Intel's 45nm chips coming next year

Toggle sidebar Toggle sidebar
old_times

old_times

Distinguished
Feb 14, 2006
248
0
18,680
Intel First to Demonstrate Working 45nm Chips

"Achieving this milestone means Intel is on track to manufacture chips with this technology in 2007 using 300mm wafers, and continues the company’s focus on pushing the limits of Moore’s Law, by introducing a new process generation every two years."

"Intel’s 45nm process technology will allow chips with more than five times less leakage power than those made today. This will improve battery life for mobile devices and increase opportunities for building smaller, more powerful platforms."

Isn't it just getting harder for AMD?!

,,
 
So now they will be able to make more chips on a single wafer.
By the time they release 45 nanos, Amd will be using 65 nanos.
Since Amd beats Intel with the current 90 nanos to intel's 65 nanos, I doubt having Amd's production that much closer will help Intel at all.
 
Intel has already built functioning 32nm parts while IBM cant even get 65nm working :roll:

@topic:
Its always getting harder for AMD unless they get some serious help from other manufacturers.
 
Isn't it just getting harder for AMD?!
Click to expand...

And why??

It's a sweet joke for me to know that Intel's 65nm aren't cool enough than AMD's tried-and-true 90nm. :lol:

Did you see the new AM2 processors (specifically the 35W X2 3800+) which will consume no more than 65W (excluding the FX-62)?
Let's not forget that these are still 90nm 8) [/code]
 
These projects -if true (link?!)- normally take a long time to get to the production line. This is important that Intel has working sample of 45nm chips.

Did you see the new AM2 processors (specifically the 35W X2 3800+) which will consume no more than 65W (excluding the FX-62)?
Click to expand...

I may agree with you on the performance of AMD chips, but still I’ll wait to see what Conroe has to offer. Finally, it comes to performance/price ratio. Due to the reduced cost with this technology, Intel can offer prices we have never seen on new chips.

,,
 
Intel might as well skip 45nm and go to 32nm, AMD has Intel's 65nm beat with their 90nm on performance and power consumption! They have consisitently demonstrated an inability to efficiently use a die at any size. There is no reason to believe that even if Intel does go 45nm they will be able to manage their heat loss issues and performance issues while demonstrating an edge over AMD. Take that you fanboys!
 
Sure they can. Intel has their 65nm cores out but with old technology that really doesn't work well unless you OC them. Conroe is a whole new ball game and should most certainly gain back plenty of respect for Intel.
 
no dvd, you're not even close. right now intel is working on 1.8121551 nanometer chips... :roll:

there's supposed to be a theoretical limit of the size which you can make transistors before you get too close to the size of atoms. so then what? will we have loose atoms floating around in our chips? will current leakage be avenged tenfold?
 
pretty much all i see is intel is jumping from milestone to milestone and amd is following with various improvements from intel's "mistakes", so really AMD has an advantage.
 
Wow, reading this stuff is really hilarious. so many taglines and key words without the faintest comprehension of what is really going on. just a bunch of people out to say why their chip is beter than the other guys.

First off AMD does have performance lead on many applications. They also have a power lead. that is known.

However, to blindly say that since AMD, using a 90nm SOI process, is lower power than the P4 EE on a 65nm process means that AMD's 90nm process is so incredibly that much better is just raising a big flag to say "look at me i haven't the faintest clue what i'm talking about!!"

the fact of the matter is that AMD's power lead has everythign to do with their clock speed, not with their process. higher frequency, higher power. look it up in any 1st-year electronics textbook.

If you want to make a process comparison, therefore, compare centrino. Intel's Yonah processor (Core Duo) to AMD. They are much closer frequency-wise.

Here all you AMD people will say "yea, but that's a mobile processor, that's not a fair comparison!". so? neither is the AMD-to-P4 comparison when measureing the efficiency of the process.

So if you think Intel isn't leading the process-technology or manufacturing revolution just cause your Athlon FX runs warcraft at 20fps faster than the P4, you're again, waving the dummy flag.

Intel has over 2M sq-feet of cleanroom space. They are spending $3.5B to build a production FAB for a fully-proven, chip-in-hand, 45nm process technology they alrady developed in their research fab. And they already have 3 full production fabs (using 300mm wafers) on 90nm technology.

AMD has 250,000 sq-feet of cleanroom space, they are just now getting their 300mm wafers online (Intel had this in 2002), they are using a 90nm Silicon-on-insulator process (that incidentally requires more-expensive wafer stock with substrate "stuck" onto silicon dioxide, thus eroding profits) . And as far as all of you saying how they are "working with IBM on 22nm process"...if you think that Intel isn't working on the next process, hell, the next 4 processes, without help from outside companies, you're sadly mistaken

See these links from 2003! Paul O. showed photos at IDF of prtotype transistors (in-hand) all the way down to 22nm
Intel.com

and the following press about it:
C-Net

Intel is the world leader in semiconductor manufacturing. period. That is their power. if you think that AMD has *anything* on them in that regard I eagerly await hearing what that might be.

Also, there is more to process techology than power, or performance. There is cost. Intel, or AMD for that matter would gladly accept the exact-same chip on a smaller process, with no improvements whatsoever. In fact, that is *exactly* what intel did going from their 90-nm to 65-nm P4 processors. there was no performance gain there, only the process change. So why do it? simple. lower cost per processor. Fabrication goes through a number of steps per wafer (lithography, implantation, NiO2 deposition, cleaning, SiO2 growth, poly-dep, etc...). the time it takes the wafer in each one of those steps dictates how many wafers you can put thorugh the fab. If you can get more processors (dies) per wafer, great, because it doesn't change how many wafers you can get through the factory - just that you get more chips per day, and thus you increase your capacity, and lowers your cost/processor (which is then sold at the same price or more).

Intel leads the world in that regard. AMD has to go on IBM's research coattails, announcing joint-partnerships and so forth about "technology breakthroughs " that intel announced years earlier. sometimes about technologies that intel already has in production.

So there are many factors at play here, but as far as the title of this thread, If nothing else, just realize that lower-clockspeeds means lower power. It does not mean better process technology. A new process technlogy usually means lower power as well, but if you're gonig to compare that directly, you have to compare apples to apples. Right now, that's not available, since CoreDuo operatesin a 35-W TDP mobile envelope. Athlons are in the 85W desktop envelope, so naturally can be run a little faster, higher power, higher performance.

The true comparison comes with Conroe later this year. I can't wait to watch and see what the 65-nm process can really do at a lower clock speed.

-Mortt
 
I can't help but agree.

The fact is, Intel has always been a manufacturing company first, which has become blindingly obvious with the failure of the Netburst architecture. I think a lot of the cynicism about Intel's manufacturing abilities was due specifically to the 90nm transition with Prescott. However, looking at Prescott alone doesn't tell the whole story. While Intel did initially have problems going to 90nm it actually isn't as much of a black mark on the process itself as people may think. Just look at the Banias to Dothan 130nm to 90nm transition. The maximum clock speed was increased from 1.7GHz to 2.1GHz and the L2 cache was doubled all the while the TDP actually dropped from 24.5W to 21W.

As well, claims that AMD's 90nm SOI technology is far superior to Intel's 65nm process is overblown. All we're seeing right now is Intel's initial production run on 65nm. The current Cedar Mill and Presler chips aren't very impressive, but they were the test-run and I believe that they're production actually started over a year ago. Yonah is the second run on 65nm process and it's managed to actually increase clock speeds from Dothan's 2.26GHz to an eventual 2.33GHz while increasing the FSB to 667MHz and adding a second core with the TDP only increasing 4W.

65nm is even now currently moving into its third run with the C1 stepping Cedar Mills and Presler. Currently the fastest mainstream dual core, the 3.4GHz 950D, has a TDP of 130W and with the new stepping will actually drop to 95W. You can argue that using mobile chips like Dothan and Yonah to look at process technology may be cheating, but the results are impressive when looking at these desktop chips. Netburst is a decidedly inefficient architecture yet with the C0 stepping Intel has managed to fit two 3.4GHz cores into a lower TDP than the original single core 550 while still doubling the power consumption L2 cache. Certainly the ability to stuff two 3.73GHz cores together into a similar TDP to the single core variety is good progress. The yields have also increased to an extent where Intel is planning 50% price reductions on the 950D. 65nm is maturing nicely and will be making it's fifth run on Merom and family. (The fourth run is the new stepping of Yonah coming before Merom, but after the Presler/Cedar Mill C1 stepping.)

It wouldn't be wise to simply dismiss Intel's 45nm architecture. You can debate how much or how little the 90nm to 65nm transition brought, but the 65nm to 45nm transition will be different. 65nm was only a refinement of existing technologies in the 90nm process, mainly in the strained silicon and various gates. However, 45nm will continue improving strained silicon and other features while adding PD-SOI. We can agree about the benefits that SOI brought to AMD's processes, and now Intel is planning on skipping SOI and going directly to it's next improved incarnation. The difference between 65nm and 45nm should be quite a bit larger than the previous transistions.

http://www.theinquirer.net/?article=25512

In fact, they go so far as to claim:
Think happy thoughts here people, from what several sources have told the INQ, the leakage problem is solved, and I mean solved, not lessened.
Click to expand...
I know we can debate over the validity of the sources and the claim, but it's worth noting the bluntness and the lengths the article goes to disuade doubters. Certainly there must be something to the claim, if they are willingly to go so far out on a limb to support it.

So, if you hear gushingly good things about 45nm coming from IDF, believe it. If you hear anyone pooh-poohing Intel and its process tech because of the debacle that was 90nm, just point and laugh. This one will be very very good.
Click to expand...
 
There's a nice chart that David Wang from Realworldtech.com put up about process technology papers from IEDM (biggest annual process technology conference).

http://www.realworldtech.com/page.cfm?ArticleID=RWT123005001504&p=14

When debating "who has the low power process", I think it's interesting (well, to me anyway) to compare the Ion/Ioff numbers for the various process technology papers. "Ion" is the saturation current when the transistor is on - so it's basically how much current the transistor can source when it's conducting... higher numbers are better. And then "Ioff" is the leakage current when the transistor is off - so it's basically the amount of current that leaks through the transistor when it's supposed to be off.. lower numbers are better.

Note that to compare apples to apples, you want to compare columns that have the same "Vdd" and the same process node. Papers 10.8 and 3.3 make a nice comparison.

The SRAM cell size is also interesting - it's basically how small one can draw a minimum sized SRAM cell in square microns.
 
basically intel has more dough :lol: intel just sells technology and "oooo-ahhhh" the consumers wheras AMD tries to provide the highest performance on what they have. So in the research and technology area, intel is in the lead wayy ahead while AMD tries to catch up. To me, they are doing a good job at it. 😀
 
Actually the 45nm SRAM is best acheived by Intel at this time. And IBM is announced this week that they will print circuits with 30nm ridges using current lithography imaging processes.
 
wow ycon, your sooo bias, go away, dvdpiddy is a little bias towards amd but they are better at the moment, so what's wrong with that, also, some of you need to shorten your posts[lt data], man i read one sentence and skip it. ALSO, conroe is mainly speculation right now, and someone said, i thinkj it was mcdonalds guy, that amd cleans up intels mistakes so they have an advantage, but now with conroe having such a low clock speed for intel, it seems they are copying amd, and since most people see a lower clock speed as crap, that may be bad for intel's sales.
 
frankly, who gives a rats ass who thinks what, i like AMD cuz they are cheap and give good preformance

you may like intel becasue...danm im stuck 😀 ..but you may like intel

this argument is liek the ATi vs Nvidia someone will come by and say ATi is better then someone will reply back saying Nvidia has a lifetime warrany on their products then someone else will say ATi si better etc and the process will go on
 
true, but after reading everything i have this to say:
i use both amd and intel for my builds, and honestly they both preform very well cuz when it comes down to it its how well it operates and not all of that crap bout how it was made, i mean if intel has a better make then amd thats great, but i want to no who is going to preform better and who is goin to last longer. and even if intels new 45nm chips are amazing then thats great, then ill use intel, but if amd is better then im gunna use amd and where does cell processing fall into all of this?? its being used in the ps3 and had herd it was coming out for pc as well,a lil off topic jw. so if u think im bias ur wrong, i use w.e i want and w.e i think is best for what it needs to do. amd has gaming, intel has multi apps and stuff of the sort. and there right u cant really compare amd and intel acurtly in build style there to differnt. amd is using older stuff and making it preform better but intel is not doing so bad and beats amd sometimes. u would need 2 cpus with same core and same socket and same everything, then we benchmark, then we no who is better, but right now they dont wanna do that, so im personaly left to speculate.

If u compare a 2.8 intel and a 2.8 amd amd is going to win becuase they refined that chip better and intel wins with ultra high clock speeds. on the other hand, amd has intel beat in FSB, 2000mhz and 1066 max for intel. they also use ddr2 for intel and ddr for amd. so again u cant really compare any of it all, i mean ya u can run tests but there are to many vairiables in it.

Heat is a big thing for some ppl, but i mean if its stayin under 60c full load shut up and let it run, thats plenty cool and within the specs, plus its free heating! YAY!! yes i have water cooling but thats cuz my parents complained bout the noise lvl, and its better for ocing, but to run it at stock speeds under 60c ur fine, when u start ocing then u need to cut that in half, but thats neather here nore there.

so more or less im with who ever i think is better at the moment or who ever im building the comp for thinks is better. so ill continue to build both and praise both. and if u have a problem with it then that sucks.
 
i think you guys missunderstood me cause i don't really like intel's stuff at the moment and really like AMD. Also, my point was that people can think what they want but they are ALL speculations until you can have a fair test on equal ground with 2 equivalent processors.
 
thats what i said, there is nothing that u can compare becuase of all of the vairbles so im just taking what i like and leaving it at that and ty for agreeing wit me im happy now, i can sleep easy 😛 😛
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom