Sign in with
Sign up | Sign in
Your question

Nehalem to cost $1100,

Last response: in CPUs
Share
April 2, 2008 9:18:28 AM

"We learned that Nehalem is the most significant and yet most dangerous marchitecture move that Intel made in the last five years.

It sure is a holy grail of AMD’s ideas wrapped around an Intel product and having the integrated memory controller and L3 cache in a single processor makes a lot of space for errors.

We don’t say that Nehalem will have problems or that it will be late, but we can say that Intel is risking like never before and most of the stuff it uses has been seen in AMD’s CPUs. The only big advantage Nehalem has over Denab is its Hypertheading support but this is something to be seen to believed.

Nehalem will launch as a single $1100 CPU in Q4 2008 and it will find its place in server market but it will take until middle 2008 to take the significant market share out of the Penryn core quad / duo generation. Intel tends to make our life easier as it tends to repeat the same pattern all over again."

http://www.fudzilla.com/index.php?option=com_content&ta...

...no point waiting then if your thinking of upgrading anytime soon.

More about : nehalem cost 1100

April 2, 2008 9:43:20 AM

Wow, that's expensive. And those CPUs aren't Extreme Edition, they are meant for servers.
a c 127 à CPUs
April 2, 2008 3:26:38 PM

Wow. Fudzilla needs to do some research. Intel developed a IMC abefore just decided it wasn't needed yet. They also have had a integrated L3 cache on the first P4 EE 3.4GHz Northwood, this I can vouch for as I have one.

Now yes it is stuff that AMD has used and we don't know if there will be any errors or not.

But the price makes sense for a server chip. They are normally more expensive. And I am sure the first end user chip will be a EE. Or maybe not since it might not need OC'ing anymore.

I just want to see the specs and performance first. Thats what interest me the most.
Related resources
a b à CPUs
April 2, 2008 3:33:59 PM

Wow, this will be a major switch. As jimmy pointed out, they have used some of these technologies before, but never in a mainstream/major chip. I'm definitely not saying they are going to screw it up, but I wouldn't be surprised if supply is low and cost is high at first. As with any new core there will be bugs to work out and sectors to be optimized. I guess we shall see.
April 2, 2008 3:51:14 PM

Still waiting on the next generation before I up grade, maybe then games will highly multi-threaded.
April 2, 2008 4:13:11 PM

The "original" IMC chip was more like a 486 lol, Nothing near what todays chips are like, so, no, its a huge step for Intel, and cant even be compared. Also, this chip is rated at 2.53Ghz? Are they having problems getting better speeds using a true quad arch?
April 2, 2008 4:17:26 PM

Half the news reported on Tom's Hardware is barely believable.
How can anyone believe anything from a website named fudzilla?
Let's check the meaning of FUD.
a c 127 à CPUs
April 2, 2008 4:20:20 PM

JAYDEEJOHN said:
The "original" IMC chip was more like a 486 lol, Nothing near what todays chips are like, so, no, its a huge step for Intel, and cant even be compared. Also, this chip is rated at 2.53Ghz? Are they having problems getting better speeds using a true quad arch?


No it was Timna which was based off of the Pentium III Coppermine.

And haven't we learned that speed isn't everything? Plus when C2D first came out it debut at 2.66GHz if I remember correctly.

spoonboy said:
http://forums.vr-zone.com/showthread.php?t=257519

(source) http://www.hwupgrade.it/articoli/cpu/1938/intel-develop...

according to this it will be at Phenom-like core clock speeds, so no revolution in single/partially threaded apps/games.


Um this might not be true. I need to get time to read the article but Nehalem is supposed to have dynamic OC'ing built in. This meaning that in order to help complete a single threaded task, the other cores will down clock and the core the thread is on will OC giving the CPU the boost it needs to complete the task faster.

Although this may have to be seen in reality first. But could you imagine a 3GHz chip OC'ing one core to 3.6GHz while the others lower to 2.5Ghz in order to get a thread completed? Might improve gaming performance more than you think.
April 2, 2008 4:29:29 PM

I wasnt referring to timna , I was referring to this "The interesting thing about Intel's original decision to build Timna is that the company had tried to create an integrated product before and it failed, Brookwood says.

About ten years ago the company created an integrated product line, Brookwood notes. Intel shipped both the 386SL and a related 486SL after some delays. But vendors stayed away, and by 1992 Intel had cancelled the line.

" But neither really came to market, and never was Intel successfull with it, so this is a huge move, and that wasnt my question about IPC, it was what it was, is Intel having problems running up the clock speeds? Using a true quad arch?
April 2, 2008 4:33:30 PM

Another application to benefit from that dynamic OC thing would be audio encoding... which to my dismay remains to be LARGELY single threaded. The only (pratical) workaround to load a multicore chip is to launch multiple "workers" to encode seperate tracks.

In anycase, I look forward to the next generation of chips (from either camp)... Its going to be one hell of a step up from my x2 3800+

JAYDEEJOHN said:
Using a true quad arch?

BLASPHEMY!!! replace that "true quad arch" bologna with something more... correct. Monolithic Quad... for example...
April 2, 2008 4:37:55 PM

Wow! Should be true, it's from FUDzilla!

Do you hear the irony in my comment!?!?!?

Like every unannounce products, wait for the real announcement before making a decesion... and never forget that the best time to buy is WHEN YOU HAVE THE MONEY FOR IT!

Quite simple... no?
April 2, 2008 4:55:52 PM

JAYDEEJOHN said:
My question holds relavence, cause since when has Intel shown a new cpu at mid clock rates? http://translate.google.com/translate?sourceid=navclien... This is the reference from FUD. It looks interesting


It still holds 0 relevance.

#1) Clock Speeds are no relevant between CPU designs as can be seen between Core2 and P4. Core2 is generally about twice as fast at the same clock speed.

#2) This is a OCTO not a QUAD processor designed for servers. As a result, # of Cores is more important that Max Clock Per CPU. Intel will also be releasing Quad Core Nehalm chips.

#3) The speed shown is necessarily the top speed Intel can currently generate and even if it was, it's not highly relevant since this was 6-9 months prior to expected ship dates.

:pfff:  Gotta play with your toy tops more, you like to try spinning but you are not quite there yet.
April 2, 2008 5:05:13 PM

Again, no ones answered my question. Whens the last time Intel showcased a cpu at mid level speeds? And are the Octos supposed to be that much lower clocked? Abd umm forget the conspiracy theory here, these are real questions, some people may actually know the answers, so dont go getting paranoid, ok?
April 2, 2008 5:16:29 PM

And as for the octo "This screen was taken from a dual Socket Nehalem system with two processors each quad core capable of 8 threads simultaneously for a total of 16 core processor available for the system. La workstation è impegnata in una simulazione di fluidodinamica, occupando tutti e 16 i core di processore al massimo delle loro possibilità. The workstation is engaged in a simulation of fluid dynamics, occupying all 16 core processor to the maximum of their ability.
" http://translate.google.com/translate?sourceid=navclien... So whos doing the spinning?
a b à CPUs
April 2, 2008 5:22:10 PM

Frist of all Intel did show anything except that fire video from the other day? All the rest is just typical fudzilla crap. Did Intel release the info on what cpus/what clock speeds/what core count/IMC or non IMC/socket types will be availble at launch?. Is there even a launch date?. Intel didnt showcase anything.

The actual GHZ of these chips means NOTHING. As in terms of how it will perform.
a b à CPUs
April 2, 2008 5:30:32 PM

And in that article it is using the age old FSB and mulit's formual to come to conslusions on the chip. Very interesting.

A frequency of 2.13 GHz is obtained by combining a bus to 1,066 MHz with frequency multiplier equal to 8x, while the 2.53 GHz is obtained with the same frequency but using multiplier equal to 9.5 x: we do not know how technology QuickPath Interconnect is recognised by the system and this contributes to the generation of clock frequency of the final processor, so these our deductions on the relationship between bus frequency multiplier and could be incorrect.

Basically just relax, Wait to Intel releases details on the cpus. Wait till they actually have ES cpus out and about(and mobos for em).

April 2, 2008 5:41:35 PM

Thats true, tho this Italian site got some good info on it, and yes, this isnt there uaual presentation. Its just been my experience that Intel, ATI, nVidia and AMD usually showcase their best products no matter what the presentation its in, ala the 3Ghz Barcelona. I know that Intel wouldnt showcase a cpu that it wouldnt sell, or couldnt make/sell, so Im not going there
April 2, 2008 5:45:13 PM

This is news, new news. Maybe great news. But watering it down doesnt make it more or less.
April 2, 2008 6:15:32 PM

Sorry to blow your bubble JAYDEEJOHN, but I remember AMD showing off 1.2GHZ A64 a year before it's real launch. All they want to do is show that the product works, altough there is still stepping to go to get it to work at the intended frequency.

Also, Anandtech (side news bar: http://www.dailytech.com/article.aspx?newsid=11350 ) mentionned totay that the NEHALEM shown (a quad-core tough) was working at 3.2GHZ. Not bad at all for a product that's gonna be available at the end of the year. Not bad at all!

April 2, 2008 6:19:45 PM

Cool , Ill look it up. And dont worry, the skin on this bubble is tough enuff thanks
a c 100 à CPUs
April 2, 2008 6:28:07 PM

JAYDEEJOHN said:
I wasnt referring to timna , I was referring to this "The interesting thing about Intel's original decision to build Timna is that the company had tried to create an integrated product before and it failed, Brookwood says.

About ten years ago the company created an integrated product line, Brookwood notes. Intel shipped both the 386SL and a related 486SL after some delays. But vendors stayed away, and by 1992 Intel had cancelled the line.

" But neither really came to market, and never was Intel successfull with it, so this is a huge move, and that wasnt my question about IPC, it was what it was, is Intel having problems running up the clock speeds? Using a true quad arch?


The 386SL certainly had an on-package memory controller, although it is not clear whether it was an on-die or MCM arrangement. There was also external, on-die cache. I tried to find a 386SL datasheet, but they were taken off Intel's site long ago and don't seem to be available anywhere else either, perhaps due to the relative rarity of the processors themselves. But from what I was able to find, the 386SL and it's companion chip were more of an early attempt at a low-power SoC for inexpensive laptops than anything else.
April 2, 2008 6:29:20 PM

NightlySputnik, I want to thank you for answering my questions, and the link. Like I said, it is what it is. I didnt ask these questions to cause controversy, the the response may cause some themselves, after my questions. Im not spinning. Im not making a quad into an octo , its just that Im very interested in this. If anyone asks a question on here, is this appropriate answers, or responses? Again , I thank you, and it does look like great news to me
April 2, 2008 6:35:03 PM

Look, I think everyone else needs to relax a bit. If you follow that link to Anands, youll see the first questions asked were the same as mine. Theres some really uptight people here
April 2, 2008 8:26:04 PM

same pricing as intel has used for the last decade

release the extreme at $1000

what else is new?

05 EE840 2 singel core cpu in one package - junker then lead to the quad

06 EE955 and EE965 - 965 is killer chip running easy at 4.4ghz in air - dual core on single die

qx6700 nov of 06 first quad

qx9650 nov of 07 first new quad 45nm
a b à CPUs
April 2, 2008 8:42:39 PM

zpyrd said:
Half the news reported on Tom's Hardware is barely believable.
How can anyone believe anything from a website named fudzilla?
Let's check the meaning of FUD.

+1

[:turpit:2]
April 2, 2008 10:20:45 PM

all cpu's run the same with in small region

e2180 e4500 or e6600 or e6700 all oc in the same region, same as pentium

same with the new design - i expect 4ghz
a c 127 à CPUs
April 2, 2008 10:57:05 PM

I have heard that Nehalem is supposed to be able to easily hit 5GHz. Maybe yes maybe no. I would probably see that with the 32nm shrink more than th 45nm but we never know right?
a b à CPUs
April 3, 2008 1:27:17 AM

^ 5Ghz would be sweet.
April 3, 2008 1:27:55 AM

humm... Fuddo...

He has the credibility of similar to Charlie D. Ok I lied, maybe Fuddo is slightly better.

The roadmap released by Intel Japan puts Bloomfield around the price of Q9550 upon launch in Q4. Of course this roadmap was released in Q407, so the time of launch may be different. However I really doubt the price will change.
April 3, 2008 1:31:29 AM

Nehalem running at 3.2GHz at IDF. Hexus also predicts a 3.6GHz Nehalem CPU this year.

Remember that this is only A1 silicon as well, there's still six to eight months before we'll see Nehalem.

Claiming Nehalem will cost $1100 at launch based on one Fudzilla article is just stupid.
April 3, 2008 1:52:23 AM

i dont believe that article...
a c 100 à CPUs
April 3, 2008 1:59:52 AM

dragonsprayer said:
same pricing as intel has used for the last decade

release the extreme at $1000

what else is new?

05 EE840 2 singel core cpu in one package - junker then lead to the quad

06 EE955 and EE965 - 965 is killer chip running easy at 4.4ghz in air - dual core on single die

qx6700 nov of 06 first quad

qx9650 nov of 07 first new quad 45nm


The 840EE was actually two cores on one die. The die consisted of two electrically separate single cores that were just cut out of the wafer as a pair rather than separated. Intel did that as they didn't have time to refit the CPU package to accomodate two pieces of silicon if they wanted to beat the Athlon 64 X2 to market (which they did by a couple of weeks.) The Pentium D 900s were two physically separate pieces of silicon as Intel had time to refit the CPU package to accommodate them.

The first "real" dual core on a single die Extreme Edition was the Core 2 Extreme X6800.

EDIT: dang spelling...
a c 127 à CPUs
April 3, 2008 2:00:01 PM

Mandrake_ said:
Nehalem running at 3.2GHz at IDF. Hexus also predicts a 3.6GHz Nehalem CPU this year.

Remember that this is only A1 silicon as well, there's still six to eight months before we'll see Nehalem.

Claiming Nehalem will cost $1100 at launch based on one Fudzilla article is just stupid.


Saw that. Thats just pretty hehe. And you are right. Its just A1 silicon and we will probably see a B or C stepping when we get it. 3.6GHz seems possible. Look at the current processors on 45nm. Dual cores 4GHz on air. Quada can get to 3.6GHz on air.

The $1100 seems feasable for a server chip. But even a EE will come out in tray prices for $999 like usual and be price gouged by the re/e-tailer.

Man now I wish I had a lot of money so I can build one when it comes out. But I wanna wait for a 8 core. That way I have 16 threads. Imagine a game that could offload everything else to the CPU and just the graphics are being processed throught the GPU. Man that would be awesome. Maybe Nehalem will help run Crysis at a good frame rate?
April 3, 2008 3:48:29 PM

"3.6GHz seems possible. Look at the current processors on 45nm. Dual cores 4GHz on air. Quada can get to 3.6GHz on air."

Its another matter going from a dual-core die to a 6 or 8 core die. Cost and heat will be a much larger factor in nehlamem than for current intel cpus.
a c 127 à CPUs
April 3, 2008 3:56:07 PM

spoonboy said:
"3.6GHz seems possible. Look at the current processors on 45nm. Dual cores 4GHz on air. Quada can get to 3.6GHz on air."

Its another matter going from a dual-core die to a 6 or 8 core die. Cost and heat will be a much larger factor in nehlamem than for current intel cpus.


Yes but it will start with a quad core. So I say let us focus on that and then look towards Dunnington(might be spelled wrong), the 6 core, and then the 8 core.

From what we can see on the 45nm process for quad cores, if you have the right multiplier its easy to get them from 3GHz-4GHz.

But we shall see.
April 3, 2008 4:21:17 PM

JAYDEEJOHN said:
NightlySputnik, I want to thank you for answering my questions, and the link. Like I said, it is what it is. I didnt ask these questions to cause controversy, the the response may cause some themselves, after my questions. Im not spinning. Im not making a quad into an octo , its just that Im very interested in this. If anyone asks a question on here, is this appropriate answers, or responses? Again , I thank you, and it does look like great news to me


It's all good. I hope I wasn't rude, it wasn't my intention.

That's the problem with writing, you get only less then 10% of the message. :) 

About Nehalem, it looks good from the "tone of voice" Intel is using talking about it. Still, we'll all know for sure only at the end of the year.
April 3, 2008 4:27:00 PM

JAYDEEJOHN said:
The "original" IMC chip was more like a 486 lol, Nothing near what todays chips are like, so, no, its a huge step for Intel, and cant even be compared. Also, this chip is rated at 2.53Ghz? Are they having problems getting better speeds using a true quad arch?

Please elaborate.
Thanks!
April 3, 2008 4:29:25 PM

Don't know why you would doubt that nehalem will cost $1100 when it first comes out. The 45nm quads were over this price when they first came out, hell, it's still hard to find the main stream 45nm quads. I woudln't expect an affordable nehalem until 3-4 months after the release of the the high end product.
April 3, 2008 4:55:21 PM

By Q2 09 there will probably be an affordable Nehalem chip.
April 3, 2008 5:42:33 PM

He was only off by a small amount, in this case one multiplier increment, it debuted at 2.93GHz with the Core2Extreme X6800, but the highest non-extreme launch CPU was the E6700 at 2.66GHz.
April 3, 2008 5:44:40 PM

Quote:
JimmySmitty, you really think C2D debuted at 2.66ghz??? LOOOOOOL
Now thats some funny chit.


says a person who has no clue... ;) 
a c 127 à CPUs
April 3, 2008 7:10:17 PM

Gravemind123 said:
He was only off by a small amount, in this case one multiplier increment, it debuted at 2.93GHz with the Core2Extreme X6800, but the highest non-extreme launch CPU was the E6700 at 2.66GHz.


Thanks for clearing that up. I thought it was 2.93GHz but I think in the non EE not overpriced arena, which by the way MrsBytch I was right so :kaola: 
April 3, 2008 7:21:05 PM

jimmysmitty said:
Wow. Fudzilla needs to do some research. ~snip~


Since when have Fudo let truth get in the way of a story? :whistle: 
April 5, 2008 6:34:33 PM

i have a question, if Nehalem has no FSB how do you overclock it?
a c 100 à CPUs
April 5, 2008 11:13:28 PM

tipoo said:
i have a question, if Nehalem has no FSB how do you overclock it?


You increase the system base clock rate or the multiplier, just like how you would overclock the FSB-less AMD K8s and Phenoms.
!