Sign in with
Sign up | Sign in
Your question

AMD Barcelona, Fusion and future

Last response: in CPUs
Share
May 11, 2007 1:02:18 PM

Good post...


Quote:
Much to our dismay and definitely against our recommendations, AMD will not follow in Intel's footsteps and let us do a performance preview of Agena or Barcelona. In fact, AMD wouldn't even let us know what clock speeds its demo systems were running at. While we cautioned AMD that a lack of information disclosure at this point would only reinforce this idea that AMD is lagging far behind Intel, AMD's counterpoint does have some validity. AMD's reasoning for not disclosing more information today has to do with not wanting to show all of its cards up front, and to give Intel the opportunity to react. We still don't believe it's the right decision, and we can't help but believe that the reason for not disclosing performance today is because performance isn't where it needs to be, but only AMD knows for sure at this point.
a b à CPUs
May 11, 2007 1:12:17 PM

I found it interesting the article points out that AMD made "a cardinal sin" by updating both the uArch and fab process at the same time. And while this double update has resulted in significant delays, it seems to me that it was necessary so AMD could remain competitive.

The Fusion project and Torrenza fascinates me. The article really didn't elaborate on Fusion and Torrenza much beyond vanilla business machines and I wonder if they plan to move that concept into the enthusiast/workstation category. I would like to see a dual socket mobo with DDR3 or DDR4 memory support and using one socket for a cpu and the 2nd socket for a gpu. A dual socket mobo like that combined with integrated cpu/gpu on one die would open the door for multiple combinations of workstation/gaming/scientific specific machines.

In light of this article, it seems that AMD will be the "tock" in Intel's tick-tick product strategy by releasing similiar products approx. 6 months behind Intel.
Related resources
May 11, 2007 2:00:04 PM

Granted, but if the latest news that Agena/Kuma is not expected on the roadmap of this calendar year, it really does seem to beg the question of exactly how they are going to compete with Intel. They can't just keep slashing prices and expect to keep the lights on. As I've stated previously I purchased a temporary gap system last week. X2 3800+, motherboard with onboard video/audio and 1GB RAM and got change back from two $100 bills. Those kind of prices are just not tenable in the long run.
May 11, 2007 2:03:58 PM

So we kinda go back to days of a co-processor?
May 11, 2007 4:57:47 PM

Quote:
I found it interesting the article points out that AMD made "a cardinal sin" by updating both the uArch and fab process at the same time. And while this double update has resulted in significant delays, it seems to me that it was necessary so AMD could remain competitive.

The Fusion project and Torrenza fascinates me. The article really didn't elaborate on Fusion and Torrenza much beyond vanilla business machines and I wonder if they plan to move that concept into the enthusiast/workstation category. I would like to see a dual socket mobo with DDR3 or DDR4 memory support and using one socket for a cpu and the 2nd socket for a gpu. A dual socket mobo like that combined with integrated cpu/gpu on one die would open the door for multiple combinations of workstation/gaming/scientific specific machines.

In light of this article, it seems that AMD will be the "tock" in Intel's tick-tick product strategy by releasing similiar products approx. 6 months behind Intel.


If you review the article from DailyTech they indicate that the Conroe competitor died off 6 months into development and Barcelona which was to be the Penryn competition was moved up. So AMD isn't 6 months behind they are actually ahead of Intel releases they just never had a Core2 competitor.
May 11, 2007 7:19:43 PM

Quote:
I found it interesting the article points out that AMD made "a cardinal sin" by updating both the uArch and fab process at the same time. And while this double update has resulted in significant delays, it seems to me that it was necessary so AMD could remain competitive.

The Fusion project and Torrenza fascinates me. The article really didn't elaborate on Fusion and Torrenza much beyond vanilla business machines and I wonder if they plan to move that concept into the enthusiast/workstation category. I would like to see a dual socket mobo with DDR3 or DDR4 memory support and using one socket for a cpu and the 2nd socket for a gpu. A dual socket mobo like that combined with integrated cpu/gpu on one die would open the door for multiple combinations of workstation/gaming/scientific specific machines.

In light of this article, it seems that AMD will be the "tock" in Intel's tick-tick product strategy by releasing similiar products approx. 6 months behind Intel.


If you review the article from DailyTech they indicate that the Conroe competitor died off 6 months into development and Barcelona which was to be the Penryn competition was moved up. So AMD isn't 6 months behind they are actually ahead of Intel releases they just never had a Core2 competitor.

Sounds like their plan with 65nm and 45nm. I guess they figure once they are far enough behind whats the point of spreading out your resources on something that will be behind upon release, and just chug along and take the damge in order to come out stronger. I hope this stratagy works, because its a mighty gamble thats for sure.
May 11, 2007 7:23:58 PM

Quote:
I found it interesting the article points out that AMD made "a cardinal sin" by updating both the uArch and fab process at the same time. And while this double update has resulted in significant delays, it seems to me that it was necessary so AMD could remain competitive.

The Fusion project and Torrenza fascinates me. The article really didn't elaborate on Fusion and Torrenza much beyond vanilla business machines and I wonder if they plan to move that concept into the enthusiast/workstation category. I would like to see a dual socket mobo with DDR3 or DDR4 memory support and using one socket for a cpu and the 2nd socket for a gpu. A dual socket mobo like that combined with integrated cpu/gpu on one die would open the door for multiple combinations of workstation/gaming/scientific specific machines.

In light of this article, it seems that AMD will be the "tock" in Intel's tick-tick product strategy by releasing similiar products approx. 6 months behind Intel.


If you review the article from DailyTech they indicate that the Conroe competitor died off 6 months into development and Barcelona which was to be the Penryn competition was moved up. So AMD isn't 6 months behind they are actually ahead of Intel releases they just never had a Core2 competitor.

Sounds like their plan with 65nm and 45nm. I guess they figure once they are far enough behind whats the point of spreading out your resources on something that will be behind upon release, and just chug along and take the damge in order to come out stronger. I hope this stratagy works, because its a mighty gamble thats for sure.

Leapfrogging is all well and good... when you're sitting on a war chest. AMD needs competitive products every quarter to generate the cashflow that will fund future CAPEX.
May 11, 2007 7:53:42 PM

Quote:
Anand wrote up a very nice article on this, the most encouraging:

For a while we had lost confidence in AMD, like many of you had as well, and although AMD's position in the market hasn't changed we are more confident now that it can actually bounce back from this. Intel seemed to have the perfect roadmap with Conroe, Penryn and Nehalem all lined up back to back, and we saw little room for AMD to compete. Now, coming away from these meetings, we do believe that AMD may have a fighting chance. Over the coming months you'll begin to see why; it won't be an easy battle, but it will be one that will be fought with more than just price.


http://images.anandtech.com/reviews/cpu/amd/roadahead/c...

There's one thing I found disturbing. Why did AMD run Cinebench demo and not let the press see the scores? They were probably embarrassing. It's a very odd way of running a PR department and doesn't really evoke confidence in a product.

AMD's 65 nm process might not be able to handle 4-cores. We saw an early indication of this when the 65 nm L2 cache was slower than the 90nm part.
May 11, 2007 9:46:38 PM

It is apparent that AMD gets where the future of processing is going. At least the Fusion guys do. Torrenza really seems like it could result in a step backward in thinking and makes me nervous that AMD might try to go the multiple socket approach rather than the system in a package approach. If Torrenza is more important to them than Fusion is, then they will lose in the marketplace. Torrenza might be useful in process servers, but will not be useful for most systems.

A lot of what is in this article reminds me of a discussion I had with verndewd earlier this week.
May 11, 2007 10:20:05 PM

Quote:
It reminds me alot of some of the things I have said concerning AMD in the past year. I believe at one point I did say they werent releasing any news as part of a strategy to not allow Intel to answer. I am just amazed at how far they have taken it.

Pushing desktop parts to Q4 says everything about what the intent is. I would hope that barcelona arrives on time. Well , ontime from the last felay announcement. They really need to pick up the pace and get back in black.


At the stockholders meeting Hector stated that they have not been disclosing their technology because the competitor has been copying their ideas.

Since Intel has larger resources they take those concepts and develop their own products faster than AMD. The current dilema was brough on by a product concept failure six months into development so Barcelona was to be pulled up from it's slot to compete with Penryn to release hopefully just after Core2. Well it never got pulled up and will be arriving to compete with Penryn per the original schedule.
May 11, 2007 10:32:33 PM

Good lord I wish AMD would just release some thing.........


Please.........
May 11, 2007 11:53:24 PM

Quote:
Good lord I wish AMD would just release some thing.........


Please.........


Well, it looks as though it won't just be a meaningless socket change with no backward compatability. That was actually the thing that most cheered me up: It looks as if they now care enough about their customers not to work them over with pointless 'well you have to throw a portion of your kit away and buy new' upgrades.
May 12, 2007 1:27:15 AM

Quote:
AMD's reasoning for not disclosing more information today has to do with not wanting to show all of its cards up front, and to give Intel the opportunity to react.
I don't know but this sounds like a load of crock to me. I didn't think these two companies were making products that could be altered "on the fly" sorta speak. I might be wrong, but from memory, I can't recall AMD being this hush-hush about their upcoming product when they were about to release the original K8.

It must be nice to have an extremely skewed view of reality and make millions.
May 12, 2007 2:15:13 AM

:lol:  I don't know
May 12, 2007 3:15:03 AM

What's the most definitive release date for barcy now that they've demo'd it (cough, define, cough, demo, cough)? I expect the announcement from Intel in June for penny. And it wont be 4Q.
May 12, 2007 3:31:48 AM

Quote:
Not sure if you read the other article, but its exceptional that THG actually hints at the fact that Intel is able to copy AMD's ideas in such a small amount of time because of the resources Intel actually has they can actually release the idea first. Makes sense as to why we havent heard much about Barcy.


If AMD is about secrets and not letting Intel know of intelligence, why did they claim 40% greater performance in FP than Core2Duo? Why all the detailed architecture diagrams? Why did they release R600 scores and not Barcelona? Does nVidia not copy and Intel does? Once they made the claims they made, they were obligated to back it up.

I have never in my life seen such a pathetic demo before. When was the last time you seen someone from any tech company run a benchmark and then TRY TO COVER UP THE SCORE USING A PIECE OF PAPER ON THE SCREEN? It's utterly laughable.

I've always been a fan of the underdog but AMD is really screwing things up right now. They need to hire a different PR department.
May 12, 2007 4:20:53 AM

Quote:

If AMD is about secrets and not letting Intel know of intelligence, why did they claim 40% greater performance in FP than Core2Duo? Why all the detailed architecture diagrams? Why did they release R600 scores and not Barcelona? Does nVidia not copy and Intel does? Once they made the claims they made, they were obligated to back it up.


So very true; if they are claiming that Intel will steal their ideas before the processor launches, that's just absolute BS imo. Barcelona/etc is going to launch in a couple of months, and soonafter, the Penryn derivatives. How the hell are you going to make changes to the fundamental design of a processor that quickly? The most Intel can really do is increase clockspeed because it takes weeks/months to test and debug the processor after first silicon. At this point, where AMD is hemorrhaging cash and Intel is profiting on their processors, it would advantageous for AMD to not give vague numbers on specific benchmarks and actually let their fans know that they have a killer product on hand, that is, if their product will live up to their hype. All they have done so far is talk it up without backing their statements with any definitive evidence for their performance claims.
May 12, 2007 5:19:02 AM

very good article...so there isstill hope...
May 12, 2007 7:13:26 AM

What I am seeing with AMD and with their marketing claims of scared to release specifics clock and performance tests because Intel is liable to glean technology from AMD demoing stbr product, and might quickly redesign Intel's well outlaid strategic plan and alter with AMD tech at last second, releasing a new better chip that is built of AMD observance of a demo showing?!?

WTF are they thinking we are??? That sounds like a shariwho or MMM comment more than an AMD statement...
It sure stinks of BS... and imo more likely reflects a mirror image of what AMD is doing due to Intel's release and future outlines, that internally AMD realized they had not counted on such a strong blowing and going from Intel, and Barc was intended to blow the doors off 4Ghz Hyper-threading tech, but C2D stunned them and quickly like a black project AMD is doing the risky quick redesign of their chips using concepts gleaned from Intel's successful C2D technology... maybe even they have C2D samples reverse engineering them to hopefully learn some ideas to implement and are risking it all on it working, and this takes delay and BS press statements to keep the cover on what is really going on behind closed doors...

It could be they produce a good chip, but in the rush, it may be clocked stock at within 5 or 10% of max overclock, which could mean it may perform close to intel, but if overclocked falls way short again...

Regardless how you see it, it is hard to see any positive aspects from the way AMD is conducting themselves.
May 12, 2007 10:18:46 AM

Quote:

Regardless how you see it, it is hard to see any positive aspects from the way AMD is conducting themselves.


It really makes you wonder what the hell they're thinking, given their stock price, cash problems, cut to the bone pricing policy, etc.
May 12, 2007 12:02:32 PM

Quote:
I wouldnt be surprised to see Intel answer in the same manner. AMD confirming my suspicions on this could put intel into a similar attitude and result in the most insane leapfrog loop of all time. :lol: 


Heh yeah. We have a new Intel. Only god knows how nehalem will perform. We know for certain they don't want to be behind AMD again.
May 14, 2007 4:43:25 PM

Quote:
get your 2900xt's ,right at the price range i was talking about !!! get your 2900 xt's :wink:
http://www.zipzoomfly.com/jsp/ProductList.jsp?ThirdCate...



As I said before overpriced compared to the the 8800 gts 320MB shave $60 bucks off and your about there. Should sell ok for a couple weeks while the ATI fanatics scarf them up then sit on the shelves for a while until they figure what the real price should be. AMD has the price cutting thing down to an art though so watch the prices drop should be down to $350 for back to school.
!