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

[SickBeast]

On the new system that I just build using the Intel® Core™ i5-3570K, Intel Desktop board DZ77GA-70K and the Corsair H100 I was hitting temps in the high 80s to low 90s at 4.6GHz. Which I do believe to be a little higher than what I would like to stay at. So my experience is that it is running hot so far.

I applaud both your own and your company's admission that heat is a problem with Ivy Bridge.

You are of benefit to our community here. Thank you for being nothing like the undercover shills employed by some of your competitors.

 

[Meghan54]

Thank you for being nothing like the undercover shills employed by some of your competitors.

Geez Louise.....can't you just let that crap stop, Beast? Or are you really an undercover shill wannabe that's afraid to come out of the closet?

 

[SickBeast]

Geez Louise.....can't you just let that crap stop, Beast? Or are you really an undercover shill wannabe that's afraid to come out of the closet?

?

What have I been saying about shills? I certainly don't like them but I have been far from the forefront in terms of persecuting them.

If you support undercover marketing on this forum then I think you need to give your head a shake.

 

[vailr]

Intel only makes 22nm with trigates. So any planar design would have to be 32nm or higher.

Couldn't Intel make a decision to produce "planar design" CPU's using 22, 28, or 32 nm lithography? The latest AMD & nVidia GPU's are 28 nm and aren't designed having trigates, AFAIK.
I think that there would still be a decent market for lower powered socket 1366 CPU's, especially if only a bios update would allow using them on older x58 chipset boards.

 

[Don Karnage]

Couldn't Intel make a decision to produce "planar design" CPU's using 22, 28, or 32 nm lithography? The latest AMD & nVidia GPU's are 28 nm and aren't designed having trigates, AFAIK.
I think that there would still be a decent market for lower powered socket 1366 CPU's, especially if only a bios update would allow using them on older x58 chipset boards.

Planar transistors were on its last legs. They're like a worn out shoe that was falling apart. In innovation you have to implement new ideas and technology's.

 

[BFG10K]

I’m not buying the thermal shrink explanation. A Radeon 5770 has a similar die size (166mm2) while having a higher TDP, yet it certainly didn’t suffer from temperature problems.

Also IB is hotter at stock speeds too; you don’t have overclock it.

Also Intel confirmed there’s a change in package thermal technology, so the OP should update the thread as it implies 22nm is the sole cause.

 

[vailr]

Planar transistors were on its last legs.

There was a time when AMD CPU's could use the same CPU socket and motherboard as originally designed for Intel's Pentium CPU.
Well, maybe AMD could make some updated 28 nm lowered-voltage socket 1366 compatible CPU's, if Intel is too focused on their crispy trigate CPU "innovation" to bother with. Combine with updated PCIe 3.0 & USB 3.0 support in a new revised "x58-A" chipset motherboard. A bios update could allow the lower powered (and cooler running) AMD CPU to work in existing x58 chipset boards, but you just wouldn't get the PCIe 3.0 or USB 3.0 when used in those boards.

Anyhow, planar transistor designs will still be around for a long time yet. Especially if Intel doesn't plan on licensing their trigate transistor patents to competitors, such as AMD, nVidia, Qualcomm and others.

 

[Don Karnage]

There was a time when AMD CPU's could use the same CPU socket and motherboard as originally designed for Intel's Pentium CPU.
Well, maybe AMD could make some updated 28 nm lowered-voltage socket 1366 compatible CPU's, if Intel is too focused on their crispy trigate CPU "innovation" to bother with. Combine with updated PCIe 3.0 & USB 3.0 support in a new revised "x58-A" chipset motherboard. A bios update could allow the lower powered AMD CPU to work in existing x58 chipset boards, but you just wouldn't get the PCIe 3.0 or USB 3.0 when used in those boards.

Anyhow, planar transistor designs will still be around for a long time yet. Especially if Intel doesn't plan on licensing their trigate transistor patents to competitors, such as AMD, nVidia, Qualcomm and others.

If amd could develop a product with close to SB's IPC that clocked to 5Ghz+ i'd def be interested.

 

[superccs]

Ivy seems to be quite the turd... Its the only peice of silicon that runs hotter than its larger predecessor while using less power.

I calll BS. they messed up and accidentily used paper towel rolls instead of heatsinks..

 

[Edrick]

Like many have already stated, IB is just SB shrunk to 22nm. The uArch was not designed to use 22nm tri-gate process. It was designed on 32nm. Just wait for Haswell before jumping to conclusion about the 22nm process.

 

[Don Karnage]

Like many have already stated, IB is just SB shrunk to 22nm. The uArch was not designed to use 22nm tri-gate process. It was designed on 32nm. Just wait for Haswell before jumping to conclusion about the 22nm process.

Going to be a long 14 months

 

[Diogenes2]

On the new system that I just build using the Intel® Core™ i5-3570K, Intel Desktop board DZ77GA-70K and the Corsair H100 I was hitting temps in the high 80s to low 90s at 4.6GHz. Which I do believe to be a little higher than what I would like to stay at. So my experience is that it is running hot so far.

I'm using a Coolermaster 212+, and my temps with a 3770K @ 4.6g - 1.3v don't look that bad..

The package has been as high as 86 but the max core temp has been 76 ..

Ambient ~ 23

I'm going to see if I can do 4.8 later..

 

[cbn]

The issue is the chip is alot smaller. So you need to remove alot more heat per mm2. Even tho you produce less overall heat.

[

Thanks for the graph.

Yes, this issue of Heat density is very interesting indeed!

I can't wait to see the strategies pan out in this area of technology.

 

[oceanside]

I still would like to know how much the TIM affects the temps over a soldered lid. Intel has already come out and said it was a contributor. Likely it's negligible... but I'd certainly like to know. AFAIK, we have nothing to go by but one or two examples by some adventuresome individuals. Hardly a good sampling.

Will a soldered IHS yield 5 degrees improvement on good air? Will Haswell or Haswell-E (next planned upgrade) also have TIM under the hood?

I don't mind spending a premium on a good air cooler if it nets me 3-5 degrees on a fully loaded CPU. I happily paid the premium for a Noctua DH-D14 on my 2500K. I paid an even greater premium for a good air-cooled case. Therefore, I would gladly pay a further premium for a soldered lid on a CPU, even if it netted me just 5 degrees fully loaded.

 

[God Mode]

I would actually prefer a soldered IHS due to the number of times Ive seen dried paste in hot running and neglected systems. The surface of the IHS is larger than the exposed core and it makes sense to me.

 

[Rvenger]

I would actually prefer a soldered IHS due to the number of times Ive seen dried paste in hot running and neglected systems. The surface of the IHS is larger than the exposed core and it makes sense to me.

I am concerned about that as well... Once that TIM dries up inside the heatspreader, I wonder if the temps are going to skyrocket even more. Probably would take a couple years to notice a difference in temps though. I guess intel figures it will last 3 years and the enthusiasts can delid it and add their own if they are still using it.

 

[Denithor]

I am concerned about that as well... Once that TIM dries up inside the heatspreader, I wonder if the temps are going to skyrocket even more. Probably would take a couple years to notice a difference in temps though. I guess intel figures it will last 3 years and the enthusiasts can delid it and add their own if they are still using it.

Ouch. Hmmm. Is this the first generation of 'disposable' processors, designed to quit working after say 4-5 years so you have to buy another?

 

[meloz]

The GPU acts as dark silicon as well, actually aiding in the cooling of the chip.

Only for those who have a discreet GPU card, though. Those who use the iGPU for casual gaming at low-res (which is all it is capable off, frankly) should be the worst off. The combined heating from CPU + iGPU might really take the tempertaures to the limit.

One thing none of the mainstream reviews of IVB have touched is overclockability with iGPU enabled and stressed with some load. Agreed, this is a bit of an edge case since people who overlock are also likely to have a GPU card, but would have been interesting to see if in this particular scenario IVB suffered.


Aside, not concerned about IVB-E. In another 6-9 months Intel might refine the 22nm process and they can also use solder to conduct heat away from the chip (instead of thermal paste as they have done with the small socket product). The large socket forms the basis for high-margin Xeons, Intel won't cheap out and risk anything in that segment.

 

[meloz]

Also Intel confirmed there’s a change in package thermal technology, so the OP should update the thread as it implies 22nm is the sole cause.

Agree, I just have a feeling (no facts) the bigger reason for high temps is down to change in thermal tech. IVB-E should be illuminating, particularly if Intel decide to use solder instead of thermal paste.

 

[ShintaiDK]

Ouch. Hmmm. Is this the first generation of 'disposable' processors, designed to quit working after say 4-5 years so you have to buy another?

No, the Ivy Bridges at stock will easily run for 25 years.

So lets pack the idiotic random rumours and stay on reality.

 

[Rvenger]

No, the Ivy Bridges at stock will easily run for 25 years.

So lets pack the idiotic random rumours and stay on reality.

Do you really know how long the TIM will last under hot conditions? How do you know that it will last 25 years? No way of telling IMO. They very well may be operable at stock speeds but I can guarantee that they won't overclock once that TIM dries up, cracks, and breaks apart.

 

[Diogenes2]

25 Years !

Are you kidding ?

Sure a lot of histrionics going on here about a problem that has not been shown to be a problem ..

 

[BallaTheFeared]

I thought 25 years was used to show how off the initial comment was, three year disposable "bic" cpus.


Do you have any idea how long sealed paste will last?!1 Doo you?!/1 Fearr!!!

 

[cbn]

What is the difference in solder vs TIM?

And as far as heat density issue goes Intel has been seeing this problem coming for many years....

ftp://download.intel.com/technology/silicon/Bourianoff-Proc-IEEE-Limits.pdf

From the article above (written back in 2003):

V. SUMMARY AND IMPLICATIONS
The 30-year-long trend in microelectronics has been to increase
both speed and density by scaling of device components
(e.g., CMOS switch). However, this trend will end as
we approach the energy barrier due to limits of heat removal
capacity. For nanoelectronics, this result implies that an increase
in device density will require a sacrifice, due to power
consideration, in operational speed, and vice versa. Thus, it
appears that we are entering a regime where tradeoffs are required
between speed and density, quite in contrast to the
traditional simultaneous benefits in speed and density from
conventional scaling.

 

[BallaTheFeared]

The solution is higher IPC, bulldozer was a lie!