists have hypothesized.
Unfortunetly this could not account for dark matter.
The reason scientists believe that there should be dark matter is because of the fact that the stars on the edge of galaxies move faster than they should. According to the measured amounts of mass in a galaxy, the stars on the edges would fly out of orbit at the speeds they are going.
Extra mass on the outer fringe of a galaxy could not contribute to this lack of gravity. I am pretty sure that more than 1% of the galaxy's mass is missing also. But I suppose this goes to show that we never know as much as we think we do.
Checkout the everything 2 node on dark matter for more information.
Why must we constantly hear about people guessing the future of Moore's Law. Is this even news anymore?
Intel itself has already said that Moore's law is over, explained in slashdot here. Of course, other people are always predicting the end of it as well. Then again, some people think it will continue.
I really wish people would get over Moore's prediction and talk about relevant stuff. There is no way to predict how long unknown scientific breakthroughs will allow Moore's Prediction to remain true. There is one absolute though, the end will come some day, you can only store so many atoms in a certain amount of space according to the rules of quantum physics - that is the absolute barrier.
Until it is actually abandoned I could do without hearing more of Moore's law.
> Think about it - the main problem in terms of pushing computing power these days is electron migration, caused by extremely high clock rates.
Electron Migration? what are you talking about. Processors continue to get faster and faster due to improved processing technology and increased parallelism. Leakage and electromagnetic interference from the clock signal are major problems today but who knows what scientists are working with nowadays.
> By doubling the word length to 64-bits, you can reduce the clock rate of the chip, and will still be able to perform more instructions per second than your top-of-the-range Athlon/Pentiums.
That is absolutely not true. Having 64 bits allows you to access a larger amount of total memory, and it lets you put more information in each instruction. The amount of data you can work on in any given clock cycle is proportional to the cache access and bandwidth and the register size (Neither of which inherently need 64 bit long instructions).
To perform more instructions per second (or instructions per clock cycle) you need instruction level parallelism (ILP). This has been a major goal of processor manufacturers for many years now. Intel had two main ways of trying to increase ILP.
Use an instruction set with inherently more parallelism - allowing you to issue multiple instructions at once - Itanium
Try executing from more than one thread at a time - allowing you to use more of the processor per clock cycle - Hyperthreading (now on Pentium 4 processors).
oh you hadn't noticed the tendency of everyone to blindly dislike Intel regardless of what they do?
Itanium has lots of cool new features that compilers could be using and people could be taking advantage of, but it doesn't have good backwards compatibility, and therein lies the problem.
You know why this got so much press? It was exactly because it was a screw up on Intel Marketing. For some reason Intel didn't have a representative in the right place, and AMD got one deal it shouldn't have. Lets compare all the large servers planning to be built with Itanium with those planning to be built with Hammer and a different picture will emerge.
The agreements revolve around the fact that court cases are long, drawn out, and involve legal fees.
They decide, okay we get use patent X,Y, and Z for $200M and you agree never to sue us for using them again. Intel gets to make processors using the patents, and Integraph gets money for nothin. This way nobody gets dicked by a long drawn out trial to try and explain advanced technical concepts to lay-persons. With a court case, you are guaranteed nothing.
>Intel should thumb their noses at both patent law, and this other company, and just scrap the project.
If you knew the amount of time/effort/money that Intel has and is putting into Itanium, you would realize that this suggestion is preposterous.
As far as paying for a loser, Intel has already dumped out $350M to Integraph, these lawsuits are about ADDITIONAL payments. Intel isn't stupid enough to bury themselves in a project they think will never make enough money to be worth it.
Hammer and Itanium are not in the same market. They may compete a little bit, but neither Intel, nor AMD plans to have them competing with each other. Remember Xeon? its still around despite the lack of press. Itanium is supposed to be very high-end computation based and Hammer is a smaller server/desktop product that will compete with Xeon and Pentium brand processors.
P4's are made all over the world, you are correct. That means not just in China! It is world-wide technology created by an American company. I don't think anyone can quantify how much of the technology comes from "America" specifically, versus other labs and offices around the world, but I still think it qualifies as "*our*".
Also, the latest and greatest fab that Intel just built in Rio Rancho, New Mexico!. This fab features larger 300mm wafers for an important reason that I tried to illustrate before.
The point is: the processor YIELD determines the cost. The cost of labor and environmental codes is not a large factor (Intel factories all follow the same environmental codes anyway!). The larger wafers will let us get more working processors out of the same wafer, thereby increasing yield and letting Intel sell processors cheaper!
It may also be worth pointing out that much of the handling in the new fab is now automated, creating a reduced labor cost.
Explanation - a rough calculation
Say intel can fit 20 Pentium 4's on a 300mm wafer. That wafer costs Intel $1,000 to process using all the cool.13u transistor technology (the price of the wafer isn't important and I just guessed). Little air particles, stress fractures, and errors in processing cause half of the Pentium 4's to have errors and get discarded. This means the Pentium 4's cost $100 ea. to make. Now imagine that they improve the process technology, and there are a tenth of the errors that there used to be. This improvement results in only a quarter of the Pentium 4's on a wafer being damaged. Now we get 15 good processors per die, and they cost Intel only $66.6 to make. Intel just saved a third of the price of making a Pentium 4. Now multiply that savings by the millions of P4's sold and you can see why the yield is so important.
I don't know much about manufacturing hardware, but I know that in the processor realm they cannot make anything cheaper than we can.
Processor prices are limited by yield (aka. how many good processors come out of the factory) which is entirely limited by technology. Their fabrication technology is not better than ours right now. If anything, it is far far behind. There is no way that they can produce chips cheaper than we can.
Of course they could possibly get some non-proprietary hardware cheaper, and make their workers work for less. However chip errors occur because of microscopic particles in the air and slight jolts and bumps from a mechanical handler. These are not things that foreign manufacturers have a better handle on than we do. The number of working dies(the chip before its packaged) per silicon wafer is the most important factor in chip price.
How many times do we (as a people) have to hear about this complaint before people stop voicing it? Well they never will.
When software starts to become usefull people say "Engineering and Math will become obsolete, programs will do all the thinking for people".
And when computers first came out "Computers are going to replace people in all sorts of jobs, soon a computer will be doing YOUR job, better than you do it!".
Anyone remember when Robot's were popular? People were saying that robots will take over all our jobs. Soon robots will do all the thinking, and man will become obsolete.
The industrial revolution, I wasn't there to experience, but anyone suppose there was similar paranoia about large machines? Printing presses? Automated factory machines?
Finally I vaguely remember a quote from history class. It was something along the lines of: "This new invention will be the end of us. Man will no longer have to think for himself, it will all be done for him" and you know what they were complaining about.... THE WRITTEN WORD.
Come on, as long as there is a need for people to know math, and engineering skills (which there always always will be) there will always be engineers. My computer isn't quite perfect yet. I still need to be able to put my stick of memory in, or have some engineer soder a new capacitor in when a chronic blue screen keeps appearing.
Some games get really greedy with their textures. Latest games such as Soldier of Fortune (I think that is the one) will load up so much of the memory, that if you happen to only have 64MB then the hard drive thrashing occurs.
And any good hardware buff knows that hardware thrashing is really bad. Once that happens it doesn't matter if your running NVidia or Voodoo or even integrated graphics, its going to run SLOW. If you want to be able to see all your blood and gore, as well as explosives, and triple that up with realistic surroundings and backgrounds you better go with 128MB.
However, I suppose if you run at the lowest levels, turning off all the special effects to try and get the highest frame rate, then the extra memory won't help. You are turning off all the effects that use the memory anyway. So if your going for plain FPS and don't really care about how it looks you will be okay with 64MB.
Intel currently has wireless networks setup throughout some office buildings. Of course not every employee can be on (too many connections), but it works well for those who need it.
Also, I recently setup a wireless router at home, and got an 802.11b card for my laptop. Now I can browse the internet on the couch in front of the tv, posting dumb comments to/. threads;)
I would just be happy if people would realize that 10Mb/s and 10MB/s are two different things. They can never achieve anywhere near 10 Megabytes a second through their network.
Thats because your network only operates at 10Megabits a second!!!.
ah, bits bytes, whatever. Its all just techno babble to them. It isn't our fault that networking equipment reports speed in bits and downloads report speeds in bytes.
If you like the theoretical physics (stuff about universes, really really big things, time travel, really really small things, etc) I would highly recommend two books by Stephen Hawking.
His first A Brief History of Time goes into standard theoretical physics with some exploration about black holes, etc.
His most recent book The Universe in a Nutshell has lots of great explanations, in layman terms, of modern theory. He covers p-brans, M-theory, lots of relativity type material. I highly recommend this one.
Professor Hawking also has several lectures on his website that can give you information on more specific aspects of his research.
this material will give you enough background in theoretical physics to wax philosophical and impress your geek friends with talk of alternate universes and multiple dimension objects. At the very least it is interesting material.
uhh its called cable, and we would pay more for it.
or they would change their methods of advertising to counter the technology that lets you skip them. Its hard to tell the difference between someone giving a quick news headline versus a commercial product pitch.
The worlds largest supercomputer is being built as we speak at various campuses around the world. Its a multipart system with various clusters linked together at the different campuses. If your interested I covered the basics of the system below.
TeraGrid is the name of the soon to be world's largest computing cluster that will be completed in 2002. It will contain approximately 3,300 Itanium(TM) and McKinley processors on IBM servers running Linux connected through a Qwest fiber-optic network. Once completed the TeraGrid will be capable of a massive 13.6 teraflops and will have access to 450-600 terabytes of data. This is a huge step (for Intel at least) in acceptance of the Itanium processor into the server market. Intel is fueling the program by providing optimized compilers and software as well as various customized tools.
It is being funded by the National Science Foundation by a $53million grant. Various researchers will have access to the system to perform a variety of simulations. Possible uses include :
-Molecular modeling for disease detection -Drug discovery -Automobile crash simulations -Climate and atmospheric simulations -any other approved scientific research purposes
The TeraGrid will be unique because it will link together various computing clusters at different locations rather than host them all at the same location. Globus is providing open-source protocols that will determine how the grids will communicate with each other. These open-source protocols will create a "plug-n-play" type effect where more machines could easily be added to the network.
The largest section of the TeraGrid will be hosted at the National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign. There will also be portions of the TeraGrid at the University of California San Diego, Argonne National Laboratory, and the California Institute of Technology.
Re:Not Really A Concern
on
Space Wars
·
· Score: 2
Making one missle hit another is not that incredible, and it is far far from being able to shoot down an ICBM.
>How we detect it is irrelevant How so? It is very relevant because we might not BE ABLE to detect it, and in the current state of affairs we can't.
>testing a guidance system any other way is stupid I'm not saying they should test it a different way, what I am saying is that they are not very far along. It is far from working.
>it was a very large success, they successfully made one missle strike another. This worked something like 1 out of 4 tries. That it not very successful in my mind when you are setting everything up to work (using a beacon etc.)
>read what they are working on before you make a knee-jerk reaction to success or failure. It was a minor, minor success, Im not saying they failed.
Lets just hope they post more positive test results soon, because they have a ways to go.
Re:Not Really A Concern
on
Space Wars
·
· Score: 1
sorry, let me qualify, the bandwith crunch isn't really a major problem for the Star Wars program.
No kidding the Global Hawk uses a ton of bandwidth, it delivers videos instead of snapshots and as we all know, videos are large. If they can't deal with the bandwidth then make the Global Hawk take snapshots instead.
Re:Not Really A Concern
on
Space Wars
·
· Score: 1
that is why I said:
"..as far as publicly released information goes"
Its my *opinion* and you should have *realized* that
Not Really A Concern
on
Space Wars
·
· Score: 2, Informative
This bandwidth crunch isn't really a major problem. The Star Wars program won't be up and running for a long time so they have plenty of time to solve this problem. As the program grows and is funded year to year they will have enough money to put up more satellites.
The reason that the program won't be done for a long time is that as far as publicly released information goes, we have only had one successful attempt to shoot down a fake ICBM and this was with a missle that was sending out a HOMING SIGNAL. I doubt the enemy will be so courteous. Also, modern ICBMs, unlike the dummy ICBMs, have many countermeasures to prevent missles from shooting them down. We are not currently prepared to deal with these countermeasures including :
Dummy warheads of same shape and size
Metal Shrapnel Grids that dissipate electric signals and make it difficult for automated identification of the missle
Hot dummys to confuse heat seekers
just to name the ones I am currently aware of. We have dumped billions into this program and had no success yet, who knows when we will actual be able to shoot something down.
We are all finally learning why Intel chose Rambus, they just maybe should have supported DDR first and weened us off of DDR, and not the other way around. I remember when Rambus first came out and everyone was preaching DDR and saying that RAMBUS completely sucked. It is frustrating now to see the big effect it has at the higher Ghz and watch Intel abandon it because of marketing and ignorance of the general populus.
Well lets add another technology to the long list of products that were better than many commonly used products, yet never got significant market share. (BeOs, Alpha Processors, etc. etc.)
Slashdot Mirror
However, increasing word size doesn't mean you will execute more instructions per cycle or per second.
>so doubling the word size, on the bus will improve performance
Maybe - keep in mind that the size of an instruction is not neccesarily the size of the data retrieved from memory. The cache could access memory 128 bits at a time while instructions are only 32bits in length.
I guess I was a little hasty with my judgement that they aren't the missing dark matter. They could be part of the problem, a little chunk off of the grand total of missing mass that we have.
The dark matter could be due to neutrinos, black holes, WIMPs, planets(non-light emitting matter like you suggested), other universes, or any other number of bizzare things that scientists have hypothesized.
Nobody knows what dark matter is yet, but there are lots of guesses.
The reason scientists believe that there should be dark matter is because of the fact that the stars on the edge of galaxies move faster than they should. According to the measured amounts of mass in a galaxy, the stars on the edges would fly out of orbit at the speeds they are going.
Extra mass on the outer fringe of a galaxy could not contribute to this lack of gravity. I am pretty sure that more than 1% of the galaxy's mass is missing also. But I suppose this goes to show that we never know as much as we think we do.
Checkout the everything 2 node on dark matter for more information.
Intel itself has already said that Moore's law is over, explained in slashdot here. Of course, other people are always predicting the end of it as well. Then again, some people think it will continue.
I really wish people would get over Moore's prediction and talk about relevant stuff. There is no way to predict how long unknown scientific breakthroughs will allow Moore's Prediction to remain true. There is one absolute though, the end will come some day, you can only store so many atoms in a certain amount of space according to the rules of quantum physics - that is the absolute barrier.
Until it is actually abandoned I could do without hearing more of Moore's law.
Electron Migration? what are you talking about. Processors continue to get faster and faster due to improved processing technology and increased parallelism. Leakage and electromagnetic interference from the clock signal are major problems today but who knows what scientists are working with nowadays.
> By doubling the word length to 64-bits, you can reduce the clock rate of the chip, and will still be able to perform more instructions per second than your top-of-the-range Athlon/Pentiums.
That is absolutely not true. Having 64 bits allows you to access a larger amount of total memory, and it lets you put more information in each instruction. The amount of data you can work on in any given clock cycle is proportional to the cache access and bandwidth and the register size (Neither of which inherently need 64 bit long instructions).
To perform more instructions per second (or instructions per clock cycle) you need instruction level parallelism (ILP). This has been a major goal of processor manufacturers for many years now. Intel had two main ways of trying to increase ILP.
Itanium has lots of cool new features that compilers could be using and people could be taking advantage of, but it doesn't have good backwards compatibility, and therein lies the problem.
You know why this got so much press? It was exactly because it was a screw up on Intel Marketing. For some reason Intel didn't have a representative in the right place, and AMD got one deal it shouldn't have. Lets compare all the large servers planning to be built with Itanium with those planning to be built with Hammer and a different picture will emerge.
They decide, okay we get use patent X,Y, and Z for $200M and you agree never to sue us for using them again. Intel gets to make processors using the patents, and Integraph gets money for nothin. This way nobody gets dicked by a long drawn out trial to try and explain advanced technical concepts to lay-persons. With a court case, you are guaranteed nothing.
If you knew the amount of time/effort/money that Intel has and is putting into Itanium, you would realize that this suggestion is preposterous.
As far as paying for a loser, Intel has already dumped out $350M to Integraph, these lawsuits are about ADDITIONAL payments. Intel isn't stupid enough to bury themselves in a project they think will never make enough money to be worth it.
Hammer and Itanium are not in the same market. They may compete a little bit, but neither Intel, nor AMD plans to have them competing with each other. Remember Xeon? its still around despite the lack of press. Itanium is supposed to be very high-end computation based and Hammer is a smaller server/desktop product that will compete with Xeon and Pentium brand processors.
Also, the latest and greatest fab that Intel just built in Rio Rancho, New Mexico!. This fab features larger 300mm wafers for an important reason that I tried to illustrate before.
The point is: the processor YIELD determines the cost. The cost of labor and environmental codes is not a large factor (Intel factories all follow the same environmental codes anyway!). The larger wafers will let us get more working processors out of the same wafer, thereby increasing yield and letting Intel sell processors cheaper!
It may also be worth pointing out that much of the handling in the new fab is now automated, creating a reduced labor cost.
Explanation - a rough calculation .13u transistor technology (the price of the wafer isn't important and I just guessed). Little air particles, stress fractures, and errors in processing cause half of the Pentium 4's to have errors and get discarded. This means the Pentium 4's cost $100 ea. to make. Now imagine that they improve the process technology, and there are a tenth of the errors that there used to be. This improvement results in only a quarter of the Pentium 4's on a wafer being damaged. Now we get 15 good processors per die, and they cost Intel only $66.6 to make. Intel just saved a third of the price of making a Pentium 4. Now multiply that savings by the millions of P4's sold and you can see why the yield is so important.
Say intel can fit 20 Pentium 4's on a 300mm wafer. That wafer costs Intel $1,000 to process using all the cool
Processor prices are limited by yield (aka. how many good processors come out of the factory) which is entirely limited by technology. Their fabrication technology is not better than ours right now. If anything, it is far far behind. There is no way that they can produce chips cheaper than we can.
Of course they could possibly get some non-proprietary hardware cheaper, and make their workers work for less. However chip errors occur because of microscopic particles in the air and slight jolts and bumps from a mechanical handler. These are not things that foreign manufacturers have a better handle on than we do. The number of working dies(the chip before its packaged) per silicon wafer is the most important factor in chip price.
When software starts to become usefull people say "Engineering and Math will become obsolete, programs will do all the thinking for people".
And when computers first came out "Computers are going to replace people in all sorts of jobs, soon a computer will be doing YOUR job, better than you do it!".
Anyone remember when Robot's were popular? People were saying that robots will take over all our jobs. Soon robots will do all the thinking, and man will become obsolete.
The industrial revolution, I wasn't there to experience, but anyone suppose there was similar paranoia about large machines? Printing presses? Automated factory machines?
Finally I vaguely remember a quote from history class. It was something along the lines of: "This new invention will be the end of us. Man will no longer have to think for himself, it will all be done for him" and you know what they were complaining about.... THE WRITTEN WORD.
Come on, as long as there is a need for people to know math, and engineering skills (which there always always will be) there will always be engineers. My computer isn't quite perfect yet. I still need to be able to put my stick of memory in, or have some engineer soder a new capacitor in when a chronic blue screen keeps appearing.
And any good hardware buff knows that hardware thrashing is really bad. Once that happens it doesn't matter if your running NVidia or Voodoo or even integrated graphics, its going to run SLOW. If you want to be able to see all your blood and gore, as well as explosives, and triple that up with realistic surroundings and backgrounds you better go with 128MB.
However, I suppose if you run at the lowest levels, turning off all the special effects to try and get the highest frame rate, then the extra memory won't help. You are turning off all the effects that use the memory anyway. So if your going for plain FPS and don't really care about how it looks you will be okay with 64MB.
Also, I recently setup a wireless router at home, and got an 802.11b card for my laptop. Now I can browse the internet on the couch in front of the tv, posting dumb comments to /. threads ;)
Thats because your network only operates at 10Megabits a second!!!.
ah, bits bytes, whatever. Its all just techno babble to them. It isn't our fault that networking equipment reports speed in bits and downloads report speeds in bytes.
His first A Brief History of Time goes into standard theoretical physics with some exploration about black holes, etc.
His most recent book The Universe in a Nutshell has lots of great explanations, in layman terms, of modern theory. He covers p-brans, M-theory, lots of relativity type material. I highly recommend this one.
Professor Hawking also has several lectures on his website that can give you information on more specific aspects of his research.
this material will give you enough background in theoretical physics to wax philosophical and impress your geek friends with talk of alternate universes and multiple dimension objects. At the very least it is interesting material.
uhh its called cable, and we would pay more for it.
or they would change their methods of advertising to counter the technology that lets you skip them. Its hard to tell the difference between someone giving a quick news headline versus a commercial product pitch.
The worlds largest supercomputer is being built as we speak at various campuses around the world. Its a multipart system with various clusters linked together at the different campuses. If your interested I covered the basics of the system below.
TeraGrid is the name of the soon to be world's largest computing cluster that will be completed in 2002. It will contain approximately 3,300 Itanium(TM) and McKinley processors on IBM servers running Linux connected through a Qwest fiber-optic network. Once completed the TeraGrid will be capable of a massive 13.6 teraflops and will have access to 450-600 terabytes of data.
This is a huge step (for Intel at least) in acceptance of the Itanium processor into the server market. Intel is fueling the program by providing optimized compilers and software as well as various customized tools.
It is being funded by the National Science Foundation by a $53million grant. Various researchers will have access to the system to perform a variety of simulations. Possible uses include :
-Molecular modeling for disease detection
-Drug discovery
-Automobile crash simulations
-Climate and atmospheric simulations
-any other approved scientific research purposes
The TeraGrid will be unique because it will link together various computing clusters at different locations rather than host them all at the same location. Globus is providing open-source protocols that will determine how the grids will communicate with each other. These open-source protocols will create a "plug-n-play" type effect where more machines could easily be added to the network.
The largest section of the TeraGrid will be hosted at the National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign. There will also be portions of the TeraGrid at the University of California San Diego, Argonne National Laboratory, and the California Institute of Technology.
Making one missle hit another is not that incredible, and it is far far from being able to shoot down an ICBM.
>How we detect it is irrelevant
How so? It is very relevant because we might not BE ABLE to detect it, and in the current state of affairs we can't.
>testing a guidance system any other way is stupid
I'm not saying they should test it a different way, what I am saying is that they are not very far along. It is far from working.
>it was a very large success, they successfully made one missle strike another.
This worked something like 1 out of 4 tries. That it not very successful in my mind when you are setting everything up to work (using a beacon etc.)
>read what they are working on before you make a knee-jerk reaction to success or failure.
It was a minor, minor success, Im not saying they failed.
Lets just hope they post more positive test results soon, because they have a ways to go.
No kidding the Global Hawk uses a ton of bandwidth, it delivers videos instead of snapshots and as we all know, videos are large. If they can't deal with the bandwidth then make the Global Hawk take snapshots instead.
"..as far as publicly released information goes"
Its my *opinion* and you should have *realized* that
The reason that the program won't be done for a long time is that as far as publicly released information goes, we have only had one successful attempt to shoot down a fake ICBM and this was with a missle that was sending out a HOMING SIGNAL. I doubt the enemy will be so courteous. Also, modern ICBMs, unlike the dummy ICBMs, have many countermeasures to prevent missles from shooting them down. We are not currently prepared to deal with these countermeasures including :
Well lets add another technology to the long list of products that were better than many commonly used products, yet never got significant market share. (BeOs, Alpha Processors, etc. etc.)