Xeon X5680s Now $45... October 2018

[mikehalloran]

Gravity Jim wrote:

I've also seen reports of re-pasting the Northbridge with Arctic Silver instead of the factory crud and seeing a 6ºC reduction just from that.

I have pulled quite a few CPUs over the years, taken off the heat sink 'pads', used a little Arctic Silver and put things back. Many times, with PPCs, that was all the fix it needed.
http://www.arcticsilver.com/as5.htm

There are copper heat sink compounds that are nearly as good but for < $8 for enough to cover 16 sq", why bother?
http://www.amazon.com/Arctic-Silver-The ... B000OGX5AM

They have detailed instructions for applying Arctic Silver depending on the CPU:
http://www.arcticsilver.com/intel_appli ... ethod.html#

[Gravity Jim]

To double back:

I thought I had it nailed looking at Temp Monitor, but now I'm concerned that I may have mis-applied the thermal paste. I downloaded Macs Fan Control, and while I don't know which program is accurate, I'm getting different readings than I do from TempMonitor. It's showing CPU A at a comfy 47ºC and the associated heat sink just a few degrees cooler at 45º. But CPU B is running at 65ºC, but it's heat sink is much cooler, at 35º.

This would seem to indicate that the heat transfer from CPU B to it's heat sink isn't working well. Am I right?

ALso, I cooled my Northbridge down another 10 degrees by setting Boost A to speed up if the chip gets hotter than 59º. Working great and the sound is barely audible.

[HCMarkus]

Because CPU B follows CPU A in the front to back airflow of the Mac Pro, B will typically run somewhat hotter. But the difference between CPU and (way lower) heatsink readings suggest something funky; I'd try some other temperature monitoring software; perhaps the fan software is messed up. Or could be a bad sensor or incomplete connection. Or you can just pull B, repaste and retest; not too hard on the 5,1. If you want to get really scientific, swap the CPUs in the A and B positions.

EDIT: I could have A & B backwards in my mind. I have an excuse: Mine is a single CPU MP.

[Gravity Jim]

It might be a bad sensor, too... The two CPUs are very close in temp at the diode and at the heat sink... just different at the core sensor. I'll likely pull it and check one of these days... I do want to redo the paste on the Northbridge anyway.

[HCMarkus]

Just saw this Jim…

http://forums.macrumors.com/showthread. ... 705&page=5

Post 109 sounds familiar. Subsequent posts may offer a solution for you.

[James Steele]

By the way… I saw this guy's video and noticed that he replaced a de-lidded processor with a lidded one. I was under the impression that's not a good idea. I don't think he added a spacer, so the risk is over tightening perhaps? I would think I'd want to do the de-lidding if I had a MacPro4,1 eight core.

http://youtu.be/Ng34AVZS8Aw

[HCMarkus]

By the way… I saw this guy's video and noticed that he replaced a de-lidded processor with a lidded one. I was under the impression that's not a good idea. I don't think he added a spacer, so the risk is over tightening perhaps? I would think I'd want to do the de-lidding if I had a MacPro4,1 eight core.

http://youtu.be/Ng34AVZS8Aw

Two ways to skin a cat.

[James Steele]

So there's nothing wrong with this? I think maybe you'd need to put in a spacer or something right? Or just not tighten down too much?

[Gravity Jim]

Researching my own upgrade, I read several accounts of users who replaced delidded chips with lidded ones with no ill effects. Seems to me that if you went slowly on the final turns, alternating bolts as you went, that it would be obvious when to stop... The bolts would be tight but not overtorqued.

[Michael Canavan]

So there's nothing wrong with this? I think maybe you'd need to put in a spacer or something right? Or just not tighten down too much?

OK well let's be strait forward here about all this, none of this is sanctioned by Apple, and they definitely would't recommend any of it. I have a good friend who spent years doing QA on Apple MacBooks and Mac Pros, I ran all this by him in December and he was like ??? The advantage of an unmodded Mac Pro is these kinds of questions as far as lifespan of your computer etc, are answered by QA guys running stress tests on them, and reading reports back, QA and engineers working with certain manufacturers that provide worthy product etc.


That said the extra power these CPUs draw isn't going to kill anything, the basic architecture of the Mac Pros can and seems to handle these mods just fine. My point was if you can de-lid the chips why wouldn't you want to? People are modding their cases to fit lidded chips and undoubtedly they run slightly hotter with that extra chunk of metal between them and the aluminum heat sinks.

The danger is roughly the same you can kill your CPU board, (it's much easier to crush the CPU into the connecting pins with lidded chips in the 09 models), or you can kill your chip de-lidding it, both around $400 to replace. I think the double sided razor blade, inner cardboard from masking tape, and iron method is relatively fool proof, plus it's matching Apple specs on what fits in that machine.

Researching my own upgrade, I read several accounts of users who replaced delidded chips with lidded ones with no ill effects. Seems to me that if you went slowly on the final turns, alternating bolts as you went, that it would be obvious when to stop... The bolts would be tight but not overtorqued.

Gravity, I would disagree about the turns, it's not that obvious when to stop, the thread paste remaining on the poles makes it a bit difficult. I at first had them both way to loose, and got a black screen at startup, had to go much tighter. Everything was fine after that, but I thought about how tight I went and realized I was really glad I didn't go with modding the case.

[James Steele]

Yeah... my gut tells me if I were to do this on a 4,1 eightcore, I'd de-lid the processors. It was obvious when I replaced my heat sink on my 4,1 quad (which does not require de-lidding) that the heatsink seemed designed so that you could no longer tighten it down further when it was in the correct position.

[toodamnhip]

I am about to do my computer. I have seen much concern about heat but no one has mentioned "tinting" your heat sink for a shorter break in period of the Aertic Silver. I am going to "tint" my heat sink and chip... Here are the instructions:
Tinting the Heatsink and Metal cap:
Why tint the heatsink and metal cap? Simply put, it will lessen the break-in period. If the break-in period is reduced you will achieve maximum performance in less time (To learn more about the break-in period for your Arctic Silver's product please see page 6).Here is the list of things you will need to tint your heatsink and metal cap:

 Your new thermal compound

 An old credit card or a piece of hard

plastic with a straight edge

 Coffee filter or lens cleaning cloth.
Squeeze enough thermal compound onto the center of this area to create a small mound. By working the plastic tool (old credit card) back and forth in all directions (See green symbol in photo HS1) you will smooth out the compound and work it into heatsink. This will ensure optimum filling of the microscopic valleys in the metal where the CPU cores will contact the heatsink.

Important! DO NOT smooth or apply the compound with your bare finger, you will contaminate the surface (skin cells, and body oil again). After you have thoroughly worked the thermal compound into the surface of the heatsink, remove the excess compound by wiping it away with a coffee filter or a LINT FREE cloth. DO NOT use any solvent or fluid to clean the surface or you will reverse what you just accomplished. Notice that the base of the heatsink is slightly discolored even after the entire compound would seem to have been removed. The discoloring you see is the thermal compound inside the microscopic valleys of the heatsink. Now tint the surface of the metal cap the same way. You won't need to cover the entire surface of the metal cap, just cover about the size of a quarter in the center of the CPU metal cap.

[danworks]

Hi there,

I've started to think about upgrade computer which I know it'll be a big pita if I consider the PCI-424 and related interfaces, UAD2, 4 internal SSD & HD to box … something not too far from €10,000. Plus, all the stress of re-installing everything from scratch and check them all. I was writing about this and I had a very unexpected but itchy suggestions from James - thank you man … the processor upgrade

I did it once ages ago on my old black Wallstreet PowerBook, installed a new G3 processor with flash firmware update, a incredibly stressful operation that at the end worked great. Now here I am again thinking at something that I didn't even know could possible to perform.

I've read many post in this thread (not all of them) then I went to Ebay, read other pages here 'n there and I've got some questions:

1- if anything goes really wrong can I safely go back to the stock processors?
2- the folks who had fan noise problems did they solved it?
3- apart the geekbench results which are amazing, is there any "large VIs orchestrator" that tested the upgraded machine on real life - real work environment?

Thank you so much fm the other side of the Atlantic ocean …

ciao

P.S.
I've got a mid 2012 - model MacPro5,1
2 x 2.4 Ghz 6 cores Intel Xeon

[James Steele]

2- the folks who had fan noise problems did they solved it?

Hey Danilo. I reported some perceived increase in fan noise, but it went away. I'm not sure if it was the clean install I subsequently did due to an unrelated issue, but I'm okay. And yes, if you keep your old processors, you can put them back in. Also, since you're on a MacPro5,1, your upgrade should be pretty easy.

HC Markus... do you know if the 5,1 needs to install the firmware hack, or because it's newer than then 4,1, will it already recognize and utilize the hexcore chips?

[HCMarkus]

5,1 does not require the firmware hack.

The 4,1 hack (EFI update) simply installs the 5,1 firmware in 4,1 Mac Pros.