Intel shows running 32nm chips.

[Logos]

don't know if the link's been posted already, here's an Anandtech article:
http://www.anandtech.com/cpuchipsets...oc.aspx?i=3513

[terrace215]

i5 is a mainstream quad. Plus C2Q. And who says you cant upgrade your i5 quad? Its just not gonna be with westmere.

This is wrong. You assume that the list of Westmere parts on that slide is exhaustive. It isn't. Intel themselves said there will be additional Westmere parts coming in 2010, and Anandtech (among others) said one of those will be a QC i5-equivalent part.

With 45nm i5 yet to launch, even, Intel has every reason not to explicitly talk about its successor just yet, lest it encourage would-be buyers to wait for the Westmere version.

[Particle]

That 6C/12T 32nm looks yummy.

[Shintai]

This is wrong. You assume that the list of Westmere parts on that slide is exhaustive. It isn't. Intel themselves said there will be additional Westmere parts coming in 2010, and Anandtech (among others) said one of those will be a QC i5-equivalent part.

With 45nm i5 yet to launch, even, Intel has every reason not to explicitly talk about its successor just yet, lest it encourage would-be buyers to wait for the Westmere version.

No..and from Anandtech.

Keep following; if you want a quad-core Westmere, your only option will be in the LGA-1366 socket with Gulftown. Core i7 will get replaced with a six-core, twelve-thread processor in early 2010. There won’t be a 32nm quad-core part on the desktop until the end of 2010 with Sandy Bridge.

Not even to mention the Intel slides.

[hairy]

that's wrong though. gulftown is six cores.

[AuDioFreaK39]

ahaa! I knew that face in the background looked familiar.

[Drwho?]

ahaa! I knew that face in the background looked familiar.

But Steve, on the front of me is one of the brains behind this amazing break through, and the man who plays with Atoms is Mark Bohr:


Working and meeting those men is the reason I love working at Intel.

Think about it, if you take a red cell, you can put many hundreds of transistors and their interconnect in the diagonal of the red cell.
Those men create technology smaller than what Nature was able to do, smaller than the smallest Cell.

Those men are Material magicians, too many people forget about this.

[ghost_recon88]

Drwho? are we gonna get a 32nm sneak peak anytime soon?

[Drwho?]

Drwho? are we gonna get a 32nm sneak peak anytime soon?

no no no no no no ...

[Nanometer]

The Classic Intel HSF with the push through hole design. It......Must......DIE !!

Perkam

I agree, it is not a good design. Feels cheap and it isn't that easy to handle.... and the placement of the holes makes it prone to board flex..i hate board flex.

[kiwi]

I think the question was mostly due to Cinebench being one of the benches that scales with basically everything you throw at it, be it clocks, cores, ipc improvements or HT, so the winner was rather clear from the start.

Exactly But since this is the first demonstration, impressing media is important. So let's wait for ES and early results and see

[Donnie27]

Still not bad for an early Demo. It is good that it was Dual Core vs. Dual Core but I'm not so sure the right DC was picked. I'd much rather have seen it vs a Wolfdale. Now if they mean they were using something in the same price range, then that's outstanding. If not, it then become WTF were you thinking Intel?

[Spacemaster]

no no no no no no ...

why not

[ghost_recon88]

no no no no no no ...

Was worth a shot

[savantu]

i5 is a mainstream quad. Plus C2Q. And who says you cant upgrade your i5 quad? Its just not gonna be with westmere.

...

That's BS.

What's the point for a Quad-Core Westmere in the next 12 months when i5 is being launched late this year ? 4 core Westmere will happen , just that it will probably ship in H2 2010.

Imagine this - Why not Westmere at the high end with 45nm I5s as the bulk in 2010 ? Later , Sandy Bridge as high end with 4-6 core Westmere as bulk in 2011.
It is tick-tock slightly altered allowing Intel to gain maximum possible from a n-1 process by extending its lifespan.

45nm will rule the middle for 3 years allowing Intel to get the $80B envisioned from 45nm products.Intel brought 4 45nm FABs online.You need huge volume to get ROI on that.While its US FABs are converted to 32nm, Intel needs to keep at least one huge 45nm FAB , FAB28 , full , thus it needs to manufacture large amounts of 45nm products ( Atom is close to irrelevant at over 2000 per wafer ) like CPUs and GPUs.

[Hornet331]

I agree, it is not a good design. Feels cheap and it isn't that easy to handle.... and the placement of the holes makes it prone to board flex..i hate board flex.

Major OEM disagree with both of you (and so do I), i never had a problem with pushpins and its more then adequate for the stock heat sink, its not intels fault that 3rd party cooler manufacturer use push-pins on there 2kg + tower coolers.

[Donnie27]

Major OEM disagree with both of you (and so do I), i never had a problem with pushpins and its more then adequate for the stock heat sink, its not intels fault that 3rd party cooler manufacturer use push-pins on there 2kg + tower coolers.

Yes and the boards are built to withstand some flexing anyway.

[Donnie27]

That's BS.

What's the point for a Quad-Core Westmere in the next 12 months when i5 is being launched late this year ? 4 core Westmere will happen , just that it will probably ship in H2 2010.

Imagine this - Why not Westmere at the high end with 45nm I5s as the bulk in 2010 ? Later , Sandy Bridge as high end with 4-6 core Westmere as bulk in 2011.
It is tick-tock slightly altered allowing Intel to gain maximum possible from a n-1 process by extending its lifespan.

45nm will rule the middle for 3 years allowing Intel to get the $80B envisioned from 45nm products.Intel brought 4 45nm FABs online.You need huge volume to get ROI on that.While its US FABs are converted to 32nm, Intel needs to keep at least one huge 45nm FAB , FAB28 , full , thus it needs to manufacture large amounts of 45nm products ( Atom is close to irrelevant at over 2000 per wafer ) like CPUs and GPUs.

That's what I think too! I only see i3 or 45nm Dual Core being dumpped for a Dual Core version. I too think Intel will ship millions of 45nm Quad Core Processors. These can also be something dirt cheap to build and they can get tons of them per wafer. Even at a moderate price, Intel still makes a killing on them. Unlike some current Dual Cores, the 32nm are Native Dual Core and very modular, right? They can ship ultra cheapie with nothing but the Core, slap on a GPU and etc...., run a couple in tandem

I've consistantly read info saying Intel would ship i5 for the before the back to school rush.

http://www.dailytech.com/Intel+Targe...ticle14228.htm

With new sockets will come new chipsets. Intel will expand their 5 series of chipsets with new models for consumers and businesses. All Nehalem and Westmere based products use DDR3.

The Q57 chipset, codenamed Piketon, is targeted at businesses, while the P55 chipset, codenamed Kings Creek, is targeted at consumers.

[Drwho?]

I am surprise that none of you guys though about this:

if you put some Ln2 on the CPU cores ... guess who is going to get some LN2 too ... and with Nehalem memory controlers, somebody will have a lot of fun ...

sorry for teasing ...

[freeloader]

I am surprise that none of you guys though about this:

if you put some Ln2 on the CPU cores ... guess who is going to get some LN2 too ... and with Nehalem memory controlers, somebody will have a lot of fun ...

sorry for teasing ...

LN2 on what's essentially integrated graphics....why?

[informal]

I am surprise that none of you guys though about this:

if you put some Ln2 on the CPU cores ... guess who is going to get some LN2 too ... and with Nehalem memory controlers, somebody will have a lot of fun ...

sorry for teasing ...

I don't get it . You're gonna put "some LN2" on a hybrid MCM chip,basically a dual core+ crappy GMA gpu? What records will that break,even if it reaches the "magical" (for intel's i7) 6Ghz??
I understand if 6 core westmere was put under ln2 and clocked to 6GHz,that machine would surely break some records. The MCM you linked in that image ,under LN2 ,could only beat a 3650 radeon+ wolfdale @on air.And that's about it.

[T_Flight]

no no no no no no ...

Is the NDA under NDA again?

[Drwho?]

I don't get it . You're gonna put "some LN2" on a hybrid MCM chip,basically a dual core+ crappy GMA gpu? What records will that break,even if it reaches the "magical" (for intel's i7) 6Ghz??
I understand if 6 core westmere was put under ln2 and clocked to 6GHz,that machine would surely break some records. The MCM you linked in that image ,under LN2 ,could only beat a 3650 radeon+ wolfdale @on air.And that's about it.

you assume it is crappy ... that is your mistake
it was running spore at medium setting during the presentation over XXfps ... (Censured )

Most of integrated graphics need more BW, here , you have a mem controler that is the Nehalem one ... plus/Minus some details ... (not talking)

The same people were telling me I was a lier when we showed Conroe numbers for the 1st time, we all know how it ended up.
Don't assume we are stupid, it take time to understand, and then, even more time to design the chip. we understood 3 years ago, here we go!

You may get surprise...

[Drwho?]

Is the NDA under NDA again?

no no ... the NDA lift date is under NDA, but since they is not NDA yet for the press on this one, this is an even stronger NDA of Death ...

[Movieman]

you assume it is crappy ... that is your mistake
it was running spore at medium setting during the presentation over XXfps ... (Censured )

Most of integrated graphics need more BW, here , you have a mem controler that is the Nehalem one ... plus/Minus some details ... (not talking)

The same people were telling me I was a lier when we showed Conroe numbers for the 1st time, we all know how it ended up.
Don't assume we are stupid, it take time to understand, and then, even more time to design the chip. we understood 3 years ago, here we go!

You may get surprise...

Here, I'm going to bail you out!

Guys, this is no BS..
Yesterday I'm talking with a guy whose working on a 32nm system..testing,etc..No, no one from here..
His words to me:
"I have never seen a system at this starting stage that was so smoothe, and the power you simply won't beleive"

That's a quote, he wouldn't give me numbers.
I asked politely and he said he couldn't..
Now that Wolftown system looks interesting.