Slashdot Mirror
Intel Shrinks Transistor Size By 30%
pinkUZI writes "Intel will announce that it has crammed 500 million transistors on to a single memory chip, shrinking them in size by 30%. " The tech details are sadly lacking in the article - but I'm sure those will follow. Indeed, the Yahoo piece gives the details that "...has created a fully functional 70 megabit memory chip with transistor switches measuring just 35 nanometers."
Actually reducing transistor size by 30% has a double inpact on the heat dissaption reducing it down by half (70%^2).
Mouse powered Chips, Open source Processors and Lego
It is likely that the new chip doesn't produce any more heat than the old one. It is a very simple effect: smaller transistors require less power to operate. Also, if they did consume the same amount of power in a much smaller space they'd end up as slag, no matter what cooling solution used. This means that if they were to make a current chip using the new 30% smaller technology, the result would probably produce about 30% less heat and use that much less power.
I don't really understand what the big deal is comparing the heat outputs of the P4 and Opteron is anyway, it isn't like these are mobile cpu's. I do have an Athlon 64 under the hood now, but heat output has never been a real concern of mine when selecting a cpu. I'll never understand the processor tribalism that has infected some computer users. Just use what's best for the job.
I submitted this story last night, and it didn't get posted.
With the switches this small, is it safe to say that they are using nanotechnology? I know it's not the cool molecule-sized-killer-robot style nanotech but this seems to fit the description of devices on the scale of a nanometer.
Blaze a trail to the New World
Like I said, I'm not looking for Intel to supply me a cooler desktop CPU. Just like I don't expect nVidia to come out with a cooler high-end graphics card.
AMD/Intel sell to the high performance crowd. They sell supercharged V8s that require a helluva radiator to keep them cool. They even handle overclocking fairly well, which would be like bolting a couple NOS bottles into the trunk.
VIA/Transmeta make little hybrid 4 cylinder engines that are good enough to push around a compact sedan, and you could probably run them for months with a dead radiator, cooling them with just the heater core. (Ie; the cars interior heater on full blast).
They're different things. I'm not shocked when I find out that VIAs stuff isn't in the same performance league as the P4, and I'm not shocked when I find out that Intels stuff is much hotter than VIAs. Just like I'm not surprised to find out that a supercharged V8 in an old muscle car runs hotter and sucks more gas than the 4-banger in my mitsubishi go-kart.
I don't need no instructions to know how to rock!!!!
Electrons move at about 3cm/s
The speed of the electron is not the speed of the signal. Think of a cardboard tube full of ping pong balls. Stick a ball in one end, it pushes a ball out the opposite end.
10 amps of current in a 1mm copper wire has a drift velocity of about 0.024cm/s. Thats how fast the electrons in the wire are moving. The thermal velocity, however, would be somewhere around 100,000 meters/sec. Thats how fast the signal is moving. And it's really close to c/3 (a third the speed of light).
The bound electron whipping around a hydrogen atom is moving pretty damned close to the speed of light.
Sometimes, electrons can move Even faster than light!
Optical computing may or may not be the future. In theory, quantum teleportation and that kind of crap could propogate even faster than a bunch of photons.
I don't need no instructions to know how to rock!!!!
I don't know where the 500M transistors mentioned in the submission come from. I don't find it in both linked articles. I doubt the 64MBit chip (~67E6 bit, marketing makes that 70) uses that much transistors. I think the 64MBit chip is just a demonstration/benchmark for the new process since it's pretty easy to scale a memory chip design to a smaller gate size due to its simplicity.
First of all, they were talking about memory modules there. The more transistors you can fit on them, the bigger memory modules will be. With 64bit computing on the horizon it's about time they increased module sizes and made 2G and 4G modules as common as 512M and 1G are today.
Second of all, you don't have to put more stuff on the chip. They just say they now can do it. They also can make smaller chips doing the same thing which means better yield and less cost.
Carver Mead would say Moore's Law is at an end.
I'm not fully understanding why 70Mbit is being treated like it is so dense. I mean we have 16 chip 512MB DDR nowadays don't we? That's 256Mbit and 256 > 70. Am I just missing some nuance or what?
Thanks.
The radiant power of the sun, at the distance of about 95 million miles (i.e., Earth orbit), is 1350 watts per square meter. The radius of the sun is about 430000 miles. The ratio of the Earth's orbital distance to the sun's radius is 95000000/430000 = 221, let's call it 220 even. Now, power decreases with the square of distance, so take 220 and square it: 48400. So, the sun's radiant power is 48400*1350 watts/meter^2 = 65.3 megawatts per square meter. Yeah, I could probably have looked that up, but it's more fun to derive it.
Now, look at the power output of a typical Intel CPU. Assume a die size of 150 square millimeters, and a power consumption of 150 watts. (Pretty typical values.) That's a power density of exactly 1 megawatt per square meter.
So, the next time you buy an Intel CPU, or any CPU for that matter, remember this: the Sun itself only produces 60 times as much power per unit area as the little chip you are about to hide under that enormous heatsink and fan.
Hell, with numbers like those, I'd wager an Intel CPU produces more power per unit area than a nuclear reactor!