frnkzks
06-13-2003, 05:39 PM
Go AMD!
Click to See Complete Forum and Search --> : Intel, AMD: Battle of the Transistors
ichorid
06-14-2003, 12:50 PM
souns cool,.............. complicated,...................expensive:rolleyes:
ne1 understand any of that article???
explain it for me??:confused: :(
:t
ne1 understand any of that article???
explain it for me??:confused: :(
:t
milkchaser
06-18-2003, 03:20 PM
AMD and Intel are the largest manufacturers of PC microprocessors (with Intel selling 4 processors for every 1 that AMD sells). Both companies also make flash memory chips which are widely used in handheld devices (e.g. cell phones, palm computers, digital cameras).
Both microprocessors and memory chips are essentially very complicated patterns of transistors. These transistor patterns are imprinted on silicon wafers using proprietary chemical processes. The imprinting process, called lithography, involves shining a light through a mask that makes a particular shadow on the silicon wafer. The wafer is coated with a light sensitive chemical. The regions of the wafer upon which the light shines will chemically react and leave a tiny residue of some kind of material or other. The shadowed region does not and its chemical covering is washed away.
This process is repeated over and over and eventually a very complicated pattern of metal oxides builds up on the wafers. This pattern has certain desired electrical characteristics, chief among them is that of a transistor. A transistor is essentially a switchable electronic gate. Switch it one way and electric current will flow. Switch it the other way and current cannot flow. Put many billions of these transistors together in the right way and you got yourself a microprocessor.
The chemicals, lights, mask and all are tightly kept trade secrets. But each company likes to crow about its advances in order to convince its customers and shareholders that it is not falling behind its rival. This article discusses the broad differences between the AMD and Intel lithography methods.
And these transistors are very, very small. One telling measurement of how small a chip is, is the width of a line drawn by the lithography process. When I was working in the field, we were drawing lines that were a little more than 1 micron wide (a micron is a millionth of a meter). To give you an idea of how small that is, using a normal light microscope to look at things that small, certain really small features are invisible because the lightwaves are so big that they bounce around the feature and not on it. Smaller than white lightwaves -- that's small.
Nowadays, manufacturers are easily and reliably drawing line widths at 0.13 microns (130 nanometers) and working there way down to 0.09 microns (90 nanometers). AMD is working on a thinner than 65 nanometer process. Go AMD!
By making the line widths smaller, they can fit more transistors in the same space. This has many advantages. First, you can make more chips on the same wafer in the same amount of time. This cuts costs and that's one reason AMD can sell its chips for less than Intel. But smaller chips also don't use as much electricity, hence, they don't heat up as much. Also, since the transistors are physically closer together, it takes less time for an electric signal to pass from one transistor to another, hence, the whole processor speeds up. However, when electricity flows it generates a magnetic field and that can disturb the electric flow in nearby transistors, so designing microprocessor circuitry is very difficult (that's why I'm in software ;) ).
Intel is going in a new direction (up) by not so much worrying about just how much it can fit on the flat surface of the silicon wafer. It's somehow building transistors on top of transistors. Prior to this, the complex circuitry patterns have all been mostly two-dimensional. Don't get me wrong, the circuits are really three-dimensional and are built up layer over layer over layer. But for the most part, the circuitry is not stacked. Each feature sits on the raw silicon at its base and reaches for sunlight on top. Intel knows how to change that and stack features on top of each other. Wow. Go Intel!
Both microprocessors and memory chips are essentially very complicated patterns of transistors. These transistor patterns are imprinted on silicon wafers using proprietary chemical processes. The imprinting process, called lithography, involves shining a light through a mask that makes a particular shadow on the silicon wafer. The wafer is coated with a light sensitive chemical. The regions of the wafer upon which the light shines will chemically react and leave a tiny residue of some kind of material or other. The shadowed region does not and its chemical covering is washed away.
This process is repeated over and over and eventually a very complicated pattern of metal oxides builds up on the wafers. This pattern has certain desired electrical characteristics, chief among them is that of a transistor. A transistor is essentially a switchable electronic gate. Switch it one way and electric current will flow. Switch it the other way and current cannot flow. Put many billions of these transistors together in the right way and you got yourself a microprocessor.
The chemicals, lights, mask and all are tightly kept trade secrets. But each company likes to crow about its advances in order to convince its customers and shareholders that it is not falling behind its rival. This article discusses the broad differences between the AMD and Intel lithography methods.
And these transistors are very, very small. One telling measurement of how small a chip is, is the width of a line drawn by the lithography process. When I was working in the field, we were drawing lines that were a little more than 1 micron wide (a micron is a millionth of a meter). To give you an idea of how small that is, using a normal light microscope to look at things that small, certain really small features are invisible because the lightwaves are so big that they bounce around the feature and not on it. Smaller than white lightwaves -- that's small.
Nowadays, manufacturers are easily and reliably drawing line widths at 0.13 microns (130 nanometers) and working there way down to 0.09 microns (90 nanometers). AMD is working on a thinner than 65 nanometer process. Go AMD!
By making the line widths smaller, they can fit more transistors in the same space. This has many advantages. First, you can make more chips on the same wafer in the same amount of time. This cuts costs and that's one reason AMD can sell its chips for less than Intel. But smaller chips also don't use as much electricity, hence, they don't heat up as much. Also, since the transistors are physically closer together, it takes less time for an electric signal to pass from one transistor to another, hence, the whole processor speeds up. However, when electricity flows it generates a magnetic field and that can disturb the electric flow in nearby transistors, so designing microprocessor circuitry is very difficult (that's why I'm in software ;) ).
Intel is going in a new direction (up) by not so much worrying about just how much it can fit on the flat surface of the silicon wafer. It's somehow building transistors on top of transistors. Prior to this, the complex circuitry patterns have all been mostly two-dimensional. Don't get me wrong, the circuits are really three-dimensional and are built up layer over layer over layer. But for the most part, the circuitry is not stacked. Each feature sits on the raw silicon at its base and reaches for sunlight on top. Intel knows how to change that and stack features on top of each other. Wow. Go Intel!
SysOpt.com
Copyright Internet.com Inc. All Rights Reserved.
Copyright Internet.com Inc. All Rights Reserved.