Ivy Bridge

john3850

Golden Member
Oct 19, 2002
1,436
21
81
The sb was the easiest to oc cpu do to unlocked multiplier and no fsb setting.
When the Ivy Bridge comes out on the 22 nm die shrink will its small size limit the vcore and oc.
 

psolord

Platinum Member
Sep 16, 2009
2,142
1,265
136
The sb was the easiest to oc cpu do to unlocked multiplier and no fsb setting.
When the Ivy Bridge comes out on the 22 nm die shrink will its small size limit the vcore and oc.

Smaller die size, usually means more frequency with less voltage.

We may indeed be limited to the max voltage we have to use, compared to the current gen. For example you may be able to use 1.4V for a Sandy@5Ghz today, but 1.4V could be out of the question for Ivy. Still, we couldn't possibly care any less, if it could hit 6Ghz for 1.35V or something for example.
 

Borealis7

Platinum Member
Oct 19, 2006
2,901
205
106
Ah, but IVB is not just a die shrink. dont forget the whole new transistor technology (3D Tri-Gate) which will also make it need even less voltage (due to decrease in leak) and allow for even faster switching speeds.

intel is on to some serious voodoo with this next gen.
 

psolord

Platinum Member
Sep 16, 2009
2,142
1,265
136
By the opportunity presented but this "Ivy Bridge" thread, I'd like to ask the following

1) Is it absolutely 100% sure, that there will be Ivy Bridge processors that will be able to work with todays Sandy Bridge motherboards? If yes, are there any features that may be locked, if the cpu is used with an older gen mobo?

For example, is there a possibilty we could face the unbelievable scenario of Ivy Bridge working on the Z68 platforms, but not be able to overclock or something? Could we have Quick Sync 2, inoperable on Z68? You see what I mean...

2) Is it true that the tri gate transistors will only be used for the cache and not for the logic circuit? If yes, will the trigate technology still provide substantial gains, even if used solely for the cache?

thanks :)
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
1) Is it absolutely 100% sure, that there will be Ivy Bridge processors that will be able to work with todays Sandy Bridge motherboards? If yes, are there any features that may be locked, if the cpu is used with an older gen mobo?

For example, is there a possibilty we could face the unbelievable scenario of Ivy Bridge working on the Z68 platforms, but not be able to overclock or something? Could we have Quick Sync 2, inoperable on Z68? You see what I mean...


Nothing in life is 100% absolutely assured.

At one point AMD said Bulldozer would be officially supported on AM3...then later on they said it won't be officially supported.

Intel could come out today and tell you it is 100% absolutely assured that IB will work on today's chipset/mobos, and one day before IB's is launched they could come out and state "well, not really...".

That said, Intel has historically (not always, but for the most part) engineered their tock platforms (sandy was a tock, as was the Z68 chipset and platform) to be forward compatible with the intended tick refresh.

So it would uncharacteristic of Intel to make the Z68 platform available today while knowing it won't be compatible with IB.

No guarantees.

2) Is it true that the tri gate transistors will only be used for the cache and not for the logic circuit? If yes, will the trigate technology still provide substantial gains, even if used solely for the cache?

That's what you get for reading too much MostlyInaccurate crap. No, trigate xtors will be used for the entire chip, not just the cache.
 

Edrick

Golden Member
Feb 18, 2010
1,939
230
106
We may indeed be limited to the max voltage we have to use, compared to the current gen. For example you may be able to use 1.4V for a Sandy@5Ghz today, but 1.4V could be out of the question for Ivy. Still, we couldn't possibly care any less, if it could hit 6Ghz for 1.35V or something for example.

+1

There is a very good possibility that we will be hitting well over 5Ghz with less than 1.4v on IB. But to answer the OP's question, I do believe that the max Vcore on 22nm will be slightly less than 32nm. (I think the same was the case for 45nm and 32nm.)