GlobalFoundries roadmap slips?

[Salvador]

http://www.eetimes.com/rss/showArtic...0200094&pgno=2

If this is true, and GF has to postpone 32nm some, it won't do much good for competition will it?

[[XC] gomeler]

Aren't we stuck on 45nm with AMD till late 2010 anyways? Waiting on the Bulldozer launch for 32nm right? I'm not familiar with AMD's latest product roadmaps but how the fact that they don't mention how far of a "slip" this is doesn't worry me too much now. Hasn't AMD always been ~1 year behind on process technology, this hopefully won't change a thing.

[EniGmA1987]

Last I heard Bulldozer was pushed back to a probable 2011 timeframe, not that it matters too much. AMD's upcoming server chips seem to be doing REALLY well. If they can possibly take a magny-cour and port it to AM3 then that will tide them over for a while longer.

[[XC] gomeler]

Last I heard Bulldozer was pushed back to a probable 2011 timeframe, not that it matters too much. AMD's upcoming server chips seem to be doing REALLY well. If they can possibly take a magny-cour and port it to AM3 then that will tide them over for a while longer.

One can only hope. Would be amazing if they launched Thuban and an MCM of Thubans at the same time. The only problem is I fear the lack of clockspeeds won't make Thuban and the Magny-cours AM3 alternative competitive. AMD already has trouble with 4 cores at 3+GHz, I can't imagine 6 cores let alone 12 would run cool at an even 3GHz AMD needs a die revision like C0 -> D0 for Intel Bloomfield.

[003]

One can only hope. Would be amazing if they launched Thuban and an MCM of Thubans at the same time. The only problem is I fear the lack of clockspeeds won't make Thuban and the Magny-cours AM3 alternative competitive. AMD already has trouble with 4 cores at 3+GHz, I can't imagine 6 cores let alone 12 would run cool at an even 3GHz AMD needs a die revision like C0 -> D0 for Intel Bloomfield.

Thuban is going to be even less competitive with Westmere than Deneb was with Bloomfield...

[[XC] gomeler]

Thuban is going to be even less competitive with Westmere than Deneb was with Bloomfield...

Do you know something I don't? Westmere's only chip that will compete with Thuban on a thread-basis will be the Core i9 Extreme Edition, estimated at $1500. There have been zero rumors that I am aware of regarding a cheaper launch product featuring Gulftown. The other chips are Clarkdale and Arrandale processors that will be more in line to compete with the Phenom II X3 and X4 processors.

[003]

Do you know something I don't? Westmere's only chip that will compete with Thuban on a thread-basis will be the Core i9 Extreme Edition, estimated at $1500. There have been zero rumors that I am aware of regarding a cheaper launch product featuring Gulftown. The other chips are Clarkdale and Arrandale processors that will be more in line to compete with the Phenom II X3 and X4 processors.

Well it's already been confirmed there will be an Xeon CPU based on Gulftown with a locked multiplier, so no EE style pricing. Also there has been speculation that Intel may release a non-EE version of Gulftown some time after the initial launch.

[zalbard]

Well it's already been confirmed there will be an Xeon CPU based on Gulftown with a locked multiplier, so no EE style pricing. Also there has been speculation that Intel may release a non-EE version of Gulftown some time after the initial launch.

Ummmm, link?
ES samples of Gulftown have unlocked multi...

[003]

Ummmm, link?
ES samples of Gulftown have unlocked multi...

Well there's this:
http://www.xtremesystems.org/forums/...45#post4000645

[zalbard]

Well there's this:
http://www.xtremesystems.org/forums/...45#post4000645

Rofl, and you call it a source?
Do you realise that BIOS update for Gigabyte board unlocked that CPU?
And it's just a ES, that's no way an official confirmation of anything.
Have a look here: http://www.xtremesystems.org/forums/...d.php?t=232109
This guy unlocked his CPU after updating the BIOS.

[003]

Rofl, and you call it a source?
Do you realise that BIOS update for Gigabyte board unlocked that CPU?
And it's just a ES, that's no way an official confirmation of anything.
Have a look here: http://www.xtremesystems.org/forums/...d.php?t=232109
This guy unlocked his CPU after updating the BIOS.

I really hope there is a non-EE version of Gulftown released. Or at the very least, a quad core version of Westmere (and no on-die graphics).

If not, Sandy Bridge will be my next upgrade.

[Hans de Vries]

http://www.eetimes.com/rss/showArtic...0200094&pgno=2

If this is true, and GF has to postpone 32nm some, it won't do much good for competition will it?

This is the "old" roadmap which says: Risk production in mid 2010.




Left edge = tape out; Right edge = risk production.....

Bad reading capabilities of the journalist become news items...


Regards, Hans

[JumpingJack]

This is the "old" roadmap which says: Risk production in mid 2010.




Left edge = tape out; Right edge = risk production.....

Bad reading capabilities of the journalist become news items...


Regards, Hans

So if I understand this graph correctly, 45 nm risk production should start in a few more months

[zalbard]

I really hope there is a non-EE version of Gulftown released. Or at the very least, a quad core version of Westmere (and no on-die graphics).

If not, Sandy Bridge will be my next upgrade.

Yeah, man, I am hoping for that as well.

[Nedjo]

So if I understand this graph correctly, 45 nm risk production should start in a few more months

nop you're reading it wrong: GloFo is existing since Q1 this year, and for them 45nm production started then. But I do admit that this type of timeframe scale is misleading, so one can't blame journalist... And regarding the "slip"... here's quote from that article:

GlobalFoundries denied its roadmap has slipped. ''Our roadmap for 32-nm SOI has not slipped,'' according to a spokesman for the company. ''Yes, the timeframe for introduction has been altered slightly compared to the roadmap we showed you in July, but that is not because of any issues with the technology. The roadmap has simply been adjusted to align with AMD's product needs. We have solid natural yields ramping up every week, and we have high confidence in our ability to demonstrate the same robust yields and manufacturing capability on 32-nm that we have historically had.''

[JumpingJack]

nop you're reading it wrong: GloFo is existing since Q1 this year, and for them 45nm production started then. But I do admit that this type of timeframe scale is misleading, so one can't blame journalist... And regarding the "slip"... here's quote from that article:

I was just teasing Hans ... those type of gant charts are actually pretty useless for any kind of real information.

[ajaidev]

Where is the new roadmap?? The link has only word no pictures

2010 forecast:-

Thuban vs Nehlem/Westmere "Clock to clock Westmere would win, Nehalem and Thuban would equal each other in most cases except single threads. Price to performance of Thuban is a important factor, if its closer to the price of Nehalem than Westmere we may have a winner, but Westmere is in a league of its own who want performance no matter the costs they would buy Westmere with that ultra sexy stock cooler "

Deneb vs Lynnfield "Clock to clock Lynnfield would win, but faster Denebs based on new stepping and platform costs is one of the important points"

Propus/Rana vs Clarksdale "Clock to clock Propus would win, Rana would be about equal to Clarksdale in performance. Since this is Mid-low end price matters a lot, Intels offerings are on a higher side but we have to wait and see how much does the H55 costs. P55 paired up with a i3 you lose the igp and also have to buy a new video card"

Regor vs Core 2 Duo "Clock to clock C2D wins but it depends a lot which C2D CPU you are using. L2 is very important in C2D, a E7xxx may bash Regor really well bu E5xxx cant do that good a job. Pricing basically defines this low end segment.

[saaya]

extended 45/40nm lifetime... they make it sound as if the market would demand that... why and how?
whats odd about the roadmap is that 40/45 lp is coming AFTER 32nm? does that make much sense...
even if its not optimized for low power, shouldnt 32nm be about the same for low power asics as 40/45 lp optimized?

[ubuntu83]

Propus/Rana vs Clarksdale "Clock to clock Propus would win, Rana would be about equal to Clarksdale in performance. Since this is Mid-low end price matters a lot, Intels offerings are on a higher side but we have to wait and see how much does the H55 costs. P55 paired up with a i3 you lose the igp and also have to buy a new video card"

Clarkdale has an integrated GPU so why would it not function when used in P55 boards?

[Hans de Vries]

So if I understand this graph correctly, 45 nm risk production should start in a few more months

You must have heard Otellini's claim last Tuesday then:

"Two years ago Intel introduced the world's first 45 nanometer process, and to date the competition - that is to say AMD - has shipped zero."



http://www.tgdaily.com/content/view/44057/135/
http://www.tgdaily.com/content/view/44061/103/


Regards, Hans

[ajaidev]

Clarkdale has an integrated GPU so why would it not function when used in P55 boards?

Simple P55 has no VGA, DVI, etc ports h55 has them. You need to get h55 with FDI or no video interface on mobo is provided. P55 + i3 means you cant use the integrated igp.

[v0dka]

You must have heard Otellini's claim last Tuesday then:

"Two years ago Intel introduced the world's first 45 nanometer process, and to date the competition - that is to say AMD - has shipped zero."



http://www.tgdaily.com/content/view/44057/135/
http://www.tgdaily.com/content/view/44061/103/


Regards, Hans

What the hell?

[SocketMan]

Aren't we stuck on 45nm with AMD till late 2010 anyways? Waiting on the Bulldozer launch for 32nm right? I'm not familiar with AMD's latest product roadmaps but how the fact that they don't mention how far of a "slip" this is doesn't worry me too much now. Hasn't AMD always been ~1 year behind on process technology, this hopefully won't change a thing.

I am no road map expert but I thought AMD was shipping their 4770 Radeons
since spring 2009,these are made on 40nm if my memory serves correctly.

[Lightman]

I am no road map expert but I thought AMD was shipping their 4770 Radeons
since spring 2009,these are made on 40nm if my memory serves correctly.

But they are manufactured at TSMC and not GlobalFoundries ....

[Nedjo]

But they are manufactured at TSMC and not GlobalFoundries ....

so? It's still competition...