Mac Pro with Westmere x5690 (3.46GHz)

[DEMinSoCAL]

I have a paid of X5690's in mine and if I reset the SMC and leave all my fans at default, my IOH Heatsink will also hover around 65C and be in the red with iStat. I have dealt with it by setting the min rpm of my intake/exhaust fans to 850 and that makes all the difference for me - at idle IOH Heatsink is now 56C and my two booster fans at idle are running at 1100rpm.

(Pro is 2009 -> 5,1)

Does this mean that X5690 equipped Mac Pro's will be louder than with something less (like X5670)?

Is the temp issue just with DUAL X5690's or would I have to raise the fan speed for a single X5690? I wanted to upgrade but not at the expense of a quiet Mac.

 

[h9826790]

I am now running a W3690, the thermal characteristic should be very close to a single X5690. There is no problem for my flashed 5,1 to handle this CPU.

I did run this machine for few months with stock fan control. No issue so far.

And now I use the Fan control software to fine tune it a bit. Make it run few degrees cooler under max loading without sacrifice the quietness.

[Full load]

 

[matthewtoney]

Does this mean that X5690 equipped Mac Pro's will be louder than with something less (like X5670)?

Is the temp issue just with DUAL X5690's or would I have to raise the fan speed for a single X5690? I wanted to upgrade but not at the expense of a quiet Mac.

I previously ran a single-CPU tray in this Pro with a W3690 and with it I never altered any minimum RPMs on any of the fans - I left the SMC control stock and never had a problem at all. (nor did I have any temps showing in the red in iStat) I think the IOH Heatsink issue is with a pair of the higher power CPUs on a dual tray.

Normally the intake/exhaust fans have a min RPM of 600 and I've personally only bumped that up to 850 - I can't tell a bit of difference in the noise myself but everyone's got a different level of sensitivity to that I suppose. For me, I notice the big difference in noise in the Pro when the 2 booster fans start getting revved up more than anything else.

 

[drewsof07]

Does this mean that X5690 equipped Mac Pro's will be louder than with something less (like X5670)?

Is the temp issue just with DUAL X5690's or would I have to raise the fan speed for a single X5690? I wanted to upgrade but not at the expense of a quiet Mac.

I doubt it. My single socket is running about 5-10c cooler with the 5680 than the weaker 3530 it replaced. I suspect cleaning up the gooey mess of factory applied thermal compound helped.

 

[h9826790]

I doubt it. My single socket is running about 5-10c cooler with the 5680 than the weaker 3530 it replaced. I suspect cleaning up the gooey mess of factory applied thermal compound helped.

45nm vs 32nm, hard to compare actually.

 

[tinycg]

I think the IOH Heatsink issue is with a pair of the higher power CPUs on a dual tray.

Ok that would be inline with what I'm seeing. I'll bump up the fans and see. Should I be worried that I'm causing lasting damage?

 

[Larry-K]

They're Finally In!

No time better to install Microprocessors than a rainy night, so I finally grabbed my pair of pasty, hairy, unstickered X5690's and slapped them in my old 12-core refurb.

So it appears Aventis is off the hook for an RMA, but one more justified swipe at them, is for having a couple of chunks of dried thermal paste in the anti-static bag of one chip, especially since the only thing protecting the chip was a thin, unattached sheet of Uline foam. I don't care how menial a job pulling out, cleaning and bagging chips is, show a little professional pride and do the job right.

Now for a swipe at Apple, looks like whoever put the thermal paste on at the factory put a thick slice of cheese on, instead of .2cc of goop. I think they were intent on coming absolutely as close as they could to shorting out the contacts without actually doing it, I spent more time cleaning the old 2.4s than doing anything else.

That being said, I used the Apple Technician's Guide method for applying thermal paste (haven't heard anyone else using that technique, curious if anyone else did it that way).

Got the Bresink System Monitor and only had time to let it idle for 15 minutes or so, but it was running at 43°, which seems awfully cool, but I'll take it. Heard this was superior to iStat, but I'm open to suggestion.

 

[DEMinSoCAL]

That being said, I used the Apple Technician's Guide method for applying thermal paste (haven't heard anyone else using that technique, curious if anyone else did it that way).

Could you briefly describe what that method is? I've read the review of the various paste application methods (rice, line, X, spiral, etc.), and wondered what Apple recommends?

 

[h9826790]

Could you briefly describe what that method is? I've read the review of the various paste application methods (rice, line, X, spiral, etc.), and wondered what Apple recommends?

Basically draw a "No go sign" on the CPU. I follow the guide and use that method as well.

 

[DEMinSoCAL]

Basically draw a "No go sign" on the CPU. I follow the guide and use that method as well.

Got it. Thanks. Here's a good link to an article comparing all the different application methods. Oddly, the "No-Go sign" is not in the review.

 

[h9826790]

Got it. Thanks. Here's a good link to an article comparing all the different application methods. Oddly, the "No-Go sign" is not in the review.

It seems Apple's idea is let the vertical line handle the inner part and leave the circle handle the outer part.

As you can see from the test, a vertical line basically will turn into a circle when compressed, but nothing to take care the outer part.

I follow Apple's method, draw the no go sign vertically with a very thin line, it works very well. As you can see from my post above, the idle heat sink temperature is just 1 degree below the CPU temperature when idle, very effective.

 

[lowendlinux]

I've always used rice, guess I'll try x the next time.

 

[DEMinSoCAL]

It seems Apple's idea is let the vertical line handle the inner part and leave the circle handle the outer part.

As you can see from the test, a vertical line basically will turn into a circle when compressed, but nothing to take care the outer part.

I follow Apple's method, draw the no go sign vertically with a very thin line, it works very well. As you can see from my post above, the idle heat sink temperature is just 1 degree below the CPU temperature when idle, very effective.

What's your choice of paste? I use Noctua NT-H1, but it's very thick and hard to manipulate.

Also, when you say "no go sign vertically", you mean a circle with a straight up/down line through it? I suppose horizontal would work too, since the heatspreader is a square? Also, does the line extend out of the circle or is it entirely in the circle?

 

[h9826790]

What's your choice of paste? I use Noctua NT-H1, but it's very thick and hard to manipulate.

Also, when you say "no go sign vertically", you mean a circle with a straight up/down line through it? I suppose horizontal would work too, since the heatspreader is a square? Also, does the line extend out of the circle or is it entirely in the circle?

I use AS5, if you check their website, they suggest a vertical line for X56xx or W36xx CPU, which is part of agree Apple's method. Even though the CPU is a square, the architecture inside the CPU (or the thermal characteristic) is not necessary rotational symmetrical. So, a vertical line or a horizontal line may make the difference.

Anyway, the technician guide suggest a circle which made by a thin line that just smaller than the CPU IHS, and a thin vertical line that entirely inside the circle (but touch the circle on both ends).

It looks like a no go sign inside the technical guide because they draw that with a CPU which rotated 90 degrees.

 

[tinycg]

I've adjusted the fan control a bit, experimenting a little bit.

Adjusting the Exhaust / Intake I was able to get the IOH Heatsink down to what other people are reporting as good numbers but...

This morning I've noticed that CPU B Core 0 ProcHeat is in the red at 73C, my understanding is that is the amount relative to how much hotter the CPU can get, I've never seen that much of a gap. The higher the better? but is too high bad? This is all new to me just trying to make sure everything stays between good norms.

 

[h9826790]

This morning I've noticed that CPU B Core 0 ProcHeat is in the red at 73C, my understanding is that is the amount relative to how much hotter the CPU can get, I've never seen that much of a gap. The higher the better? but is too high bad? This is all new to me just trying to make sure everything stays between good norms.

No need to worry the red, it's iStat's problem.

 

[matthewtoney]

I've adjusted the fan control a bit, experimenting a little bit.

Adjusting the Exhaust / Intake I was able to get the IOH Heatsink down to what other people are reporting as good numbers but...

This morning I've noticed that CPU B Core 0 ProcHeat is in the red at 73C, my understanding is that is the amount relative to how much hotter the CPU can get, I've never seen that much of a gap. The higher the better? but is too high bad? This is all new to me just trying to make sure everything stays between good norms.

Someone *please* correct me if I understand this wrong, but I think iStat's "black versus red" for the "Relative to ProcHot" sensors is backwards. My understanding of "Relative to ProcHot" is that the higher the number the *cooler* the processor is, so it going red in iStat when it gets *above* a certain level doesn't make any sense to me. (mine does that same darn thing)

I've tweeted at the iStat support folks and asked for clarification on this and whether its actually something that might need to be fixed in iStat but I always get ignored...

 

[audio_inside]

That being said, I used the Apple Technician's Guide method for applying thermal paste (haven't heard anyone else using that technique, curious if anyone else did it that way)

Got link to that Apple Technician's Guide? Is it publicly accessible?

 

[tinycg]

Someone *please* correct me if I understand this wrong, but I think iStat's "black versus red" for the "Relative to ProcHot" sensors is backwards. My understanding of "Relative to ProcHot" is that the higher the number the *cooler* the processor is, so it going red in iStat when it gets *above* a certain level doesn't make any sense to me. (mine does that same darn thing)

I've tweeted at the iStat support folks and asked for clarification on this and whether its actually something that might need to be fixed in iStat but I always get ignored...

Thanks matthewtoney and h9826790 really appreciate all your help. Had the same thought that iStats was somehow backwards I appreciate the confirmation and will watch with interest if someone corrects us with new info.

 

[Larry-K]

Basically draw a "No go sign" on the CPU. I follow the guide and use that method as well.

Cool, I thought it looked too simple. although a real "no go" symbol would have a diagonal thread down the center, whereas the guide one is parallel to the long edge.

----------

Got link to that Apple Technician's Guide? Is it publicly accessible?

I'll find it, but I think if we post it we get killed or something. They took Lou's down.,

If you do a search for "2010 mac pro apple technician guide" in google, you should see an Australian site with a PDF on in near the top.

Even gives you the order to loosen/tighten heatsink screw in, and the torque setting.

 

[Larry-K]

Got it. Thanks. Here's a good link to an article comparing all the different application methods. Oddly, the "No-Go sign" is not in the review.

I think I'd avoid the "Happy Face".

I don't think it occurs to people there is actually a tech manual from Apple with this info, I heard it mentioned or I wouldn't have thought to go looking for it.

 

[audio_inside]

If you do a search for "2010 mac pro apple technician guide" in google, you should see an Australian site with a PDF on in near the top.

Awesome, thanks, found and saved it. Search instructions as good as a link!

For those of us upgrading Macpro4,1, the 2009 version of the manual is also available on that same site - not sure how it differs, but at least the number of pages is different

 

[flowrider]

Got link to that Apple Technician's Guide? Is it publicly accessible?

As was mentioned above, I posted the link here, but my post got removed and the mods slapped my hands. Really makes no sense why they would do this?

Lou

 

[Larry-K]

As was mentioned above, I posted the link here, but my post got removed and the mods slapped my hands. Really makes no sense why they would do this?

Lou

Hey Lou, sorry you got slapped on this, but I wanted you to know you really helped me and I appreciate it.

The guide is a great resource, and I wish I had one for every Mac I've ever owned.

 

[Macshroomer]

Glad the processors are working out for you Larry, mine is too and I am over the mess left by Aventis when they pulled it for sale.

I got Bresink's Hardware Monitor yesterday and all seems really good, no big differences between the heat sink and CPU temps and no higher than 124F under the heaviest sustained loads with SMC fan control off.

These sure are great machines to mod and get to know inside and out...