Ivy Bridge heat problems remain even after IHS removal

/ 4 years ago

We recently brought you the news that I.M.O.G from had come to the conclusion Ivy Bridge was a hotter chip due to the fact there was an additional layer of inefficient thermal paste between the processors Integrated Heat Sink (IHS) and the processor itself which affected cooling efficiency. Naturally the only way to test this hypothesis is with some empirical testing and that’s exactly what a PC EVA forum member has done. Using a Core i7 3770K processor, with a Noctua NH-D14 CPU cooler and Prolimatech PK-1 thermal grease there was testing conducted with the IHS on and the IHS off to see if it gave any differences to thermal performance. The testing methodology involved using AIDA64 Extreme Edition for idle and load average temperature monitoring with Prime95 being used to generate the load temperatures. Testing was done at 4.5GHz with 1.2 volts on the core.

The results are summarised below in tabulated format:

As you can see even with the “cheap” thermal paste and additional outer IHS layer removed the cooling performance is still identical within a traditional margin of error of 5%, therefore its safe to say the high temperatures have nothing to do with IHS and are simply a consequence of the way Ivy Bridge’s revised manufacturing process works. In practice this now means Ivy Bridge will typically reach lower stable 24/7 clock speeds than Sandy Bridge but will offer similar performance (for example a 3570K at 4.8GHz is the same as a 2500K at 5GHz) and with slightly lower power consumption.

Source: My Drivers

  • this is bad news……. rl bad

  • solomonshv

    good thing i got a 2700k and didn't wait for this. looks like intel's success got to their head, again. at the prices they charge for these chips, you'd think they wouldn't go cheap on the construction

    intel needs to realize that the world won't revolve around them for much longer, especially since GPUs can now do all the heavy lifting in business application, and do it better (Tianhe-1A), while ARM based silicon will soon be in every mainstream device (ipad and future windows 8 tabs).

    • well i will switch to those arm once they can run games 30+fps in 3d…

      • solomonshv

        Your opinion doesn't matter. More people play games on mobile devices than on tricked out computer. My computer runs an intel i7 2600k with a pair of GTX 680s. But i am a tiny minority. Intel will not survive on people who demand 30FPS.


        • VirtualMark

          People use desktops for more than gaming. I7's are the most powerful chips around, and some of us need them. To say 'your opinion doesn't matter' to someone is beyond ignorant.

  • GeNuINE

    Then Intel 3rd Gen Epic Fail. Very Sad News 🙁

  • Pingback: Ivy Bridge Heat Problem Persist Even After IHS Removal | Maximum-Tech()

  • So you stuck a heatsink directly on the CPU die? Well that has to be the stupidest thing I've ever seen. Of course temperatures won't be any better. CPUs do not use an Integrated Heat Sink, they use a Heat SPREADER. Re-do the test with proper thermal past between the CPU and the heatspreader, and a heatsink (like anyone with half a brain would have told you) and then we'll talk.

    • On Ivy Bridge the quality of thermal paste between the IHS and CPU die clearly makes a difference. We are NOT talking about an external heat sink/ CPU cooler this is purely the CPU component ONLY, these temperatures issues will be reflective of whatever CPU cooler you choose to use. So next time you want to go trolling, use your brain, get your facts right and show people that you have at least half a brain before you start criticising others.

      • HowDareYouSir

        What on earth are you on about? Richard is spot on… The heat spreader is an essential part in cooling the CPU. The whole point was to see if replacing the TIM with a better after market solution would drop the temps, but they also removed the spreader from the equation -> making the test irrelevant. Every test that I've come across where the the TIM has been replaced (and the IHS re-attached – why the heck would you not?) has resulted in a significant drop in temps. So yeah – like you said the quality of the thermal paste between he IHS and the CPU die clearly makes a difference, only that is not the conclusion the OP arrived at (again, because he didn't put the IHS back). So next time maybe follow your own advice.

        • Mcik

          The IHS is mainly there for protection, think about it if you put a huge heatsink directly onto the core the heatsink itself acts as the heatspreader that's what it's designed to do, absorb heat and dissipate it.

          Having the IHS sandwiched inbetween with twice the amount of TIM layers simply hinders the transfer of heat from the core to the heatsink.

          Other sources have shown that removing the IHS does improve cooling so I'm not sure why these remain identical.

          • anon

            Richard and Ryan; logic fail, maybe?

            Chip surface -> TIM -> IHS -> TIM -> Heatsink


            Chip surface -> TIM -> Heatsink.

            You remove the extra layer of TIM and the surface of the heatsink, property wise, will do exactly the same thing as the IHS. The IHS isn't some mystical super-conductive contraption. The base of the heatsink is just the same.

            Chips prior to Ivy Bridge being; Chip Surface -> Solder -> IHS -> TIM -> Heatsink

            There are no benefits to be gained in removing a soldered IHS, whereas an IHS using TIM is a whole different matter.

          • wtfruon?

            IHS removal and mounting a HS or WB directly on the die is far superior to any IHS configuration. IT IS NOT for the faint of heart or the retarded, Iv been de-capping proc’s since they started putting those things on, and they started putting them on when computer building started becoming mainstream, so that all these under educated idiots wouldn’t break dies, so as to not discourage people from building more computers and spending more money. anyone who honestly believe that an IHS makes for better temperatures needs to go do some reading. IHS is just a cute name for die protection, its not there to make temps better. its there for the sole purpose of spreading the load over the entire silicon, so when botards with too much money and not enough brains try to over tighten the piss out of an aftermarket cooler without ensuring a even mount is achieved, they don’t break the die. i have de-capped a good 10+ procs and never once have i seen temperatures NOT drop significantly.

    • Shinapo Smith

      FVCK ME IN THE BLEEDING ANUS, YOU’RE A FVCKING DUMB CUNT! Get a clue and we’ll talk!

  • Andrius St

    Replacing cheap TIM makes 5-7C difference. Slimming the sandwich down should have further positive impact. Results don't make sense. I'd take this with a pinch of salt.

  • B. Zell

    In taking off the IHS for direct CPU die cooling will result in a mandatory modification to the socket latching system and heatsink mounts as to ensure proper fit and pressure for accurate cooling!

  • justletmepostdammit

    You sure you applied internal heat grease correctly?

    oh wow wtf am I doing replying to year old article

    • Ha, I didn’t do the testing someone else did. But yes the application was correct i think, removing the IHS doesn’t improve temps much because it then makes the mounting kit less efficient (removing an 8mm layer of metal [the IHS] means the mount is less tight and efficient)