AMD and 65nm

[conroe]

AMD road map surfaces, http://www.xbitlabs.com/news/cpu/display/20060207225940.html

According to slides which resemble those from the roadmaps of Advanced Micro Devices, the first desktop chip from the world’s second largest chipmaker made using 65nm process technology will be released in the first half of 2007 to substitute the mainstream dual-core 90nm products. Another 65nm microprocessor is slated to be released in the second half of next year and target entry-level computers.

What, the middle of next year? http://www.dailytech.com/article.aspx?newsid=418

AMD is expecting that its ramp up to 65nm fabrication technology is going to happen fairly soon.

Well next year is not "soon." And yet another delay for AM2 because performance is bad. Can we realy hope to see 65nm just over one year behind Intel?

Looks like Intel will be selling 45nm about the same time AMD gets 65nm realy going. This spells trouble. If Intels h2 offerings are anything like the scuttlebutt, AMD may be doomed.

 

[dvdpiddy]

I don't think so conroe i think amd will pull a rabbit out of the hat. I think that they are delaying m2 on purpose to make intel feel confident that they will win the fight.then when time is right amd will release m2 with atleast a 20% performance increase then conroe with same specs amd people will be like wooh did you see that a 2ghz m2 with a gig of ram is beatin a 2ghz conroe with the twice the amount of cache holy crap and it's cheaper too gotta get an m2!

 

[conroe]

Did you read An And's bloq? He said AM2 is performing lower than 939, and An And is very respectable. That would be awful if AM2 came out like that.

 

[conroe]

...performance of AM2 is significantly lower than Socket-939 today.

-- http://www.anandtech.com/weblog/default.aspx?bid=276
Months after is should have been on the market, and stll nowhere near to a good product.

 

[dvdpiddy]

strangestranger is right conroe soi is whats kicking intel's silicon ass you see soi was developed by ibm and amd licensed it but intel is to stupid to accept technology from another company cause they want people to think intel doesnt need to ask for help cause they are the best :x and thats not true amd can atleast admit when it's procs need working on look at the opty 150 beats a xeon 3.6 in nearly all tasks and it takes half as much power and produces less heat.

 

[conroe]

http://www.the-inquirer.com/?article=29504
http://www.theinquirer.net/?article=25623
http://www.theregister.co.uk/2004/03/17/intel_plots_4mb_l2_64bit/
http://www.digit-life.com/articles2/cpu/p6-nexgen.html
http://www.infoworld.com/article/05/08/24/HNconroeintel_1.html?s=feature
http://www.xbitlabs.com/news/cpu/display/20050512111032.html
http://www.intel.com/pressroom/archive/releases/20020919tech.htm
http://www.electronicstalk.com/news/iel/iel115.html
http://www.intel.com/pressroom/archive/releases/20050823corp.htm
http://www.dailytech.com/article.aspx?newsid=418
http://www.theinquirer.net/?article=29342

The list goes on an on. No one honesly believes Meron/Conroe/Woodcrest are going to be bad, It's 3 to 3.5 the performance-to-watt of Presler. A $530 2.67 chip will out-perform and use much less power than a FX-60! Add the fact AMD's 90 nm will not likely ever be able to produce much over 3ghz in any volume, AMD needs 65nm soon, just like Hector Ruiz said.

I can find more links, need them?

 

[dvdpiddy]

http://www.the-inquirer.com/?article=29504
http://www.theinquirer.net/?article=25623
http://www.theregister.co.uk/2004/03/17/intel_plots_4mb_l2_64bit/
http://www.digit-life.com/articles2/cpu/p6-nexgen.html
http://www.infoworld.com/article/05/08/24/HNconroeintel_1.html?s=feature
http://www.xbitlabs.com/news/cpu/display/20050512111032.html
http://www.intel.com/pressroom/archive/releases/20020919tech.htm
http://www.electronicstalk.com/news/iel/iel115.html
http://www.intel.com/pressroom/archive/releases/20050823corp.htm
http://www.dailytech.com/article.aspx?newsid=418
http://www.theinquirer.net/?article=29342

The list goes on an on. No one honesly believes Neron/Conroe/Woodcrest are going to be bad, It's 3 to 3.5 the performance-to-watt of Presler. A $530 2.67 chip will out-perform and use much less power than a FX-60!

I can find more links, need them?

no way conroe i think amd is purposely doing this to make intel feel confident then bam intel's dead

 

[luminaris]

I respectfully disagree. Intels current 65nm chips are still based on the P4 architecture which Intel now knows, has been doomed. What they did was transition their new process based on old technology just to more or less, real world test it. Its already proven to run cooler and much more efficient which is what Intel wants.

As far as conroe goes, it will be a new ball game regardless. The new chips are based on pentium M platform which has already made a name for itself. When Intel goes to 45nm, that will certainly be a huge advantage as well considering 65nm has already made big gains for them. Ok, maybe not so much as performance goes, but efficiency, yes. Its a good starting point for them to be very competitive.

Intel knows they have to produce better products now as they are in trouble with the whole anti trust thing, which I strongly do not support in any means. Better products means better performance and competition between the two.

 

[conroe]

http://www.the-inquirer.com/?article=29504
http://www.theinquirer.net/?article=25623
http://www.theregister.co.uk/2004/03/17/intel_plots_4mb_l2_64bit/
http://www.digit-life.com/articles2/cpu/p6-nexgen.html
http://www.infoworld.com/article/05/08/24/HNconroeintel_1.html?s=feature
http://www.xbitlabs.com/news/cpu/display/20050512111032.html
http://www.intel.com/pressroom/archive/releases/20020919tech.htm
http://www.electronicstalk.com/news/iel/iel115.html
http://www.intel.com/pressroom/archive/releases/20050823corp.htm
http://www.dailytech.com/article.aspx?newsid=418
http://www.theinquirer.net/?article=29342

The list goes on an on. No one honesly believes Neron/Conroe/Woodcrest are going to be bad, It's 3 to 3.5 the performance-to-watt of Presler. A $530 2.67 chip will out-perform and use much less power than a FX-60!

I can find more links, need them?

no way conroe i think amd is purposely doing this to make intel feel confident then bam intel's dead

Show just ONE link to confirm your possition. :lol:

I'll get you ten more on mine, ha ha ha.

 

[luminaris]

Well, your talking server products there man. We all know AMD has the throne as far as servers go.

 

[conroe]

Yah, I wish I got one when they were low-balling them last fall.

 

[DuxSyagrius]

Damn dude! Your arguements sound like they would come from an Intel fanboy.... I dont think AMD is seriously trying to fool Intel into beleiving they are going to have faster chips and then.... BAM! They get their hearts broken. Intel has plan and I am sure it will be succesful. Even if Conroe is on par with AM2 it will be a victory for them. If we look closely at Intel's roadmap, all of the new technologies will converge and make a very fast low power/low heat chip. Will AMD answer? I am sure. But no one can stay on top forever. AMD did have the performance crown before too you know...

 

[luminaris]

So true. I mean, Intel is a much bigger company with more plants, more development, more money and more resources than AMD does. Why Intel has been so behind these last few years is beyond me. The only thing I can say is, let us wait and see what lies ahead

 

[DuxSyagrius]

Its not like we havent seen this before. The original Athlons were a little faster (if not a lot hotter to counter) than the P3. Intel released the P4. It sucked. It was slower than the tualatin P3s. but Intel had a roadmap that called for a series of upgrades that spanned a few years. The athlon XP was an upgrade as well and a good one at that. But the upgrades just werent as far reaching as Intel's. I'm sure everyone remembers just how much better the 800mhz FSB HT P4s were compared to the athlon XP. Then came the athlon 64...... This time the tables were turned. Repeat what I wrote above.

Here is my theory. Intel knew that the last generation of P4s was what the P4 was meant to be all along and I bet they had the technology to do it. But why let the cat out of the bag all at once? They had a captive market that would keep buying new mobos/RAM/ and CPUs again and again. I guess as long as they made tons of money they were willing to trade the performance crown with AMD during the 400mhz FSB to 533mhz to 800mhz to 800mhz plus HT upgrades. When it comes to the athlon 64 era, I think they gambled too much and ran out of steam. You know AMD wallowed in misery for many months with the K6 CPUs while cooking up the athlon. They suffered just as much during the later stages of the XP while desining the A64. Intel is sucking tit right now just as AMD did in the past. They have been here before and came back. I just hope AMD doesnt get as complacent as Intel has in the past. Competition is good. I dont want to see AMD have 6 to 8 upgrades before AM2 realizes its true potential. For once I just want to see a company ket the cat out of the bag at the beginning at release a truly compelling product that we just simply must have. Something along the lines of the 9700 Pro or the 4600Ti.

 

[luminaris]

First off, let me just say that the P4 Prescotts are not that bad. In reality, the only real arena that AMD has beaten it to death is efficiency. Performance is another area that only seems like it does on paper. Yes, the intels use more power and you get less performance per watt however, you OC these prescotts and even preslers, the performance margin narrows considerably (as shown on paper) and in some areas, beats the AMD hands down. So I for one, would never say that AMD has the stronghold entirely. To me, its the angle of how you look at the whole picture.

Judging by what I've read here in the forums, some AMD users only choose to travel down a one lane road. I use both processors and stand by what I see here in my world.

 

[brainysmurf]

Well, I followed the link to Andand's blog.
I think you're overreaching and misrepresenting what he said.
He hasn't gotten his hands on a "real" AM2 or "real" AM2 motherboard.
What he's seen are very early, pre-production experimental components that don't give an accurate, real-world view of what the finished CPU and board's will do.
He was basically apologizing for not yet presenting benchmarks, because those benchmarks would not be representative of the actual product.
There's no way that AM2 could possibly perform worse than Socket 939 upon release. AM2 basically IS socket 939. There are some minor revisions and updates to the core, along with a new memory controller to allow the use of faster DDR2 ram, and a change in the socket design to avoid confusion (so newbies won't fry the processor by trying to use it in the wrong motherboard).
AM2 is NOT a major new archictecture, it is not comparable to Intel's transition from Pentium III to Pentium IV.
AMD's next really new chip is coming out after the switch to 65nm, possibly late in 2008.

 

[conroe]

That road map may very well be a fake.
http://www.hardwaresecrets.com/article/282?www.dailytech.com
What I said last November at our annual analyst meeting still holds. We have had 65nm preliminary silicon running in our Dresden Fab 36 since last June; we plan to begin 65nm volume production in the second half of 2006, and we plan to be substantially converted to 65nm in Fab36 by mid-2007.--Phil Hester VP and CTO AMD

 

[svyerkgeniiy]

Why is this necessarily a bad thing? It is not clear to me that Intel has a superior position over AMD merely because they offer a superior process technology sooner. Will Intel's 65nm process give them a market leap over AMD, merely because AMD has "only" the 90nm process? I think they have more problems to overcome (design, for one) than just process. Did their use of DDR2 allow them to wollop AMD over the head? Not really.

AMD has always been slow to put new technology into place, mostly because new technologies don't always provide an advantage. Now that DDR2 can provide more measurable or at least equivalent benefits, they now incorporate it.

While they don't have 65nm for another year, I'm not worried yet. It does give Intel an economic advantage on a per-unit level, but only after investing a huge amount of capital to make it happen, so even that benefit is mitigated.

--dv