Ivy bridge overheating ???? Poor cooling or what ?

Hi

I was looking at recent ivy bridge notebooks reviews on notebookcheck.de

Notebook Test, Laptop Test und News - Notebookcheck.com

What surprised me is that most of them overheat like hell, i mean 95°C-100°C for the cpu even the Core i7 3612Qm which has 35 tdp.

It was already the case with Sandy bridge, throtthling stories etc...but it seems even worse with Ivy.

Can't imagine the temps with dust, these rigs will fry soon.

So weak manufacturer cooling or hot cpu ?

 

They are just a hot CPU, same power consumption as the last gen but on a smaller core with lower surface area.

 

It's not so much the architecture as it is poor thermal design on the OEM's part. They put almost NO thought into thermal dissipation.

 

Also note they are running prime95 and furmark in those tests, that is as hot as it is physically possible for these chips to run.

 

thorttling has been around for ever...the fact that a cpu gets that hot is because of poor design. my CPU is 45 TDP and it never got that high with the stock cooling system

 

On the contrary, throttling hasn't been around all that long, several years maybe. Before that, you had CPUs literally bursting into flames and, more recently, simply motherboards shutting down.

 

Intel screwed up on the cooling design for IVB. Intel used lower cost thermal grease instead of solder underheat the integrated heatsink spreader. IVB generates just about as much heat as SB but in a smaller area, and then Intel used crappy materials that are not effective at sucking the heat away.

Ivy Bridge proven to suffer from poor thermal grease by VR-Zone.com

 

But how does that relate to notebook CPU's that have no IHS?

 

impressive find +rep....can someone make a DIY of how you remove the laptop version? It looks like my 720qm has a IHS

 

Well, I guess that depends on the laptop manufacturer .

 

Well, most ppl do not need the power of i7 so a little throttling shouldn't be noticeable ? And if people do need the power, people will find a way to keep it not throttling.

win win for OEM.

Edit: cute article there, I believe people will start ripping out their desktop CPUs. If you go to the link of the japanese site, all they used is a 99c cutter ~~

 

the problem with throttling is it can kill game performance if it is a cpu intensive game like CS Source...its single thread so any throttling will murder fps

it is not like it locks a core to a single thread speed....it locks it to the lowest multiplier which is murder

 

time for tech support or new machine . OEM says, and that's why you buy warranty, so they clean out your fan regularly .

 

Why would you want to? Pre-IVB Intel CPUs used solder, not paste (if they had a IHS at all), so that's not a problem for them.

 

Which further backs up what I said... OEM's don't give a damn about thermal dissipation. They just stick some parts in a shell and hope it doesn't melt. If it throttles, it throttles. They'll tell you to just deal with it.

In the last 2 years, for instance, I can count on ONE hand the number of systems that actually have an efficient cooling design.

 

i have a m17x r4 on order....thats why i asked....not referring to my g51j

 

You mentioned the 720qm... And mobile IVB doesn't even use a IHS.

 

What I have heard is that Ivy runs about as cool as Sandy Bridge with normal use. These hot temperatures only comes when overclocking.

I could remember wrong though

 

you dont read posts very well do you? HT said a PLAIN statement of "what about notebook cpus that dont have an IHS" I made a reply saying my 720qm has one from what i can tell. In neither of those statements was IVB stated. Also I have not seen an IVB cpu yet. This is news to me that they dont have a IHS....never seen a cpu without one before....at least from what i remember. Can you repaste multiple times with no IHS? doesn't it ruin a chip to paste directly on it?

 

Keep out the personal attacks, those aren't necessary here at NBR.

Really? I haven't seen an integrated heat spreader (IHS) on a mobile processor in recent years. Just so we're on the same page, this is what an IHS looks like on an Intel and AMD processor (desktop). This is what an IHS-less notebook processor looks like (your i7-720QM).

Yes, you can repaste multiple times on an IHS-less CPU with no issue. Just carefully clean off the residue between repastes.

Greg's link is relevant mostly to desktop CPUs, not to mobile processors.

 

This is how it looks underneath the IHS

 

Good point. That being said the link does illustrate quite well that higher quality TIM is going to be needed on IVB. Period. Laptop manufacturers are going to have to upgrade their TIM relative to what they used on SB if they don't want even hotter CPUs.

 

which is not likely to happen

 

As if that'll be happening anytime soon

 

That is why I always repaste my CPUs after I recieve my notebooks. I`ve seen some pretty horrible paste jobs, and I`m pretty shure the manufacturers isn`t actually using the best paste

 

Might be necessary in some notebooks, but my T500 is still going strong after 3.5 years on Lenovo's stock thermal compound. Overall temperatures are in line with what they were when I bought it: about 33-37 C on idle at 19-21 C ambient, load still stays below 70 C.

 

weird so the actual cpu is metal? I thought that shinny metal surface was the IHS....the chip seems every sturdy.

 

That would be the silicon die.

 

huh you learn something new everyday

side note i thought it would be black like other chips.....i bet your laughing with that newb comment

 

Black?

The one in the picture above is slightly brushed, it seems. Most have a mirrored finish that looks black if it's not sitting in the light at the right angle.

 

It's no different than the old pentium 3's and athlons that had the chip exposed. I wonder if one could just remove the IHS, and either use some kind of compatible cooler, or just repaste it?

 

You can remove the IHS on a CPU that has only TIM between the die and the IHS fairly easily, though that will definitely void any warranty left.

However, based on previous information, the E6600, which had flux solder under the IHS, could not have its IHS removed without damaging the die.

 

you know like other chips on a computer...they are solid black....lol like on graphics cards and motherboards and ram ^^ lol like i said i bet your laughing

 

Chips that don't have their die exposed have a protective layer of ceramic-compound around them. When they don't have their die exposed, they don't need external cooling. Like this:

Spoiler

The actual dies of the GPU core and CPU core are exposed since they have to have some serious cooling solutions. Their heat comes right off the silicon die through the TIM and out the heatsink.

 

That's only part of it.

The 3D transistors also generate more heat than the previous generation components and they didn't realize how much more heat they produced until they had a fully fledged CPU. However, they aren't less durable than the previous designs. If anything, these chips run at 90*C on the desktop counterparts and that's still within the safezone. Laptop components are the same as well.

Not to mention they did tests with a direct contact plate on the CPU itself, taking off the cover plate and putting the heatsink onto the chip, and it produced no different temperatures. We're well aware of this over at Overclock.net

 

Ticks (die shrinks) usually come with heat and power consumption issues at first. Arrandale had the same problems when compared to Penryn.

I expect Intel will fix everything when they release a new stepping and/or refresh the lineup in 3-6 months.