LGA2011 for SNB-E sneak peek

[Peter Griffin]

more some info
http://hoc.hu/Szamtech/CPU/Sandy-Bridge-E-akar-36-kal-gyorsabb-mint-a-990X.html

 

[karlotta]

Got pictures or at least model number on those LGA2011 board, I want to pre-order mine ?

http://wccftech.com/asus-x79-chipset-based-p9x79-deluxe-motherboard-unveiled/

 

[Thorburn]

CPUZ says C0/C1, dunno which. Or if it's one and the same. It's Model A, Stepping 4, bought the CPU in february or march 2009 I believe. Not sure when it first came out though.

C0 = Core i7
C1 = Xeon

They didn't change the Model/Stepping code for some reason, so CPU-Z can't differentiate, C1 was validated for DDR3-1333 which i7 never officially supported.

Desktops went straight from C0 to D0.

 

[Grall]

Got pictures or at least model number on those LGA2011 board

Like I mentioned, I don't follow LGA2011 that closely since my current system is still more than good enough, but if you can stomach the Swedish, here's MSI X79A-GD65 8D, or the Gigabyte X79 G1.Assassin 2, they both look pretty cool (don't ask me WTH is up with those weird product names though, because I don't have the first clue... )

Actually, on closer look the Gigabyte board only supports 4 modules, so scratch that one off your list, unless you got a fetish for its black/neon green color scheme.

What's the deal with PCIe 3.0 ? I thought performance wise, it's pretty small improvement with those PCIe iteration. Or can't PCIe 3.0 board fit in PCIe 2.0 ?

Mainly just future proofing. I don't like the idea of buying a new, very expensive CPU that doesn't support the latest connectivity tech.

They didn't change the Model/Stepping code for some reason, so CPU-Z can't differentiate, C1 was validated for DDR3-1333 which i7 never officially supported.

Ah, I see. Well, it's an i7 for sure, so C0 it is then. It's running very well with 1660MHz DDR3 by the way, validated or not. I gotta say, Intel absolutely builds some really top-notch memory controllers.

Too bad they can't design a discrete GPU if their lives depended on it though, but maybe both ATI and NV ought to thank their lucky stars Intel can't perform in this regard; if it could get it up, they'd get screwed so bad...

 

[Blazkowicz]

as there are news about delayed sandy bridge Xeons, I've just learnt of a LGA 1356 socket. that feels mad. three sockets now!
it's basically another LGA 1366.

I assume it doesn't target the consumer, as the spot is taken by single socket 2011.
http://en.wikipedia.org/wiki/LGA_1356
http://www.cpu-world.com/news_2011/2011101001_Launch_schedule_of_Intel_Xeon_E5_microprocessors.html

 

[DarthShader]

I thought that socket rumour died ca. 1 year ago, didn't it?

 

[fellix]

i7-3960X vs. i7-3930K (15M vs. 12M L3)

 

[denev2004]

Anyway I'd like to ask a question, do somebody exactly know how much and what kinds of ALU are there in one single SNB-E Core?

 

[DarthShader]

Try here: http://www.realworldtech.com/page.cfm?ArticleID=RWT091810191937&p=6

 

[mczak]

Asking for alus in "SNB-E" doesn't really make much sense though as it's certainly exactly the same for any Sandy Bridge cpu...

 

[Grall]

The cores and ALUs of SNB-E is just plain ol' SNB AFAIK, there's just more cores, and more cache. And more I/O, of course.

 

[denev2004]

Asking for alus in "SNB-E" doesn't really make much sense though as it's certainly exactly the same for any Sandy Bridge cpu...

Well I missed the information about the latest one...

 

[homerdog]

Why is there so much cache on this beast? Wouldn't that area be better spent on more logic? It seems like there's hardly any difference at all between the 12 and 15MB variants.

 

[fellix]

Why is there so much cache on this beast? Wouldn't that area be better spent on more logic? It seems like there's hardly any difference at all between the 12 and 15MB variants.

It's for the server/HPC applications mostly. With the increase of the memory bandwidth, the data pre-fetch becomes more efficient and that requires more cache to utilize it.