[Unconfirmed] Sandy Bridge-EP [8C16T] benchmark results leaked

[sniper_sung]

Source (you'll need to have replied to the OP to be able to see the pictures there): http://bbs.siluhd.com/thread-474123-1-1.html

Disclaimer: the OP was posted a week ago. If it's already been mentioned here and identified as photoshop'ed crap then please ask a mod to delete this post.





















If it's real, it makes the i7 3960X [6C12T] completely pointless, and also smashes Bulldozer.

[hersounds]

well,so sorry but with ivybridge on the horizon that may worry bulldozer?

[Origin_Unknown]

well,so sorry but with ivybridge on the horizon that may worry bulldozer?

as soon as i saw this thread i wondered how long a comment like that would take

[Lanek]

Lol Ivy around the corner ?` not before june 2012 ... well you can maybe see faaaar on the horizon .. lol ..


If i was AMD i will not be worry for Ivy vs Bulldozer, but the question is open with Piledriver vs Ivy.


Anyway back on topic: i cant read CPU_Z but if this Xeon is avalaible on the LGA2011 in addition of SandyBridge-E... it will cost an arm ( basically 1200-1400$), and i don't see why AMD will have to worry about this..This concern so little consumers and sell. Then on professionnal side, this is a Xeon, so i more interested on how this cpu can impact the 12-16cores Orochi ...

[Pestilence]

Lol Ivy around the corner ?` not before june 2012 ... well you can maybe see faaaar on the horizon .. lol ..


If i was AMD i will not be worry for Ivy with Bulldozer, but the question is open with Piledriver vs Ivy.

Ivy should be out in mid march and not june.

[Lanek]

Ivy should be out in mid march and not june.

my bad if it is Q2, i was remember H2 ....

Anyway a 8cores 16threads start to be really interessant for my personal worksation at home ( Autocad 3D+Inventor).

[Pestilence]

my bad if it is Q2, i was remember H2 ....

Anyway a 8cores 16threads start to be really interessant for my personal worksation at home ( Autocad 3D+Inventor).

All good bro. I fully expect an 8 Core with Ivy Bridge E in Q1 of 2013 which would be awesome.

[informal]

That's 2P system right?

[mAJORD]

1P

Overclocked (?) should be 2.7ghz

[PatRaceTin]

score look great

[god_43]

what case is that? it looks really nice.

[Lanek]

Corsair one no ?

[iboomalot]

my i7-920 has to run nearly 3.8ghz to equal that chip in SuperPi at 3.4ghz if it can clock to 5ghz WOW ***drool***

[Tenknics]

If it's real, it makes the i7 3960X [6C12T] completely pointless

Why?

[Pestilence]

Why?

Because we'd rather pay 1K for an 8 core Xeon then 1K for a 6 core I7

[trn]

Except the 6c will clock higher on each thread and have unlocked multipliers, no results yet on SB-e non-multiplier OC'ing.

[JkS]

That is one sexy chip. /droooool

[Pestilence]

Except the 6c will clock higher on each thread and have unlocked multipliers, no results yet on SB-e non-multiplier OC'ing.

If what has been leaked so far is true i'll gladly take a 8 core xeon as long as i could use a 150mhz bus.

35 X 150 = 5250 would be plenty

[Lanek]

150mhz on a sandybridge ? i know they use something special on the models X, a part who permit to separate the gpu, memory controller, fsb bus... but yet we don't know how it work ( Anandtech have fail to use it ) ... and what max fsb is available with it. i mean SB start have some realy problem at 105-110 fsb.. we don't know what can be the max at this point...

[Pestilence]

150mhz on a sandybridge ? i know they use something special on the models X, a part who permit to separate the gpu, memory controller, fsb bus... but yet we don't know how it work ( Anandtech have fail to use it ) ... and what max fsb is available with it. i mean SB start have some realy problem at 105-110 fsb.. we don't know what can be the max at this point...

From what I read 2011 will have selectable ratios... 125/250/175/200/225. Guess well have to wait and see

[Hornet331]

150mhz on a sandybridge ? i know they use something special on the models X, a part who permit to separate the gpu, memory controller, fsb bus... but yet we don't know how it work ( Anandtech have fail to use it ) ... and what max fsb is available with it. i mean SB start have some realy problem at 105-110 fsb.. we don't know what can be the max at this point...

The use multipliers and buffers to circumvent the limitations of normal desktop SB. I still think the only reason why normal SB can go very high with Bclck is because the PCI-E controller and the DMI bus only has a 1.1 multiplyer.

[radaja]

dont know about SB-EP but SB-E should be fun?
with the i7-3820 4c/8t locked part the reported specs say a 3.6GHz base and 3.9GHz turbo
and with the x79 reference clock of 95MHz that means the i7-3820 uses the x38 multiplier,so...
using the 1.67x ref ratio we can do 95x1.67=158.65x38@6028MHz and if we can use the turbo
multi's like with SB then 95x1.67=158.65x41@6504MHz which should be fun for a locked chip

[hurleybird]

If it's real, it makes the i7 3960X [6C12T] completely pointless, and also smashes Bulldozer.

Sandy-EP's competition is Interlagos.

[JumpingJack]

Sandy-EP's competition is Interlagos.

On the desktop, people will compare the performance of SB-E and BD (8 core)... on performance, SB-E shows BD the door, but on price BD will reign supreme.

[K404]

If the X3960 is $999 at launch, this will cost way more. Think QX9770 Vs QX9650

Volts/MHz. TDP look much better than the '3960.... wonder whats going on there??