Do you have a link?

I could use some laughs...

-VBG

Edit: looks like chipguy beat me to the punch

True, with the Intel(R) Anteng(R) processor, how could it only be only an 8 socket system. LOL!

Seriously, what convincing evidence is there that this isn't a bunch of 8 socket Itanium 9300s connected by Infiniband?

Was it the phrase "Simplify system expansion and update; support smooth expansion of nodes"? No?
How about "Adopt cluster architecture"? Probably not.
"Create cluster single system image"? Maybe that is what is throwing you.

If it supports 2~1024+ nodes and 4~2048+ Intel® Xeon® processors, this indicates to me that this likely isn't even a cluster of 8 socket systems, but a cluster of 2 socket systems. BWDIK?

-VBG

All Atoms or Netbook Atoms?

Why does it seem like Intel is always partnering with also-rans?
Or at least not the #1 player in a market.
Sony and TVs instead of Samsung.
Nokia and smartphones.
Sprint and WMAX.

Or am I off-base?

-VBG

I think I quoted verbatim from the article.

Much earlier if you belive the rest of the article:

Commercial production of 32-nm SOI HKMG is due to start at Dresden in July 2009 and 28-nm bulk HKMG is due to start at the end of September, the same time as the foundry operation of Samsung Electronics Co. Ltd. intends to start

http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=223500059

Must be a rehash of an earlier piece. Or an error.

-VBG

Maybe the mac launch is timed with the Westmere EP launch.

Why couldn't they use Xeons?

-VBG

Any worries in terms of the toxicity of any of these?

No.

First of all, wrt lead, there is a difference between lead in solder joints and material that is buried under 10+ metal layers and a SiN passivation layer.

Second, think for a second how much arsenic we are talking about here when it is a layer prob < 200 nm thick relative to the thickness of an entire wafer (775,000 nm).

Third, arsenic has been used as a dopant for 40+ years of semiconductor manufaturing.

Fourth, do you hear anyone worrying about the arsenic in RF components like the power amplifiers for cell phones?

I could go on but I think you get the point.

-VBG

AMD Poodles vs Intel Doberman

No dogs, but how about cats?

Anil is copping a plea mañana and is rolling over on poor Raj.
Maybe I'm just being dense, but I don't see the connection with Ruiz being the paid source. If anything, the article seems to imply that Kumar, having been accused of tipping off Raj on the ATIC deal, was also the source in the ATI deal. Connect the dots for me here.

-VBG

Nothing on Bloomberg says Hector is the tipster...

link from Tecate

It is just a repeat of the October story. It is in a completely separate section labelled "AMD Spinoff"

Separately, an unnamed AMD executive was cited in the criminal complaint against Chiesi. The AMD executive allegedly told Chiesi about an impending spinoff. That person is Hector Ruiz, AMD’s former chief executive officer, according to a person familiar with the investigation. Ruiz stepped down as CEO of AMD, the world’s second-largest maker of personal-computer processors, in 2008. Jeremy Fielding, a spokesman for Ruiz, declined to comment

-VBG

Wow, so AMD needs 300M transistors and 169mm2 of die area to match what Intel is providing with 228M transistors and 82mm2 of die area!

And Westmere will shrink that die area, offer more performance and reduce system costs with higher levels of integration.

That sounds like a recipe for success. GO AMD!

-VBG

On the flip side Westmere probably won't be hitting <$90 price points right out of the gate.

Has Tom or anyone revealed Westmere pricing? or die size info?

How is that funny?

The plan from the start was to ship the 4S part later than the HEDT or DP part. Just like Tulsa (P4-based) was shipped around the time of Woodcrest and Tigerton wasn't lauched until a year later.

Why would Intel have bothered to design Dunnington if it was planning to launch a 4S Nehalem derivative at the same time as DP?

Your assertion that it will take a year to see meaningful sales of Nehalem is your typical nonsense.

From PSO himself:
http://seekingalpha.com/article/148790-intel-q2-2009-earnings-call-transcript?page=-1

We launched our Nehalem EP server processors at the end of Q1 and in just one quarter from launch, Nehalem already makes up over one-third of all of our dual socket server shipments. We expect the ramp to continue with Nehalem shipments crossing over the 50% market in August.

Or is 30% of the market not "meaningful" enough for you?
Or is Otellini lying to the analysts?

Instanbul better gain some traction while it can. It has a slight performance lead now v.s Dunnington, but from the same source:

We remain on track to launch our Nehalem EX processor line, which is the multi-processor version of Nehalem, in the second half of the year. Nehalem EX is expected to redefine the performance standard in that segment.

-VBG

Commentary: Did TSMC delay high-k--again?

http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=219000144

TSMC appears to have delayed its high-k/metal-gate technology--by at least three quarters or more. Originally, TSMC's high-k/metal-gate offering was slated for the first quarter of 2010, according to the firm.

From 8/4, appologies if this was posted and I missed it.

-VBG

Maybe by "beat" he meant time-to-market?

Our 130nm beat our competitor's 90nm
Our 90nm beat our competitor's 65nm
Our 65nm was on par with our competitor's 45nm


-VBG

Done

Quite the ramp there...

Last year, TSMC rolled out its 40-nm process. In Q1 of 2009, the company's 40-nm process represented about 1 percent of its overall sales, which is better-than-expected. In Q2, TSMC expects to have 2 percent of its overall sales in the 40-nm arena.

I know TSMC has a different business model with a significant percentage of volume on older technologies, but I wonder what this represents in absolute terms like WSPM?

Fixed.

All of your posts smack more of playground politics or adolescent fantasy than the business world...

-VBG

the second largest chipmaker is overpriced?
I think you meant "second largest PC chipmaker"
IIRC, AMD has only been in the top ten right after they bought ATI.

Heck, they produce more processors than IBM!
That's not saying much...

-VBG

65nm

-VBG

ROTFLMOA!! (Wipes away tears...)

Ah, Pete, thanks for the comic relief!

Can you post to this thread more often?

I think chipguy put it best when he said "Correcting you could be a full time job if one let it."
http://siliconinvestor.advfn.com/readmsg.aspx?msgid=24467551

Seriously though, I'll make a prediction to go with yours and we can revisit them in future years. Except you have never been able to go back to a successful past prediction to improve you credibility, so why would I expect that to change?

Anyway, I predict that Intel will ship Westmere (32nm) for revenue in Q4'09. Proof will be either product available in retail by EOY '09 or Quarterly earning conference call where an Intel official states "shipped for revenue".

I will also predict the same for Ivy Bridge (22nm) in Q4'11.

For AMD I will predict that the first availibility of 32nm parts or systems in retail will not occur before Q1'11. No engineering samples in the hands of a review site, no statements by Dirk that "32nm is in production", as these metrics are meaningless. This will put AMD clearly at least 1 year behind Intel at 32nm.

I won't bother to point out that the 22nm announcement from August wasn't a functional chip, it was a functional cell. Oops, I just did. Sorry, but they are not quite the same thing. IBM/AMD is more like 4+ years away from 22nm CPU shipment. Here I will predict Q2'13 for available product.

Cheers!

-VBG

Asus Eee PC Boots Up In 5 Seconds

http://www.ubergizmo.com/15/archives/2008/10/asus_eee_pc_boots_up_in_5_seconds.html

How fast does your notebook or netbook take to boot up? A couple of Intel engineers managed to drop the boot up time of an Asus Eee PC to just 5 seconds, with the Fedora desktop ready and raring to go in that time. The secret of their success? Removing various boot elements that are deemed to be dead weight for netbook users as well as repeating functions, with the crowning glory of a specially written patch ...

-VBG

Linus on Intel SSDs..

http://torvalds-family.blogspot.com/2008/10/so-i-got-one-of-new-intel-ssds.html

"That thing absolutely rocks."

-VBG

That would be just silly. If I had to guess, they are claiming they will use Computational Lithograph techniques at 22nm.

ftp://download.intel.com/pressroom/kits/research/computational_litho_poster.pdf

http://www.solid-state.com/display_article/296374/28/ARTCL/none/none/Intel-promotes-computational-lithography-capabilities/?dcmp=SSTGlobal_ARCH

But that is just a guess.

-VBG

How much does AMD performance improve in moving from Barc to Deneb? How much does Intel in moving from Penryn to Nehalem? Which is the greater improvement factor... I vote for Intel.

I get the sense that Nehalem is going to be a huge improvement on some apps, but not much or none on others, hopefully this is wrong. Deneb will bring 15%-25% improvement, but Nehalem's range will probably be wider.

Regardless of which company "leap-frogs" further, if AMD can push 45nm up and down the product stack faster than Intel, that will improve their competitiveness v.s Intel. And I believe there are reasons Intel can't and won't push Nehalem derivatives down the product stack as fast as AMD might be able to (if they can manufacture them).

I think a little competitve kick for AMD can go a long way right now from two standpoints. One, 45nm may be able to help them be more cost competitve even without any performance benefit. Two, if AMD's product performance gets much worse, I think this would start to impact design wins and they would have to start worrying about getting shut out of accounts again. Dell is already down to a few AMD SKUs.

But I was mainly talking about desktop. Server is not looking good for AMD except for a 4S window that will close with Beckton. I have given up trying to understand the mobile market. Intel has a commanding performance lead and very strong MSS, but it seems that when I look at the Sunday ads, I see lots of AMD laptops and they are not in the bargain basement.

Having said that, desktop is AMDs stronghold at the present, so better competitiveness there is most important.

-VBG

Another data point that 45nm is going to launch at 3GHz.

Looks like another paper lauch from AMD. Launch in December, but under embargo until Jan 8th. To me that says "no product until Jan., if then"

It will be interesting to see how quickly Deneb and Nehalem move down into the mainstream during 2009. I would guess Deneb will be much faster. Deneb will have the advantage of compatibility with an existing infrastructure. AMD will be more motivated (Deneb should be cheaper to produce, if yields are reasonable). And Deneb will still probably be lower performing, so it will not start at the rarified air of the very top price points. Whereas Gainstown will depend on X58 mainboards, DDR3, and Intel will want to keep the halo around Core i7 and will prob be significantly more expensive to produce due to "native" quad-core.

AMD may be at the nadir of their competitive disadvantage now and will make up some ground in 2009, at least till Westmere in 2010, but there are still a lot of unknowns.

BWDIK

-VBG

New intel web site:
http://www.intel.com/pressroom/legal/

About time, IMO.
Intel needs to stop letting AMD control the message.
-VBG

Got to admit, this is brilliant though...

"We've moved to a different pricing cadence," said Buxton. "Now we're making our price moves at the beginning of the last month of each quarter, so don't expect another move until 1st December. This is designed to help the channel to plan in advance and to be as clear as possible about pricing. We should also be in good shape to fulfil the demand we expect in Q4."

Now they can stuf the channel in the last month of the quarter. This also will tend to shift buying patterns toward the end of the quarter so they will have an excuse for lower visibility and late warnings of missed earnings.

-VBG

Dunnington is MP (4-way) whereas Nehalem ES is only DP. But from the standpoint of leveraging existing FSB infrastructure, you are correct. Dunnington is a drop-in compatible replacement for Harpertown in the Caneland platform.

The MP version of Nehalem (EX) is still a year off (only just now showing wafers).

-VBG

you maybe disappointed in the end result. It could be that all the core improvements, faster L2 and faster/wider memory will counter the effects of the slower L1/L3 in all apps but it is not a given as you so confidently expect.

You just keep telling yourself that.

http://www.anandtech.com/printarticle.aspx?i=3326

Preliminary data suggests otherwise however.

-VBG

New Barc 2.5GHz spec results

http://www.theinquirer.net/gb/inquirer/news/2008/05/28/dell-reveals-amd-barcelona

On the two processor SPEC2006 CPU integer tests, AMD's 2.5GHz 2360 SE Barcelona CPUs came in almost 30 points behind Intel's 3.2GHz Xeon 5482 CPUs (92.4 vs 121).

AMD's quad-core Opterons caught up with the Intel Xeons somewhat in the four processor SPEC2006 CPU integer tests, exhibiting better scaling performance there. AMD's 2.5Ghz 8360 SE Barcelona CPUs posted results only 10 points behind Intel's 2.93Ghz Xeon 7350 CPUs (167 vs 177).

But AMD's 2.5GHz 2360 SE quad-core Barcelona CPUs just edged Intel's 3.2GHz Xeon 5482 CPUs in the two processor SPEC2006 CPU floating-point tests (82.1 vs 80.4).

AMD's quad-core Opterons further extended their lead over Intel's Xeons in the four-processor SPEC2006 CPU floating point tests. AMD's 2.5GHz 8360 SE Barcelona CPUs simply outclassed Intel's 2.93GHz Xeon 7350 CPUs (152 vs 108).

Will be interesting to see how Dunnington will compare when it comes out. Checked out Dell's site and I couldn't find 2.5GHz Barcs for sale, not that I tried real hard. I wonder how close we are to seeing Dunnington SPEC scores. Should be getting close to 90 days.

-VBG

Interesting how the side by side comparison shows Shanghai coming before Dunnington. I wouldn't bet on that.

-VBG

Does anyone know what on earth he was thinking, or meant..or said?

Ruiz also said, "And absent a clear and predictable path to profitable growth, we will exit those businesses."

Which I would take to mean they will be exiting the CPU business...

-VBG

Of course. With a population <2000, Mangy-Cours is always popping up on lists of major world cities along with Barcelona, Budapest, Shanghai, Montreal and Istanbul...

-VBG

From the I'll-believe-it-when-I-see-it Dept.:

Not sure if this has been posted yet.

http://www.pcpro.co.uk/processors/news/192804/amd-to-debut-45nm-quadcores-this-quarter.html

-VBG

OT: Journalism
Reading the rest of the parapragh, I don't see the issue.

The $25 million plant in Madison, Pa., will make 40,000 gallons per year. At that size, it's meant to demonstrate the process at commercial scale. Its plans also call for a full-scale facility, producing 50 million gallons to 100 million gallons a year of ethanol, by 2011.

-VBG

Sun Microsystems Buys Montalvo

http://online.wsj.com/article/SB120907829128743161.html?mod=yahoo_hs&ru=yahoo

SAN FRANCISCO--Sun Microsystems Inc. said it has purchased the technology assets of Montalvo Systems, a much-discussed Silicon Valley semiconductor startup.

Terms are not being disclosed because they are not material to Sun's earnings per ...

What is this, buy-a-small-chip-company week?

-VBG

Re: so even if they all strike (and most of them should), the dilutive effect will be much smaller

With RSUs there is no strike price. They are grant of shares of stock, period. If you get a RSU of x shares, on Apr 1st 2008, they vest equally over 4 years. (i.e x/4 on 4/1/09, x/4 on 4/1/10, etc.)

-VBG

http://online.wsj.com/article/SB120792982404908185.html?mod=yahoo_hs&ru=yahoo

Advanced Micro Devices Inc. said Phil Hester has resigned as senior vice president and chief technology officer.

Mr. Hester, a 23-year veteran of International Business Machines Corp., joined the chip maker in 2005. The decision to leave was his decision and was not associated with AMD's recently announced job cuts, said Rob ...

-VBG