Sandy Bridge model numbers revealed

[IntelUser2000]

http://www.xfastest.com/cms/tid-46719/

The new model numbers simply add "2" to so the new processors go by 2x00. The model numbers vary depending on clock speed, core/thread count, and L3 cache size.

Core i3 2100 3.1GHz 2C/4T 3MB
Core i3 2120 3.3GHz 2C/4T 3MB
Core i5 2400 3.1GHz 4C/4T 6MB
Core i5 2500 3.3GHz 4C/4T 6MB
Core i7 2600 3.4GHz 4C/8T 8MB

 

[khon]

Hmm, not loving that if true.

First of all I think they should get a new name since its a brand new architecture.

Secondly there is only one 4C/8T model which is what I want, but being the top of the line it might be a bit costly (I was hoping for ~$300).

 

[Idontcare]

Hmm, not loving that if true.

First of all I think they should get a new name since its a brand new architecture.

Agreed. Is this a tick or a tock? (facetious question)

If Intel does little more than recycle the existing Nehalem architecture's brand name then that hardly yells "new and exciting stuff inside" when its sitting on the shelves of Best Buy and Frys.

Marketing fail at a minimum.

 

[Nemesis 1]

I like the naming scheme . I believe the idea here is for people to identfi.

low range i3----
low middle i5----
low high i7----

Server high i9 don't know 2011socket.

 

[taltamir]

Agreed. Is this a tick or a tock? (facetious question)

If Intel does little more than recycle the existing Nehalem architecture's brand name then that hardly yells "new and exciting stuff inside" when its sitting on the shelves of Best Buy and Frys.

Marketing fail at a minimum.

totally agreed there... intel has a very stupid naming scheme.
buuut... I can see them trying a readeon like scheme, where you have the 2xxx followed by 3xxx, 4xxx, 5xxx... problem is, they precede those with the i3, i5, and i7.
mmm... it seems they kept the core/thread numbers though from nehalem. i3 is 2C/4T just like nehalem. i5 is 4/4, and i7 is 4/8.

 

[NecessaryEvil42]

So these will all be for socket 1155? Any info on the other new socket (2011, I think) that will be released?

 

[Maximilian]

Brand new naming scheme, people complain about it being confusing.

Sticking with the same familiar naming scheme, people complain about it not being new and exciting.

No pleasing some people

 

[Idontcare]

Brand new naming scheme, people complain about it being confusing.

Sticking with the same familiar naming scheme, people complain about it not being new and exciting.

No pleasing some people

I don't want a brand new naming scheme...I want to see a brand new brand name.

Sandy is supposed to be a tock...i.e. a brand new architecture. Not a tweaked junior version of an existing architecture (that is what tick's are for).

Core i3/i5/i7 was Nehalem (and westmere) architecture.

You don't release a mini-van and a pickup truck and call them the same product name (Sienna Tundra?) and give out paychecks to the marketing team for slapping an iterated number in the mix.

"If you and your 6-person family loved our 2010 mini-van, the Sienna Tundra i5 940, then you are just going to love our all new completely redesigned Sienna Tundra i5 2940 line of pickup trucks!"

Imagine if AMD elects to name Bulldozer products as Phenom II and they just tack on another number to model number scheme...brilliant!

 

[taltamir]

exactly what idontcare said... and it WAS confused...

AMD currently has the most straightforward and least confusing naming scheme... I only wish intel and nvidia followed suit.

Intel should have called it the Core4. with Core i# being relegated to a one time botchup of the core 3 brand.

 

[DrMrLordX]

They should rip off Lone Gunmen and call it the Octium.

 

[khon]

Makes you wonder if the marketing guys at Intel are BMW fans. Same 3, 5 and 7 series for 30+ years now.

 

[taltamir]

Makes you wonder if the marketing guys at Intel are BMW fans. Same 3, 5 and 7 series for 30+ years now.

could be... this whole drek could be a gimmick to capitalize on BMW reputation.

 

[jaydee]

I just wish naming convention followed the specs somehow. Like:

3.1GHz 2C/4T 3MB = Core i2 3130T
3.3GHz 2C/4T 3MB = Core i2 3330T
3.1GHz 4C/4T 6MB = Core i4 3160
3.3GHz 4C/4T 6MB = Core i4 3360
3.4GHz 4C/8T 8MB = Core i4 3480T


edit: obviously they wouldn't follow this exactly because it would seem the newer i2 series is "worse" than the older i3 series.

 

[Chiropteran]

I don't want a brand new naming scheme...I want to see a brand new brand name.



You don't release a mini-van and a pickup truck and call them the same product name (Sienna Tundra?) and give out paychecks to the marketing team for slapping an iterated number in the mix.

Huh? These slightly updated slightly faster processors will be used for the EXACT SAME PURPOSE as the similarly named processors they are replacing. Your analogy is terrible.

From the early 90's on we had Pentium, Pentium Pro, Pentium 2, Pentium 3, Pentium 4, even though each generation had a new architecture the name remained similar except for a numeral (or "Pro" designation). These names follow that trend perfectly. Think about how the names will be used in common speech.

Core i7 2600 3.4GHz 4C/8T 8MB

I don't think people will actually say "core i7 2,600" I think it's more likely it'll be referred to as the "core i7 2 600", which leads to a naming scheme similar to when we went from the Pentium to the Pentium 2.

Call these new CPUs "core i7 2", or whatever.

 

[khon]

Something I was wondering: The Nehalem processors all shared a clock of 133MHz, which doesn't seem to be the case for these Sandy Bridge processors, so what do you suppose the new clock is ? 100 MHz ?

 

[Idontcare]

Your analogy is terrible.

Thanks!

 

[IntelUser2000]

Brand new naming scheme, people complain about it being confusing.

Brand naming never made sense. I think the two worst examples are:

Pentium III-How did the addition of SSE instructions that needed to be recompiled to take advantage of justify a brand new naming? The earlier Pentium III's were just that

Core i7 970-Rumored to be a cheaper Gulftown. The 980X made a bit of sense at least because the addition of "X" at the end made it different. Core i7 970 should be the non-extreme variant of 975, just like the 960 was the non-extreme variant of 965. How does this work at all?

Something I was wondering: The Nehalem processors all shared a clock of 133MHz, which doesn't seem to be the case for these Sandy Bridge processors, so what do you suppose the new clock is ? 100 MHz ?

Yea, it is 100MHz. Don't know the exact reason for the change.

 

[khon]

You'd think they could at least go from core i3,i5,i7 to i13,i15,17. Would still be easy to understand due to its similarity to the older models, and it would signify a clear change.

 

[taltamir]

You'd think they could at least go from core i3,i5,i7 to i13,i15,17. Would still be easy to understand due to its similarity to the older models, and it would signify a clear change.

oooh, that would work. TBH I was busting my head on how to make that big mess work until you came up with that.

mmm, or maybe Core 2i3, 2i5, 2i7... but that is still silly and confusing.
Should be Core 4 i2/4, Core 4 i4/4, Core 4 i4/8

 

[edplayer]

the names mean nothing to me. All I care about is it performance relative to other cpus and its price.

They could call it the Buttsplatter 9000 and I would still consider it.

 

[aigomorla]

arent you guys curious why no one has yet to pull OC?

Cooler even always usually does OC.

Yet no one has brought out an OC?

Makes you guys think no?

 

[edplayer]

I'm not curious. I'm using a Q9550 and I'd like that to last till at least Ivy Bridge or maybe Haswell. Only upgrades I plan on are a SSD and graphics cards (and more hard drives like always).

I assume its not that good (at overclocking)

 

[cbn]

http://www.xfastest.com/cms/tid-46719/

The new model numbers simply add "2" to so the new processors go by 2x00. The model numbers vary depending on clock speed, core/thread count, and L3 cache size.

Core i3 2100 3.1GHz 2C/4T 3MB
Core i3 2120 3.3GHz 2C/4T 3MB
Core i5 2400 3.1GHz 4C/4T 6MB
Core i5 2500 3.3GHz 4C/4T 6MB
Core i7 2600 3.4GHz 4C/8T 8MB

I wonder if we are going to see the same IPC discrepancy with these dual cores/quad cores as we do with the current Clarkdale vs Lynnfield situation?

 

[Idontcare]

arent you guys curious why no one has yet to pull OC?

Cooler even always usually does OC.

Yet no one has brought out an OC?

Makes you guys think no?

I've been assuming that it is because either (1) coolaler (a sanctioned leaker, viral marketer really) hasn't been authorized to start making such things public at this time, or (2) the platform itself (new socket, new chipset, new bioses, etc) is not really stable enough at this point to make OC'ing a worthwhile venture.

Considering that one of the major features of SB is the on-die integrated GPU and Coolaler wouldn't/couldn't post any benches for the GPU I am going with option 1 above. His masters simply haven't signed off on making it public yet.

Otherwise even if option 2 were true he would have still attempted it and posted the results with the usual bevy of caveats declaring that the results are early and likely to improve, etc etc.

 

[IntelUser2000]

I wonder if we are going to see the same IPC discrepancy with these dual cores/quad cores as we do with the current Clarkdale vs Lynnfield situation?

Shouldn't be, aside from how much the larger L3 caches would impact it. Every Sandy Bridge will be a fully integrated chip, think of it as Lynnfield with graphics.

arent you guys curious why no one has yet to pull OC?

Well, think of the situation before Nehalem was released. There was a rumor that said Nehalem would be made really hard to overclock, or factory locked. How is that going 2 years after release?