Mac Pro 2008 (3,2 GHz) CPU gasket

I decided to test Collaborate Lab Liquid Ultra on a Mac Pro 4,1 with dual X5570 2.93 GHz CPU's. The machine was troubling since I've noticed CPU temperatures reaching over 95 C on one of the CPU's.

Both tests were monitored using Macs Fan Control, while running about 20 processes of "yes > /dev/null &" at the command prompt in Terminal.

Here are the results using MX-2:

Then I tried Collaborate Lab Liquid Ultra on CPU B:

As the above clearly indicates, there's no difference. And, btw, I did it twice and applied CLL Ultra on both sides, CPU and heatsink, to make sure.

Preparing the CPU for Collaborate Lab Liquid Ultra using the supplied steel wool removed the shiny mirroring off the heatsink and the golden coating on the CPU die - the CPU's inside the Mac Pro 4,1 do not have an Integrated Head Spreader.

Removing the stuff was nasty and it was difficulty to completely get rid of the silver residue and polish the surfaces again. Not to mention the risk of using liquid metal, which is electrically conductive and can damage components.

Conclusion:

I don't understand why Apple bothered with liquid metal, CPU gasket and krytox grease on the 3.2 GHz CPU. It doesn't make any sense to me, other than looking more expensive and technical. Artic MX-2, for example, was already available in 2007.

Will I be using liquid metal or Collaborate Lab Liquid Ultra again? Certainly not!

P.S.:

Regarding the CPU temperature, the fans on the Mac Pro 4,1 kick in once the CPU reaches 95 C. According to Intel however, the max operating temperature is 81 C. I've decided to adjust the fan speeds. The system also has an Apple RAID card installed and 4 SAS drives, which sits very close to the drives in slot 4, not leaving any room for cooling, which noticeably increases the PCI fan speed once installed. Not the best design if you asked me.

I guess that's it. Maybe someone will find this information useful.

I already bought Collaborate Lab Liquid Cooing Ultra. I've seen some report from 2006, comparing CLL Pro with Artic Silver 5 showing that under load there can be as much as 8 degree Celsius less using CLL Pro under CPU "load".

Meanwhile I found an interesting picture of a CPU, which may shows the gasket surrounded by Kryptox grease. I think the grease is supposed to seal the CPU and heatsink. I have not seen anyone else doing this and I wonder.

Any ideas?

Thanks for responding. I was also reading some other material discussing the use of grease vs. liquid metal and apparently liquid metal does help to provide some better cooling, although only a few degrees that may not be worth it considering the disadvantages of using liquid metal.

I was wondering though why Apple decided to use a gasket and liquid metal for the 3,2 GHz model. Does the gasket perhaps prevent that anything runs over contacts or because the liquid metal gets damaged when in contact with air? Maybe it works with some other stuff, but just for fun I would like to experiment with liquid metal. It would help however if I would know more details about the gasket.

Here's a link with more pictures:

http://forums.macrumors.com/threads/remove-liquid-metal-thermal-paste-on-2008-3- 2ghz.1914250/

It looks to me that more liquid metal is required in order to fill the gap between the CPU and heatsink, compared to what needs to be applied without the gasket.

So perhaps I could cut out some square gasket from 0.5 mm or less copper sheet. Some of the videos that demonstrate how to apply Collaborate Lab Liquid however do not use a CPU gasket and do not seal it with Krytox grease, and instead advice to provide only a very thin layer of the liquid metal. The stuff that Apple used apparently goes bad after 30 min. exposure to air. Does the same happen to Collaborate Lab Liquid and why do they not advice to use a gasket?

Good point, but the same argument may also apply to what Apple used. CLL Pro/Ultra and the stuff Apple uses looks the same, both are liquid. Wouldn't be a thin layer best in any case? What I have not seen anywhere is that CCL reacts to air, or like Apple has a 30 minute limit. Perhaps that could have been a reason for sealing the stuff with Krytox.

I've sent the manufacturer or distributor of CLL Ultra an email, questioning the matter of using a cpu gasket and sealing the heatsink, but have not received a response yet.