Barcelona/Shanghai Smackdown on Nehalem

AMD unveils two-, eight-way Opterons clocked at 2.5GHz
http://techreport.com/discussions.x/14900
Tech Report
by Cyril Kowaliski
June 9, 2008

Since its late arrival in April, AMD's quad-core Opteron lineup has been missing one key element: chips clocked at the same speeds as the fastest desktop Phenom X4 CPUs. No longer. AMD has rolled out four new Opteron SE processors with clock speeds of 2.4GHz and 2.5GHz—just like the Phenom X4 9750 and 9850.

The new Opteron SE line includes 2358 SE, 2360 SE, 8358 SE, and 8360 SE models. The first digit in the model name denotes the maximum number of supported sockets, while the last two digits denote clock speed. x358 models run at 2.4GHz, while the x360 parts are clocked at 2.5GHz. AMD doesn't appear to have revealed thermal envelopes for these chips yet, but their desktop counterparts are rated for up to 125W. Prices are $873 for the 2358 SE, $1,165 for the 2360 SE, $1,865 for the 8358 SE, and $2,149 for the 8360 SE.

According to AMD, these new CPUs are already "widely available," and servers based on them are on their way from "global OEMs and solution providers including Hewlett-Packard, Sun Microsystems, Dell and IBM."


Intel Nehalem Launch Expected to Be Delayed
http://www.xbitlabs.com/news/cpu/display/20080609204928_Intel_Next_Generation_Microprocessor_Launch_May_Be_Postponed.html
xbilabs
by Anton Shilov
06/09/2008

The delay of the mobile Intel Centrino 2 platform introduction may not be the only one for Intel Corp. this year, says an analyst. Apparently, the world’s largest chipmaker may also postpone mass release of its next-generation Nehalem microprocessor for unknown reasons.

Our checks suggest Intel’s Nehalem server parts will initially launch in single-socket only, which we believe will be viewed by customers and investors as a delayed ramp of the company’s second generation 45nm server architecture due out in late 2H ’08. We believe this may be viewed as the second launch delay in recent weeks and is likely to call into question Intel's solid record of on-time execution,” said analyst Doug Freedman of American Technology Research, reports EETimes web-site.

It is interesting to note that chief executive officer of Intel, Paul Otellini, said last year that the company would ship octa-core Nehalem processors in 2008, whereas now Intel claims that the firm will be able to produce code-named Bloomfield quad-core implementation of Nehalem design this year.

Earlier this month some other unofficial sources indicated that mainstream and performance-mainstream chips based on Nehalem micro-architecture will be available only in the second half of 2009, considerably later compared to server or high-end desktop oriented versions.

Still, Mr. Friedman believes that delays of Intel Nehalem launch may be positive for AMD.

“We believe the Nehalem launch plan opens the door for AMD’s Shanghai server chip (45-nm shrink of Barcelona) in 2H ‘08/1H ‘09 should AMD be able to execute to its present schedule,” the analyst said.


Possible shift in 'enterprise' market share looming in the next 12-18 months. Score one for Hector over Otellini on the server side.


 

ryman554

Distinguished
Jul 17, 2006
154
0
18,680
Hardly a smackdown...

1. If, indeed the MP nehalem are delayed, this is a huge misstep for intel and a beloved patriot in their armor. Considering this is the only place left where AMD holds any sort of an advantage, this is a ray of hope for AMD.

2. a 2.5 GHz K10 is hardly earthshattering. Recall, they were claiming the ability to produce a 3.0 GHz part by the end of last year. This falls in the bucket of "finally." However, this will not turn the tides -- not over a 10% clockspeed bump -- but serve to stem the losses.
 
We are not talking about single socket systems ...

We are talking about 4 and 8 socket servers here ... where AMD is King and Intel runs out of bandwidth ... so the cores are starved and just sit there spinning.

Extreme apologies to Yomama who I respect immensely ... this was meant to be funny.




 


And we are not talking about CPUs with a IMC vs CPUs with a FSB.

We are talking about Nehalem that in the desktop market so far has layeth the smakethdown (as the Rock would say) on everything in its path so far. Considering how fast it is in an unfinished product I wouldn't be suprised if Intels new MP systems give the "King" a run for their money easily.

 


Where might one of these be purchased ???
 

timaahhh

Distinguished
Nov 5, 2007
279
0
18,790


R u saying Intel cores can spin?!
 


It ain't the clock - it's the instruction set :)

When it comes to raw SSE performance, the Intel architectures are 3% to 14% faster in the add/subtract/multiply scenarios. When there are divisions involved, Barcelona absolutely annihilates the 65nm Core architecture with up to 80% better SSE performance, clock for clock. It even manages to outperform the newest 45nm Xeon, but only by 8% to 18%. Notice once again the vast improvement from the 2nd generation Opteron to the 3rd generation Opteron when it comes to SIMD performance, ranging from 55% to 150% (!!).

AMD's 3rd generation Opteron versus Intel's 45nm Xeon
http://www.anandtech.com/IT/showdoc.aspx?i=3162&p=6
Anandtech


 

keithlm

Distinguished
Dec 26, 2007
735
0
18,990


If you are referring to the anand preview then that might be a VERY good thing for AMD.
 
Um Wisecracker, arent you the one who started a thread about AMD smaking Nehalem? What they are doing is just denying the obvious. When Nehalem comes out they will no longer have the "We have an IMC so we scale better" excuse.

It will basically be 2 of the same type of chips and one will sclae and perform better.

What I don't understand is how you talk about an AMD chip smaking down Nehalem and yet I mentuion that so far Nehalem is a beast in performance and you ask about getting one.
 


Um wait....Nehalem kreplows Penryn at a clock per clock basis and its a good thing for AMD?

What are you taking and where can we get some so we can all feel the same that Intel having a killer chip on their hands means puppydogs and rainbows and kittens galore for AMD and all will be happy and shiny and pretty.

Sorry dude but I don't live in a fantasy world where I think that way. I prefer to look at results. Nehalem so far is showing a lot of power. If it keeps at this pace I don't know where AMD will be.
 

sailer

Splendid
Another possible, and emphasize the "POSSIBLE" reason for a Nehalem delay could be the number of anti-trust lawsuits coming down on Intel lately. Intel is already under investigation for dumping chips below cast, illegal payoffs, and a number of other things. And this doesn't take into account that Intel is so far ahead of AMD that the only company its competing against is itself. In any case, if Intel plows ahead and buries AMD, such action could bring the wrath of governments around the world down upon its head. It makes better sense to hold back Nehalem a few months and let AMD die on its own, rather than plunging in the technological knife and giving a good twist for effect.
 

keithlm

Distinguished
Dec 26, 2007
735
0
18,990



Let me rephrase my comment to make it more understandable to you:

Are you referring to the anand preview that has been shown to have Penryn numbers that are 17%-20% off to make Nehalem look better? The preview that doesn't show anything more than a few cherry picked reviews with those bad numbers?

That preview?

The preview where if you use the SAME PENRYN NUMBERS that the site had in a previous review for Penryn makes you realize that Nehalem is running about the same speed as Penryn?

That preview?
 


Obviously you are someone who does not pay attention and read everything in order to get a clear view. This is from another thread on here about Anand "Duping" the benchmarks.


Hmmm.... whats that you say? 64Bit gives about a 10% advantage? Just like it gives AMD chips the same advantage? And they ran the Nehalem review in 32Bit. I wounder why they did it in 32Bit..... could it be that they had very little time and wanted to give people something to feed on?

Seriously man....the way you talk all the time about AMD and how Intels chips are worse this I would expect better from you. You know like you reading stuff instead of just jumping to a conclusion.

People these days.
 

sailer

Splendid


Not quite sure how you intend that. Yes, I know that AMD is dieing, as it hasn't turned a profit in a long time and is nearly bankrupt. Without a great infusion of money, from someplace, AMD will run out of money within a few quarters.

Or is it that you mean to laugh at the idea that AMD is dieing? If that's the case, you haven't been looking at their financial picture, one that's bleeding red ink each and every quarter. Add in the effects of the recession, and AMD's chances for a long life look very bleak indeed.
 
The Intel "Blue sticker licker" crowd still want to talk about their single socket gameboys now ... or something that isn't out yet.

Sounds like stalling tactics to me.

Cough up a response or shutup.

*notices it has gone blue with the cold*
 

yomamafor1

Distinguished
Jun 17, 2007
2,462
1
19,790


Shanghai will "pre-frag" Nehalem.

Happy?
 

keithlm

Distinguished
Dec 26, 2007
735
0
18,990


Thank you for the insult. (I guess you were desperate and frustrated and needed to find something to combat those nasty things called "facts".)

However your insult doesn't change the facts. What is not known is whether it was deliberate or an accident. However it does bring the question of credibility of all reviews on that site into question.

Until a review or preview is done by reputable site that doesn't have shifty numbers then we won't really know for sure.

Until then... I laugh at the people that will quote this preview as anything substantial.

(NOTE: I didn't see a thread on this site about how Anand duped the benchmark. I'll go find it. I got my information from Anand's site after reading somebody else mention that previous Penryn results from that site should be reviewed.)
 

keithlm

Distinguished
Dec 26, 2007
735
0
18,990
desktop will be a long ways down the road so I dont know where you get that Nehalem will lay the smackdown in the desktop segment.

This is where Intels 'cheating' comes into play. Its hard to play fair against a cheater that steals its technology from companies like transmetta, AMD, etc. and used unfair business practices to increase its market share.


Come on... we were giving him the benefit of the doubt. **WE** know that Nehalem desktop chips won't be available until 2010... but hey...

Based on what will be available I'm sure we can say with certainty that a when a Desktop Nehalem becomes available it will put smack-down on any Phenom that is currently available at this time.
 


I don't see how it has shifty numbers or is it any site that gives a Intel chip a good name shifty to you? Both the Q9450 and the Nehalem were at 2.66GHz. Bothe ran on Vista 32Bit. Both used the same memory and GPU. So it was a fair test on the same setup just one being Penryn and one being Nehalem. So as far as I can tell those are far circumstances.

In most results like the one with a Q9450 doing better in 64Bit, it would have been the same for Nehalem. Hell Nehalem might get a bigger boost in 64Bit. All we do know that is in 32Bit Nehalem gives a 20-50% boost in performance clock per clock.

But of course you never like to admit when Intel does do well. Hell even if THG also posted the same you would still deny it.