Intel Admits That Ivy Bridge Runs Hotter Because Of 22nm Shrink

[Don Karnage]

CHIPMAKER Intel has admitted that its latest Ivy Bridge processors run hotter than its previous generation Sandy Bridge chips.
Intel launched its Ivy Bridge line last week with the firm touting its 22nm process node as its biggest achievement. Intel's smaller process node along with changes in the interface material between the die and the heat spreader have led to reports of higher temperatures.
Overlockers lifted the lid on its Ivy Bridge sample to find that the chip's die had thermal paste on it to interface with the heat spreader, whereas Sandy Bridge processors used solder. Now Intel has confirmed that it is using "a different package thermal technology" on its Ivy Bridge processors.
Perhaps unsurprisingly, Intel said the shrink to the 22nm process node leads to higher temperatures due to increased thermal density, adding that "users may observe higher operating temperatures when overclocking". The firm added, "This is as designed and meets quality and reliability expectations for parts operating under specified conditions."
Intel's process node shrink has concentrated heat generation, meaning it needs improved heat dissipation, which judging by reports and our own tests, it doesn't have. Although our tests were conducted on the statistically irrelevant sample size of one, put simply the same heatsink and fan that worked on our Sandy Bridge Core i7 2700K could not cut it on the Ivy Bridge Core i7 3770K.
Intel will claim that Ivy Bridge thermal design power is the same as Sandy Bridge and in this case it is true that temperature isn't linked to power usage. Nevertheless, going by Intel's own admissions overclockers might want to steer away from Ivy Bridge. µ
Source: The Inquirer (http://s.tt/1aq3n)

Source

http://www.theinquirer.net/inquirer/news/2171299/intel-admits-ivy-bridge-chips-run-hotter

If you think IB runs hot wait till Broadwell and 14nm

 

[BallaTheFeared]

Cascade or TEC?

 

[dma0991]

Intel did the right thing and admitted the truth instead of pretending that it never happened. :thumbsup:

 

[Diogenes2]

Was there ever any doubt ?

Do we need Intel to tell us water is wet ?

 

[BallaTheFeared]

Next we're going to hear that it actually has less transistors than initially reported.


I like how Ivy runs hot, but for some odd reason it can tolerate cold whereas Sandy can not. At least they got that part right, the cooler running chip can't run cold but the hot chip can, er…

 

[Joseph F]

Cascade or TEC?

Cascade, fo'shizzle. TECs are garbage.

Intel did the right thing and admitted the truth instead of pretending that it never happened. :thumbsup:

This. :thumbsup:

 

[Don Karnage]

Was there ever any doubt ?

Do we need Intel to tell us water is wet ?

Yes we do.

 

[dma0991]

Next we're going to hear that it actually has less transistors than initially reported.

Bulldozer got that right.

 

[mikeymikec]

Intel did the right thing and admitted the truth instead of pretending that it never happened. :thumbsup:

I agree. I wonder if 14nm will be delayed, as (IMO) Intel haven't quite nailed 22nm yet (at least as well as SB 32nm).

 

[Don Karnage]

I agree. I wonder if 14nm will be delayed, as (IMO) Intel haven't quite nailed 22nm yet (at least not to the quality of SB).

Broadwell wont be here till 2014 so why would they delay it? Also ivy bridge is amazing at stock speeds and volts. It just gets hot when overclocked.

 

[pelov]

Broadwell wont be here till 2014 so why would they delay it? Also ivy bridge is amazing at stock speeds and volts. It just gets hot when overclocked.

Ivy was delayed due to fab process issues and SB-E was delayed due to chipset problems. Then there's the Cougar Point issues last year and Ive-E being pushed back. Needless to say even chipzilla runs into problems, it's just that we don't really notice them as much because AMD is so far behind

 

[grkM3]

Im so tempted to buy an ivy just for fun but I just ran intel burn test at 4.9ghz and almost 1.5 volts and my 2600k never went over 80c full load,it will bench at 5.4 ghz at 1.55-1.6 volts too.

I know 100% that any retail ivy wont out do my current setup and its killing me not trying one out.

 

[Edrick]

Intel haven't quite nailed 22nm yet

No kidding. They just released it. Haswell should offer better 22nm designs.

 

[God Mode]

I'm more worried about cold/hot cycles and degradation than overall operating temps.

 

[mikeymikec]

- edit - I'll wait for the figures to come in for more mainstream / less expensive IVB CPUs before I comment further on the architecture. The 3770K may just be their first attempt at pushing it as far as possible.

 

[pittguy578]

Maybe multiprocessing will make a comeback either by putting more than one chip on a motherboard or by putting more than one processor on a single die? is that possible?

 

[nyker96]

I like the quote " Although our tests were conducted on the statistically irrelevant sample size of one"

 

[Don Karnage]

Ivy was delayed due to fab process issues and SB-E was delayed due to chipset problems. Then there's the Cougar Point issues last year and Ive-E being pushed back. Needless to say even chipzilla runs into problems, it's just that we don't really notice them as much because AMD is so far behind

Ivy was delayed because of no competition from Amd. There is nothing wrong with Ivy. Ivy was designed for low power not high end overclocking

 

[pelov]

Oh? Because AMD can't compete or because they ran into 22nm issues?

Even the best fabs in the world run into problems every now and again.

Initially they were meant to release late Q1 early Q2 but that was bumped up to late Q2. The i3 chips faced yet another delay and they won't be here for another month+.

I think you're overestimating Intel's capacity here. But I do agree with you; Ivy is a great chip and particularly for mobile. The only place it really disappoints is the enthusiast desktop space.

 

[AtenRa]

Ivy was delayed because of no competition from Amd. There is nothing wrong with Ivy. Ivy was designed for low power not high end overclocking

SB was designed for high performance at 32nm, IB is just a shrink(Except the iGPU) at 22nm. Intel's 22nm Tri-Gate was designed for high performance at lower Vcc.

An 8-core IB-E could have lower operational temps than 4 core IB.

 

[JAG87]

Ivy was delayed because of no competition from Amd. There is nothing wrong with Ivy. Ivy was designed for low power not high end overclocking

Exactly, just as a general eye opener, nothing intel designs is for overclocking. The article is completely misleading, intel hasn't admitted to anything because there's nothing to admit. Ivy Bridge doesn't run hot out of the box, it runs just as cool if not cooler than Sandy Bridge.

They just confirmed that Ivy Bridge uses paste instead of solder, that's not a design fault.

 

[dave_the_nerd]

Maybe multiprocessing will make a comeback either by putting more than one chip on a motherboard or by putting more than one processor on a single die? is that possible?

wat

 

[Don Karnage]

SB was designed for high performance at 32nm, IB is just a shrink(Except the iGPU) at 22nm. Intel's 22nm Tri-Gate was designed for high performance at lower Vcc.

An 8-core IB-E could have lower operational temps than 4 core IB.

Tri Gate was designed for high performance under 1.0v

Intel didn't design IB for the 1% of people who want high clocks. Intel designed IB for the 99% of OEM's. They added K sku's so we wouldn't go buy Amd processors.

 

[formulav8]

Ivy was delayed because of no competition from Amd. There is nothing wrong with Ivy. Ivy was designed for low power not high end overclocking

Thats a crock. AMD has very little to do with what Intel decides. AMD had nothing to do with the delay.

 

[pelov]

Don, I think his point was that IB-E would have a larger die thus better heat dissipation. The issue with Ivy isn't the voltage but rather that it can't dissipate the heat with that added voltage. A larger die could potentially alleviate some of those issues and we may see some higher clocks.

We saw the same with SB vs SB-E where the larger SB-E chips would idle and load at lower temperatures given the other conditions are equal. It's different this time around --as you've already mentioned the prime operating voltage with that chart -- so it remains to be seen how much or if a larger die would benefit.