backtop


Print 38 comment(s) - last by mpgenius.. on Dec 17 at 3:10 AM

Processors built with the 32nm process will be available in Q4 2009

One key to getting more performance and better battery life from notebook computers and other devices that use a processor is to build the chip with smaller circuitry. Moving to ever-smaller production methods allows chipmakers like Intel to reduce costs for manufacturing, provide more performance, and to offer processors that are more power efficient.

Current Intel processors are built on the 45nm process, but Intel reported today that it has completed the development phase for its next generation 32nm process. The chipmaker says that it will be producing chips using 32nm processes in Q4 of 2009. Intel has maintained since December of 2006 that its 32nm process was on track.

Intel will provide technical details of its 32nm process technology and other topics at the annual International Electron Devices Meeting (IEDM) in San Francisco next week. This announcement that its 32nm process is developed means the firm has kept to its tick-tock strategy that has it introducing a new manufacturing process and new microarchitecture on alternating years.

Intel says the production of 32nm chips will mark the fourth consecutive quarter it has been able to adhere to the tick-tock strategy. Intel will outline its technology at IEDM and will talk about the second-generation high-k+ metal gate technology and its 193nm immersion lithography process for patterning layers on the chips. These technologies will all work together to deliver Intel microprocessors that deliver reductions in power usage and higher performance.

Intel's Mark Bohr, Senior Fellow and director of process architecture and integration said in a statement, "Our manufacturing prowess and resulting products have helped us widen our lead in computing performance and battery life for Intel-based laptops, servers and desktops. As we’ve shown this year, the manufacturing strategy and execution have also given us the ability to create entirely new product lines for MIDs, CE equipment, embedded computers and netbooks."



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Soon...
By chmilz on 12/10/2008 12:15:06 PM , Rating: 2
Chips will be small enough and powerful enough to be implanted in our BRAINS! Bwaghahahahaha!

Actually, I could use a memory upgrade. I can't remember names, faces, or things I did or said earlier in the day.




RE: Soon...
By omnicronx on 12/10/2008 12:24:37 PM , Rating: 3
I just wonder if they are going to hit a wall soon with the current ways of making chips. Apparently 22nm and smaller is going to be a huge challenge for chip makers.


RE: Soon...
By arazok on 12/10/2008 12:32:43 PM , Rating: 3
Not according to the UK's National Physical Laboratory.

Research shows there could be no end in sight for Moore's Law:
http://www.physorg.com/news148054154.html


RE: Soon...
By masher2 (blog) on 12/10/2008 2:00:11 PM , Rating: 5
It depends on how you define the word "wall". Below 22nm, we will likely have to abandon current lithography in favor of EUV. Also, we lack a suitable high-k material for gate dielectrics at those dimensions, meaning substantial reengineering there as well.

As your article points out, entirely different approaches such as germanium nanowires may eventually come into play, but they're a decade or more away.

The industry hit a little bit of a speed bump at 90nm; the 65 and 45 nodes were smooth by comparison. By many accounts, there are likely to be serious delays pushing past 22.


RE: Soon...
By bridgeman on 12/10/2008 7:21:00 PM , Rating: 5
There is a real wall due to quantum mechanics. Around 10nm (8nm, IIRC), an electron has a 50% chance of simply tunneling through a potential barrier. For any device with those dimensions, the off-state becomes nearly indistinguishable from the on-state. In effect, we can't turn the transistor off.

In addition, these nanoscale devices are exhibiting new behaviors, such as ballistic transport. For digital, that's not a huge problem; all we need is a working switch. However, analog and mixed-signal circuits suffer and SRAM noise margins degrade significantly. In many cases, analog/mixed-signal/SRAM circuits will not scale much, or at all, with each new process node.

As for new devices, the manufacturing guys have been pushing bulk CMOS (and sometimes SOI) as far as possible. Michael's absolutely correct that nanowires and the like are easily a decade away.

(I'm a grad student in circuit design working on nanoscale CMOS circuits, dealing with this stuff every day.)


RE: Soon...
By Regs on 12/10/2008 10:41:09 PM , Rating: 2
I hope AMD pays you well.


RE: Soon...
By bridgeman on 12/11/2008 4:32:31 PM , Rating: 2
With all the hiring freezes across the industry, I'd be happy to just get hired when I graduate next year.

Besides, I could be a complete incompetent in real life. The internet has this funny way of masking the truth :)


RE: Soon...
By flipsu5 on 12/12/2008 12:03:23 PM , Rating: 2
The lithography will not be critical; iterated exposure is just as tolerable as, if not more so than, overhauling to EUV, which is still too slow.

You are right about high-k, I believe, since Intel had to change to a "second-generation" HKMG.


RE: Soon...
By omnicronx on 12/10/2008 2:06:52 PM , Rating: 2
Thats for integrated circuts in general. The processor market is much more limited unless some big changes are made.

Here is what I am talking about:
http://www.eetimes.com/news/semi/showArticle.jhtml...


RE: Soon...
By slickr on 12/10/2008 1:05:49 PM , Rating: 2
There is no stopping for Intel it seems.
Though i think they need to bring their new 32nm in Q3 2009 I think.


RE: Soon...
By flipsu5 on 12/12/2008 12:00:34 PM , Rating: 2
But where are the 32 nm fabs?


RE: Soon...
By quiksilvr on 12/10/2008 3:40:28 PM , Rating: 3
Its ridiculous how ahead Intel is. AMD really REALLY needs to push ahead fast.


RE: Soon...
By 16nm on 12/10/2008 7:17:53 PM , Rating: 2
AMD is broke.


RE: Soon...
By murphyslabrat on 12/11/2008 3:34:10 PM , Rating: 2
America is broke


RE: Soon...
By Headfoot on 12/15/2008 5:05:26 PM , Rating: 2
I don't see why....

Why does AMD have to have a better performing part? It makes no sense. All they need is a part that is better for the money . How many people drop $1k on the latest Extreme Edition? Very few. How many drop as little as possible to get the job done? The entire "business class" computer category. Not workstation, but just average pre-packaged word processor.


RE: Soon...
By Shig on 12/10/2008 12:28:49 PM , Rating: 2
I'm glad Intel is still adhering to their tick-tock timeline, given the recent lack of AMD competition. Pushing that timeline back would really suck for the consumer.

I'm waiting for Sandy Bridge on the 32nm process.


RE: Soon...
By HrilL on 12/10/08, Rating: 0
RE: Soon...
By Samus on 12/10/2008 3:00:28 PM , Rating: 3
"Son, you've smoked yourself retarded."


RE: Soon...
By Tacoloft on 12/11/2008 11:44:45 AM , Rating: 2
Did he smoke himself halfway retarded or did he go full retard?


RE: Soon...
By swampjelly on 12/10/2008 3:02:16 PM , Rating: 3
LSD-25


RE: Soon...
By wrack on 12/10/2008 10:44:48 PM , Rating: 3
Pineapple Express!


Bad news for AMD
By HrilL on 12/10/2008 12:24:34 PM , Rating: 3
This is not good news for AMD because they are just now starting to get 45nm parts out. And intel will once again gain a large advantage is size and be more profitable while fighting a price war with AMD. Just as AMD is starting to get some parts that can compete.

I'd like to see AMD with 32nm if they were on able to manufacture at the same size as intel I think they would be able to have a good competitor or even something better like when Athlon 64s came out even though AMD has always been a year or more behind intel on size.




RE: Bad news for AMD
By TemjinGold on 12/10/2008 12:38:37 PM , Rating: 2
I seriously doubt AMD will put out something better than Intel for a while (better at a certain price point possibly.) A lot of people like to talk back to the days of the A64 but most fail to realize AMD was winning then not so much because they were so good (not saying they were bad) but more so because Intel was so BAD. Intel has learned its lesson it seems so I seriously doubt that kind of upset again.


RE: Bad news for AMD
By 16nm on 12/10/2008 7:28:56 PM , Rating: 2
And when AMD was on top, it demanded top prices. Now, Intel is on top and one can buy a very high performance chip for next to nothing. AMD had its opportunities. They are where they are because of their own doing.


RE: Bad news for AMD
By KernD on 12/10/2008 11:32:29 PM , Rating: 2
What opportunity did they have?
They were selling 100% of there production at those price?
You think they should have lowered there price?
even at the price they asked for, they still didn't manage to pay of the new factory they needed to compete.
So were there price wrong, no, not at all.
What they did wrong was there R&D planning and execution.

They made the Athlon, then a die shrink, then the dual core, then AM2 version, then a scraped design or 2, then the Phenom. It was too little inovation, but then again they have what, 10% of Intel's R&D budget, so who can blame them for pricing at what it's worth? You need profit to grow and survive.

Keep in mind that Intel sold it's god damned Pentium D Extreme edition at 1000$, and the processor wasn't even respecting it's spec for heat dissipation.
Intel relied on the fact that in those days there would be no app that would use the processor at 100%, but if you did, oups, your computer would shut down because the processor got too warm. Did they get any bad press for that? no, AMD should have used that in adds to take some mind share, but then again, AMD's marketing is 10% of what Intel has, and clearly they are clueless.

No one can compete against a company that is this big, unless the big company does bad for a long time by it's own fault, like American car companies, GM lost it's #1 spot by it's own fault, Toyota is good, but with a good GM they would still be small. Intel just realize fast enough that they had gone the wrong way, and turned there ship around really fast.


RE: Bad news for AMD
By Reclaimer77 on 12/10/08, Rating: 0
RE: Bad news for AMD
By KernD on 12/11/2008 4:09:25 PM , Rating: 2
They would overheat with the stock cooler, or with better cooler, the only solution was to add a large fan in the DVD drive bay.


RE: Bad news for AMD
By psychobriggsy on 12/10/2008 1:15:52 PM , Rating: 3
Intel typically ramps a new process early, but slowly. 45nm was launched in late 2007, yet was a tiny portion of production for two quarters afterwards. It allows Intel to get halo products out.

In addition it comes down to performance in certain price ranges. AMD can still compete here with their 45nm process, which with SOI, immersion lithography and AMD's continual improvement program will not be so far behind (as long as they don't do a 65nm again).

You have to be impressed with Intel's ability to hit a new process on time, all the time.


RE: Bad news for AMD
By sxs537 on 12/10/2008 1:43:45 PM , Rating: 3
Also remember Intel has High-K+Metal Gate technology in its 45nm process which AMD (IBM process) does not have yet. SOI gives them an advantage but immersion lithography is just a technique. It doesn't allow any performance advantage except for allowing them to print 45nm. Intel was able to avoid that till 32nm and hence it gains a cost advantage (Immersion is much more expensive).


RE: Bad news for AMD
By flipsu5 on 12/12/2008 12:12:55 PM , Rating: 2
I think this is overstated, since they did not reuse their 45 nm HKMG; it's like starting over for 32 nm, which AMD will probably do. Plus inserting immersion for 32 nm is much harder and costlier than for 45 nm (the tool is more expensive).


RE: Bad news for AMD
By eye smite on 12/10/2008 3:15:30 PM , Rating: 2
I have to agree, the P4s and netburst were pretty sad. The A64 really turned it on with the X2 though. Who I feel sorry for is all the intel fans spending all the money on the new core i7 45nm cpu's only to have to turn around and dish it out again for the 32nm parts, and you know they will. lol


RE: Bad news for AMD
By MarchTheMonth on 12/10/2008 5:07:49 PM , Rating: 2
quote:
Who I feel sorry for is all the intel fans spending all the money on the new core i7 45nm cpu's only to have to turn around and dish it out again for the 32nm parts, and you know they will. lol


I bought a Athlon64 3500, then got a X2 4600 after awhile (long after it was released), and just now got i7 920, and your position is generally speaking that I'm an Intel fanboy and that I can't wait to shell out money for the 32nm chip that comes out next year? are you kidding?

I bought the i7 920 because, as far as any information presented today says, it's a platform that will last many years with many processors to upgrade to. I'm not anticipating that the Westmere will be the upgrade I choose, as most likely it will be Sandy Bridge, that is if I even choose to upgrade, it took a couple of years to really bog down my athlon 3500 with everyday, more intensive things, I really don't foresee that happening again as quickly with my i7 920.


RE: Bad news for AMD
By xsilver on 12/10/2008 8:03:16 PM , Rating: 2
What you should have read was that there is some jealousy for people who can afford bleeding edge parts all the time. Those people dont need to be sympathized for, they have the money - let them spend it however they want.

Regarding i7 though - Im not exactly sure how intel regards this chip atm. Do they consider it a mainstream chip or is it more a server part? eg. wasnt the mainstream 32mn chip going to be on a different socket?


RE: Bad news for AMD
By Reclaimer77 on 12/10/2008 11:54:38 PM , Rating: 2
quote:
I have to agree, the P4s and netburst were pretty sad.


I don't know why everyone gives the P4 such a bad rep. What they lacked architectually they MORE than made up for in sheer data throughput and memory bandwidth. And hyptherthreading, for the most part, provided a benchmarkable boost.

The AMD crowd is all too willing to point out flaws with the design and what not, but they always forget the most important aspect. To the end user, in real world application, the P4's were damn good CPU's.


RE: Bad news for AMD
By tigen on 12/11/2008 1:26:56 AM , Rating: 2
A "good" CPU really means price vs. performance relative to competition. For the most part the AMD offerings were simply a better deal, and that's what matters for the end user.


RE: Bad news for AMD
By Reclaimer77 on 12/11/2008 7:58:45 AM , Rating: 2
quote:
A "good" CPU really means price vs. performance relative to competition.


What competition ? Intel's market share frankly dwarfed AMD's in the P4 days. And most OEM PC companies didn't even offer AMD cpu's.

" End user " doesn't just mean the super geek who builds their own PC.


RE: Bad news for AMD
By Headfoot on 12/15/2008 5:08:51 PM , Rating: 2
They used tons of power and weren't fast for it. They were expensive and they weren't fast for it.

Sounds like a failure to me.


Like AMD and Intel?
By mpgenius on 12/17/2008 3:10:10 AM , Rating: 2
Get them here with free shipping! http://www.thepcpartshop.com




"Game reviewers fought each other to write the most glowing coverage possible for the powerhouse Sony, MS systems. Reviewers flipped coins to see who would review the Nintendo Wii. The losers got stuck with the job." -- Andy Marken

Related Articles













botimage
Copyright 2015 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki