Will the Yorkfield 45nm native Quad Core CPU's be LGA775?

[Psinet]

I have read that Intel are planning to introduce another CPU socket architecture - Socket B or LGA1366, for the Bloomfield CPU's. I want to find out what is likely to be the fastest processor available for my motherboard, and what I want to know is this - does anyone have any info on whether Intel will be releasing the next gen of Yorkfield 45nm native quad core cpu's on the LGA775 socket? I have a Gigabyte GA965P-DQ6 mobo that is "Quad Core Ready"", but I wonder if that is going to stretch to the native quad cores which will only apear under 45nm fabrication processes. I have read somewhere on a forum that some guy thought they would need a new mobo, but if the Bloomfield Socket B doesn't come in until mid 2008, and the Yorkfield CPU's are supposed to be released well before the end of this year........doesn't that mean they will be LGA775?

 

[enewmen]

Yorkfield will run on socket 775., That does not mean the p965, 975x, or 680i will support it. ( The 680i is more likely because of its 1333 fsb). That depends on the voltage and BIOS.
The x38 will support yorkfield. There are lots of threads on this here at THG.

 

[papi4baby]

Well with the latest bios, the DQ6 officially supports FSB 1333, so maybe it will, we will just have to wait and see. Here's to hoping, since i have a DQ6 too

 

[Psinet]

Yeah the new F9 BIOS revision (11th Jan 2007) on the DQ6 Rev 2.0 board allows for selection of FSB settings of up to 2800 using the Memory Multiplier, multiplier of 9X and a CPU host clock of 700 (6.3ghz!!!) lol heres hoping.

Gigabyte have been really impressing me with their BIOS development of the DQ6 - the F8 was only 2 weeks before the F9, and the F7 was only 6 weeks before that. And I havent had any troubles with any of them, but they sure do keep adding functionality. For those of you who dont know, press cntrl-F1 at the main BIOS screen and a tonne of new options will appear in each screen.

It sure would be ice to be able to drop a 45nm native quad core Yorkfield alongside 4gb of new ram into it.

 

[qcmadness]

Yeah the new F9 BIOS revision (11th Jan 2007) on the DQ6 Rev 2.0 board allows for selection of FSB settings of up to 2800 using the Memory Multiplier, multiplier of 9X and a CPU host clock of 700 (6.3ghz!!!) lol heres hoping.

Gigabyte have been really impressing me with their BIOS development of the DQ6 - the F8 was only 2 weeks before the F9, and the F7 was only 6 weeks before that. And I havent had any troubles with any of them, but they sure do keep adding functionality. For those of you who dont know, press cntrl-F1 at the main BIOS screen and a tonne of new options will appear in each screen.

It sure would be ice to be able to drop a 45nm native quad core Yorkfield alongside 4gb of new ram into it.

The main problem is VRM but not FSB.

 

[DavidC1]

That does not mean the p965, 975x, or 680i will support it. ( The 680i is more likely because of its 1333 fsb).

Tell that to the users of 945 chipset boards. My 945G chipset mobo claims 1066FSB support but it has no support for any CPUs that runs on the 1066FSB. 680i will probably support it unofficially, but Nvidia will obviously market it as official support.

To: Psinet

-No, nobody knows whether its a "native" dual core or its like Kentsfield
-As for 45nm CPU support, maybe if you don't care about overclocking your chipset to support the 1333MHz FSB CPU, and no if you care about official support. I say maybe because it may not support it just like Core 2 Duo didn't run on the 975 boards without a new revision
-Yes they are LGA775, that obviously doesn't mean compatibility though, as the most mainstream/high end ones are on the 1333MHz FSB
-Even if you can't support the 1333FSB CPU maybe you can run the lower end CPUs that runs on the 1066MHz FSB, there are plans of faster Core 2 Duos on the 1066MHz FSB, and I think its even true for the 45nm cores

 

[IcY18]

Yorkfield will be a native quad core, intel would not release kentsfield then release another quad core that was not native...just like they did with a core 2 duo...

 

[DavidC1]

Yorkfield will be a native quad core, intel would not release kentsfield then release another quad core that was not native...just like they did with a core 2 duo...

You are connected to Intel by any means and releasing information that is supposed to be not released but risking it anyway?? Or are you just speculating??

Some rumors point out Kentsfield-like implementation while other point out Conroe-like implementation. Why do you think there is such a disparity?? Cause no-one knows for sure.

 

[IcY18]

Well unless there are idiots running inside intel i think anyone with half a brain would realize the obvious benefits of a native quad core...yes we will see kentsfields at 45nm, we will see conroes at 45nm, and then we will see a 45nm native quad core, what its codename is is irrelevant.

intel would be highly scrutinized if it were to do the glue two cpus together method again. not only would a native quad core end up being more efficient but just like with core 2 duo it will allow intel to release a 2 cpus glued together for octo core...

 

[pottymonster]

if the 680i supports yorkfield, and yorkfield does turn out to be a native quad-core, then i will be a very happy individual. solid upgrade potential.

if it doesnt, well theres always the current core 2 quad.

 

[DavidC1]

intel would be highly scrutinized if it were to do the glue two cpus together method again. not only would a native quad core end up being more efficient but just like with core 2 duo it will allow intel to release a 2 cpus glued together for octo core...

LOL, that's so wrong. The only reason "native" multi-core parts are better is because they spend additional time optimizing the CPU, not that "native" really brings anything by itself otherwise, well maybe small performance benefits with core-to-core talk. The only thing that X2 brought in terms of "nativeness" is connect two cores by the HTT bus.

I'd like to think Yorkfield is a native part, but with conflicting rumors and guys like Charlie at the Inquirer reporting that the first "native" quad core CPU is Bloomfield, I doubt it.

 

[HotFoot]

Everything I've read suggests that Intel won't be releasing a native quad core chip until Nehalem in mid-2008. Anything we see this year is a die-shrink of what Intel already has. Changes included added SSE4 instructions and possibly larger caches.

 

[IcY18]

ok well like i said i didnt know its true code name but like i said that is irrelevant. i already posted that you would see die shrinks of both core 2 quad and core 2 duo, yes we will not have one yet but when they shrink those to lines of cpus they will release a native quad core at 45nm to do so otherwise would be shooting themselves in the foot.

 

[HotFoot]

The point is that it's not happening for another 1.5 years. The other point is that the other changes in Nehalem will be more significant than the idea of making a monolithic die.

On another note related to this thread. It seems that while AMD has been bashed for changing sockets too frequently, Intel has been praised for keeping the LGA775 around for so long. However, as far as the user experience goes, I see little extra comfort when buying an LGA775 motherboard, since chipset/VRM issues mean compatibility/upgradability is still a big issue. My 1.5 year old mobo was still obsolete 1 year after I bought it, since I couldn't put in a C2D chip.

 

[enewmen]

if the 680i supports yorkfield, and yorkfield does turn out to be a native quad-core, then i will be a very happy individual. solid upgrade potential.

if it doesnt, well theres always the current core 2 quad.

There will be a version of the yorkfield using the 1066 fbs anyway. I think you are safe. Just prey you have the correct VRM (still unknown)

 

[IcY18]

i didn't get your point, but all i was saying is that intel would not develop a new architecture at 45nm only to be the glued together version of a quad core. we all know that intel is pushing more cores for more power and especially when amd's quad core comes out they'll have 8 cores on the 4x4, so you know intel will then combine to of their native quad cores together to release their first octo core.

 

[IcY18]

my only point is the enthusiast choice of octo core will be like quad core now, people will still be on dual core for the most part(single core now) gamers and mid level enthusiasts/gamers will have quad cores(dual cores now) and the ultimate enthusiast will have an octo core (quad cores now). so you know intel will need that ultimate enthusiast option

 

[DavidC1]

i didn't get your point, but all i was saying is that intel would not develop a new architecture at 45nm only to be the glued together version of a quad core. we all know that intel is pushing more cores for more power and especially when amd's quad core comes out they'll have 8 cores on the 4x4, so you know intel will then combine to of their native quad cores together to release their first octo core.

Except nearly all rumors point out to being Kentsfield like implementation.
http://www.vr-zone.com/?i=4140
http://www.theinquirer.net/default.aspx?article=31649
http://www.theinquirer.net/default.aspx?article=36010

"The Yorktown rumours that were flying around at IDF were and are wrong, there is no converged quad core parts until Nehalem."

Shared L2 cache rumor:
http://www.hkepc.com/bbs/itnews.php?tid=675864&starttime=0&endtime=0

As for the "native xxxx core" bullshit. People remember the cache-to-cache transfer rate that Xbitlabs tested?? Apparently according to Xbitlabs, Athlon X2 is not "native" as they claim.

 

[IcY18]

ok whether its called nahalem, or yorkfield or jupiter i dont care, all i said was their new architecture would be a native quad core, so this would enable intel to come out with a octo core shortly after...i'm not talking about specific code names...

 

[DavidC1]

ok whether its called nahalem, or yorkfield or jupiter i dont care, all i said was their new architecture would be a native quad core, so this would enable intel to come out with a octo core shortly after...i'm not talking about specific code names...

LMAO. Isn't it OBVIOUS to you why the code name matters?? BECAUSE WE ARE TALKING ABOUT A SPECIFIC CPU WITH A SPECIFIC CODE NAME. And the discussion is that whether Yorkfield is a so-called NATIVE quad core or not. Obviously Nehalem is different and is one year into the future.

And you mentioned yourself "new architecture". Penryn is NOT a new architecture. No/Nay/Nada. Maybe some enhancements but basic core is derived from Merom. Nehalem is a new architecture.

 

[IcY18]

Read here

http://forumz.tomshardware.com/hardware/Penryn-Confirmed-shocker-Metal-Gate-ftopict220653.html