IBM stretches the limit on Moore's Law

IBM today announced a breakthrough chip-stacking technology in a manufacturing environment that paves the way for three-dimensional chips that will extend Moore's Law beyond its expected limits. The technology – called "through-silicon vias" -- allows different chip components to be packaged much closer together for faster, smaller, and lower-power systems. The IBM breakthrough enables the move from horizontal 2-D chip layouts to 3-D chip stacking, which takes chips and memory devices that traditionally sit side by side on a silicon wafer and stacks them together on top of one another. The result is a compact sandwich of components that dramatically reduces the size of the overall chip package and boosts the speed at which data flows among the functions on the chip.

"This breakthrough is a result of more than a decade of pioneering research at IBM," said Lisa Su, vice president, Semiconductor Research and Development Center, IBM. "This allows us to move 3-D chips from the 'lab to the fab' across a range of applications." The new IBM method eliminates the need for long-metal wires that connect today's 2-D chips together, instead relying on through-silicon vias, which are essentially vertical connections etched through the silicon wafer and filled with metal. These vias allow multiple chips to be stacked together, allowing greater amounts of information to be passed between the chips.

Link: Forum Discussion (Thanks brianshapiro)
View: Full Story
News source: IBM Research

Report a problem with article
Previous Story

Viridian and Virtual Server Timing Updates

Next Story

Cisco fixes wireless security holes

27 Comments

Commenting is disabled on this article.

IBM makes a "break-through" every damn day, but when will anyone actually adopt the tech into the consumer products industry. It's been about 2yrs since I heard they discovered a way to make transistors that work at like 9000ghz, but it doesn't matter if it's 9billion 9million or just 9, if no1 adopts it, the tech goes to waste.

So onto my point...

This is just to boost Stock speculation for IBM and get young geeks hungry. By the time Intel or AMD decide it's ready to transition chips into something much better, there'll be something MUCH better.

Do you think they don't make chips any more or something?
All the consoles have IBM chips in them..... and they make CPU's which are in many servers and super computing grids. Their Power7 chip is expected to be socket compliant with AMD Opterons.

me personally i always think AMD/intel etc.... are already capable of putting much more powerfull CPU's to the end user already but they "slowly" do it so there profits continue to go higher and higher... cause if they actually gave to the public what there capable of then they would have nothing left to sell and in between each CPU boost would probably be alot longer which would mean less profits for them etc etc.

p.s. i know some stuff aint cheap to produce but i still think they do that to guarantee to consistently have stuff to sell... if im wrong ill still bet they release stuff slower than they could.

This isn't the breakthrough that they're making it sound like. This sounds more like parallel processing where they link many CPUs together. The only real difference is that they are stacking them into one small area. I remember when they linked many Sinclair ZX81's together and got the same processing power of a supercomputer of the time.

The only real difference is that they are stacking them into one small area.

Uh yeah, that's the breakthrough. If you don't think being able to pack the components closer together is extremely important you're really missing the point. It has nothing to do with linking processors together. The article explains it pretty clearly actually.

Well, they can stack them like a Rubik Cube You can play with the CPU on your hand to "overclock" it the way you want, with different "colors" Then you can have 3-d fans rotating around to cool it, or each rubik side has a fan (like a cage-fan) , or a cylinder fan rotating around it like a motor *) Patent this!

Depends on how you interpret it. By megahertz, then yes, it already has faded. Multi-core dies and pipeline efficiency are the cornerstones of processor performance rather than clock cycles these days.

Well to be fair it never did have anything to do with megahertz. Moore's Law states that the number of transistors on an integrated circuit will double every two years.

TRC said,
Well to be fair it never did have anything to do with megahertz. Moore's Law states that the number of transistors on an integrated circuit will double every two years.

Yes, your correct. Been a while since I cracked open the CS textbooks. I guess it holds true or at least until Intel goes 80-core cpu and sets the bar to a new high! Does it apply to a single physical core or comprise multiple core cpus?

Budious said,
Microsoft makes processors now?

I presume he/she meant that instead of allowing us to do things quicker, Microsoft'll just use the extra power to cut programming corners, and add cruft.

TenebraruM said,

I presume he/she meant that instead of allowing us to do things quicker, Microsoft'll just use the extra power to cut programming corners, and add cruft.

Microsoft Windows on IBM PowerPC or Cell processors? Think they stick to ia32-x86/x64 architecture and server support for ia-64 for Itanium processors. Unless you mean Microsoft Office for Mac... I guess they could bloat that some more but Apple has switched to all Intel chips now.

how will they manage heat on a 3 dimensional CPU core? Seems like they would need "holes" filled with copper or something to conduct heat away from the center.

I imagine it will still be extremely thin, it won't be like a big block or something. Probably as thin as CPUs are currently.

TRC said,
I imagine it will still be extremely thin, it won't be like a big block or something. Probably as thin as CPUs are currently.

Also the lower power, obviously I don't know how much lower, will result in reduced heat... making whatever cooling method easier to implement.

...great innovation but - it took 'em a decade of research to come up with THAT?

Okay granted I never thought of it either, but it doesn't strike me as being particularly tricky to pile your PCBs up into a tower, link 'em vertically, and chuckle a little bit.

I may just be overly cynical!

One little breakthrough leads to another. I doubt they have been working for a decade on this one particular idea, it's just a result of all that work.

This may be the next big step for multi-core processors. Intel currently manages to get four cores onto a 2D layout. Just lay another four cores on top, rinse, repeat.

The Intel Q6600 still costs above $800, only 4 cores, so let's not even begin to think about 80 cores. If 3-D chip stacking is unneeded to add more cores, at the very least it could possibly lower the cost of developing multi-core processors, passing the savings on to the consumer.