IBM Leapfrogs Intel With 22nm Chips

Slatterz writes "Intel may be touting 45nm CPUs, but IBM says it can go much further with a strategy to produce future chips using a 22nm fabrication process. The company is adopting a technique called 'computational scaling' in order to manufacture circuits small enough to deliver more powerful and energy-efficient devices. Intel plans to introduce 32nm chips in 2009, but chipmakers have hit a problem in that current lithographic methods are not adequate for designs as small as 22nm owing to fundamental physical limitations. IBM claims to have solved this problem." Unfortunately the phrase "computational scaling" doesn't actually convey any information about how they've solved it.

Comments

Why can't you skip a generation?

[Entropy98]

I know its getting harder and harder, especially considering these things are only a handful of atoms across, but why can't they ever skip a generation? Why work on three generations of chips simultaneously? Why not just skip one?

--
  find my ip

Re:Why can't you skip a generation?

[TheRaven64]

Say it costs AMD 100$ to make, maybe they can sell it for 110$. Enter an Intel 45nm, they produce it for 50$ and still sell it for 110$

True, to a point. This depends on Intel getting the volumes up, however. The vast majority of the cost of a wafer of chips is the cost of building the fab in the first place. Each 45nm fab cost around $1-1.5bn. Intel aims to sell 100m of these by 2011. If they are the entire output from one fab in this time then they have a $10 cost just from fab creation (in practice, they will be the partial output from several fabs - not sure of the exact numbers). If they only sell 50m, then the investment cost is $20 each. In contrast, AMD can keep running their old fabs which have the up-front costs spread over a load of chips they've already made. If they have lower volumes than Intel, then the older process technology will keep being cheaper.

This is why Intel is so hard to compete with. They have such massive sales that their per-unit cost for a new fab is very low, while for their competitors it is much higher. This is why most other chip designers get their cores fab'd by a third party - if you're building chips for a dozen or more other companies then it's easier to get the volumes high enough that the unit cost of production is low.

AMD's partner IBM?

[fishyfool]

Does this mean the Phenom will be produced on 22nm scale? Could be a very interesting development in the AMD/Intel chip wars.

Re:AMD's partner IBM?

[Hal_Porter]

The writeup is misleading. 45nm is in production now, and 32nm is due in 2009. The work at IBM is basic research which will be used by both Intel and IBM to make 22nm chips later on.

At least I think that's how it works. I guess Intel and IBM license patents from each other to allow them all to use the same level of technology. It certainly seems unlikely that IBM will be ahead of Intel in introducing smaller feature sizes since Intel is usually at the head of the pack.

how about something new?

[bussdriver]

I'd like to see somebody do something new besides just get smaller. CELL for example.

Most users are just fine with a fixed system on a chip with no PCI. (ram too if you could pull that off) If you want to reduce power and cost you'd place as much as possible on a single chip. (using crazy IP games they could buy designs for parts on the chip-- consolidating manufacturing as well.)

How about a working variation of Hyperthreading? have 1.5 CPUs and manage it so almost runs like 2 full CPUs? (since pipelines are still problems.)

At least AMD is going to combine GPUs. But next they need to think about how to better integrate the vector processing that GPUs are taking over - instead of the weak MMX/SSE/etc features which have a lot of overlap in their uses.

How about hardware accelerated stacks? MMUs that can handle a driver memory space (not just kernel and user.)

Advances in clockless processing?

Just slapping more cores on chips is the lazy way out. Most people could use a business-class computer on a single chip with a stick of ram. maybe even a slower cheaper but larger secondary ram...(since GPU ram would get used a lot doing all that fluff that every OS now has.)

Catch?

[Tablizer]

Maybe they did achieve 22, but perhaps there's are tiny catch: They don't work. They only claimed 22nm, not working 22nm. Watching all this Nov.2008 campaign coverage has taught me to read between the lines.

Like Intel doesn't have labs working on this?

[Gldm]

What a joke of an article. Every semiconductor manufacturer has several generations of process in various states in the lab. Woo IBM's showing sneak peaks at 22nm!

I met with an Intel VP for an interview a while back and talked about where things are going. He had some nice lab-pr0n of what the photos claimed were 11nm transistors. I believe it was said that was "about 15 years out", and meant to offer reassurance that Moore's Law still had a bit more time left to go.

Actually here, let me go dig up my transcript so I can get a proper quote:

You're going to see that platforms are going to continue to evolve. We're moving to a faster cadence. The processor cadence is about a two year cadence, in terms of process technologies. By the way this is interesting. We know how to do Moore's Law for about another fifteen years which we've never had that kind of length of projection before. ...it sort of takes 3D transistors and all that, but we know how to do these things. It's all using standard silicon, it's CMOS it's extraordinarily well charictarized right? But we've got transistors running at 11 nanometers, I can show you photographs of them. We have the leakage issues but we've got a very good plan.

That was 2 years ago, early October 2006. Who leapfrogged what now?

Re:Like Intel doesn't have labs working on this?

[Gldm]

Umm... and that's relevant how? "Fabrication process" isn't a valid object for future speculation? How about FTA:

"claiming that the process will enable the production of smaller, more powerful and energy-efficient devices "

Will enable? Not does enable? Now enables? Has enabled?

How about "unveiled their fabrication process"? Oh wait it's "unveiled their strategy" for the fabrication process. Not the same thing.

The entire thing is future tense. It's not out yet. Hence why I compared it to other things which are also not out yet. When does the article say this 22nm process will be out? Oh wait, it doesn't! It just says that Intel plans to have 32nm out in 2009, and then something about how "chipmakers have hit a problem in that current lithographic methods are not adequate for designs as small as 22nm owing to fundamental physical limitations" without really anything to back that claim up.

Re:Like Intel doesn't have labs working on this?

[Anonymous Coward]

It's relevant as opposed to "technology" development.
It's not about building a one-shot sample to see if it works or not. It's about developing the way you will mass-produce something that already works.
To rephrase the article:
We _have_ developed a method that _will_ allow us to mass-produce 22nm chips. The method consists in {vague description}.
We're cool because 22nm chips are so-important. They will provide {list of cool features you want today}.

Re:Still not a good idea

[John Betonschaar]

It's not about only fab's, it's also about R&D on the production technology, the machines that perform the 'deep magic' also need to be developed, tested and put into production.

I'm working for ASML myself, which makes more than half of the lithography gear on the market, and I can attest that a surprisingly LARGE number of people on-site here know all the ins and outs of ASML scanner technology, both the stuff already on the market as well as the bleeding-edge stuff that no-one outside is supposed to know about.

ASML has 6500+ employees, so it's a pretty safe bet knowledge leaks out. I don't see why this would be different for IBM.

Re:Well duhhhh....

[Thyrteen]

Prior art.