Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

x5690 running 55c under normal usage

W

wiski15b

macrumors regular
Original poster
Apr 20, 2016
118
8
Hey everyone,
I have a 2009 Mac Pro with dual de-lidded x5690's. I installed them using a single line of arctic silver 5. Everything runs fine, however, my temps during normal usage (i.e. browsing the web) stay around 55c on CPU A and about 45c on CPU B. Anyone else have this setup running at these temps? Think I should reapply the arctic silver using a different method?
Thanks,
Matt
 
The temperatures sounds very good to me. Is there any special reason makes you believe that you should re-apply the paste.
 
h9826790 said:
The temperatures sounds very good to me. Is there any special reason makes you believe that you should re-apply the paste.
Click to expand...
Only because the old CPU's were running around 38c under normal use, whereas when I first installed the x5690's about a month ago, they both stayed in a range of about 40 to 45c. Im assuming by your response these temps are normal for this setup?
 
My (delidded) dual X5680 and dual X5677 systems both ran at around 40° under light use. I would reapply the paste.
 
The average temperature depends of where you are in the world, when you stay in a warm place 45/55c is not bad.
Most of the 8-core or 12-core MP's show a 10-14c temp difference between CPU A and B.

In my opinion reapplying thermal paste on your Northbridge will improve on that...it will however never be equal, mainly because the Northbridge module is placed under the heatsink of CPU A.
 
carpsafari said:
The average temperature depends of where you are in the world, when you stay in a warm place 45/55c is not bad.
Most of the 8-core or 12-core MP's show a 10-14c temp difference between CPU A and B.

In my opinion reapplying thermal paste on your Northbridge will improve on that...it will however never be equal, mainly because the Northbridge module is placed under the heatsink of CPU A.
Click to expand...
Good info thanks. Yeah, I knew CPU A will normally run warmer than B. Think maybe 2 lines of paste would help?
 
No, I think taking off the Northbridge and changing the completely baked and dried thermal paste, for fresh one, is the solution!
And while doing that, you have to reapply the cpu anyway...just smear a light layer across the cpu...easy enough..if you don't know how big a line or how big a pea!!
Its no rocketscience you know.
 
carpsafari said:
No, I think taking off the Northbridge and changing the completely baked and dried thermal paste, for fresh one, is the solution!
And while doing that, you have to reapply the cpu anyway...just smear a light layer across the cpu...easy enough..if you don't know how big a line or how big a pea!!
Its no rocketscience you know.
Click to expand...
Yes I know its not rocket science. However, one of the reasons I made this thread is because they were only installed a month ago, so the grease is not old. Was just wondering if others with the same setup had the same temps as me. Thanks.
 
scott.n said:
No, with the delidded processors, I just covered the processor with a thin layer of paste before screwing down the heat sinks.

You might be interested in looking at this thread.
Click to expand...
Thank you. Just read through it. Didn't see any x5690's but some temps show close to mine while not under load.
[doublepost=1463615852][/doublepost]
carpsafari said:
If you don't want to hear the solution, fine with me.
Click to expand...
I do and I appreciate your input! Thanks.
 
Around 50C is actually a good figure for me. And you light use is not equals to my light use.

A better test is to stress the system.

In warm area, the CPU should reach 80-85C in a minute or two. The fan should spin up to keep them at around 83C.

If your temperature never reach 80C. Regardless because if you live in cold area, or the paste is good. Still nothing to worry about.

My W3690's Tcase max is 10C lower then the X5690 (regardless they are identical on speed and power consumption), but can still run at 80C range for 24/7 (without thermal throttling). Therefore, I will say anything below 80C for X5690 should not be a problem at all.

Also, unless your NB temperature is really high (e.g. Constantly above 80 without stress). I can't see why I would like to touch that part. The NB heatsink rivet is the well known Achilles' heel on the 4,1. Better not to touch it, unless you already have the backup plan on how to fix it.
 
h9826790 said:
Around 50C is actually a good figure for me. And you light use is not equals to my light use.

A better test is to stress the system.

In warm area, the CPU should reach 80-85C in a minute or two. The fan should spin up to keep them at around 83C.

If your temperature never reach 80C. Regardless because if you live in cold area, or the paste is good. Still nothing to worry about.

My W3690's Tcase max is 10C lower then the X5690 (regardless they are identical on speed and power consumption), but can still run at 80C range for 24/7 (without thermal throttling). Therefore, I will say anything below 80C for X5690 should not be a problem at all.

Also, unless your NB temperature is really high (e.g. Constantly above 80 without stress). I can't see why I would like to touch that part. The NB heatsink rivet is the well known Achilles' heel on the 4,1. Better not to touch it, unless you already have the backup plan on how to fix it.
Click to expand...
Thanks h9826790 thats a wealth of information. Did not know the NB heatsink rivet was an issue. Don't think ill touch it.
 
Simple:

Bring all cores to 100% load and wait around 60 minutes, then open the side of your Mac Pro and look left and right near the CPU Heatsink. If you see red LED there, your CPU goes to hot and you must re-apply the paste on the CPU with the red LED.
 
h9826790 said:
Also, unless your NB temperature is really high (e.g. Constantly above 80 without stress). I can't see why I would like to touch that part. The NB heatsink rivet is the well known Achilles' heel on the 4,1. Better not to touch it, unless you already have the backup plan on how to fix it.
Click to expand...

Well I got news for you!

This is the view you read about,...however...I recently discovered a bigger problem with many Northbridge modules.
A 'newly' made 12 core 3.46 kept crashing after ±10minutes of rendering.
After trying several options I wanted to tackle the NB, despite that the temperature of the NB never got over 70C...and what do you think? Right, no more crashes!

The thermal paste was baked liked cement and didn't have any function, the rivets where OK (I changed them anyway).

The second 12core I made for the same guy, crashed exactly like the first one!
Applied the fix there and voila no more crashes.

Also this heatsink was loose on top of the NB.

I think more than 90% of 4.1 MP's will have thermal paste with no function...most will not crash because they are not used as intense.
 
carpsafari said:
Well I got news for you!

This is the view you read about,...however...I recently discovered a bigger problem with many Northbridge modules.
A 'newly' made 12 core 3.46 kept crashing after ±10minutes of rendering.
After trying several options I wanted to tackle the NB, despite that the temperature of the NB never got over 70C...and what do you think? Right, no more crashes!

The thermal paste was baked liked cement and didn't have any function, the rivets where OK (I changed them anyway).

The second 12core I made for the same guy, crashed exactly like the first one!
Applied the fix there and voila no more crashes.

Also this heatsink was loose on top of the NB.

I think more than 90% of 4.1 MP's will have thermal paste with no function...most will not crash because they are not used as intense.
Click to expand...

Thanks for your sharing. I trust you. However, it doesn't make any sense to me. Unless the number is not accurate, or when re-apply the thermal paste, you accidentally fix something else (e.g. By reseating the CPU / RAM?).

The NB max temperature rated over 100C, 70C is relatively nothing, and many post in this forum shows their 4,1's NB running over 70C without any problem (my 4,1 did this for many years, until I recently buy a duster to clean it).

On the other hand. Unless the OAT is really low, or the fan spin fast. If the NB heatsink is loose, and the thermal paste is totally dried. I don't think the temperature can maintain below 70C.

Anyway, I think we better advice OP that if he want to touch the NB, he better has the plan to replace the rivet (like what you did). Otherwise, this may create more problem then solve anything (his machine doesn't suffer from any crash at this moment, nothing need to be fixed).
 
I have tested a few thousand GPUs in a variety of 4,1s with a variety of CPU upgrades.

Don't even know where the NB heatsink is.

And I do mean 1000's of tests. Each and every one a minimum of one run through of Uningine Valley. With the AMD "Space Heater" GPUs I upped it to a min of 2 runs before I finally realized they were dying too fast to sell and removed them from our store.

If there is a widespread problem with 4,1 NB chips I have been blissfully unaware with 4 different machines.
 
  • Like
Reactions: flehman and leon771
I know this is new to most.
What MVC says doesn't mean anything,...as I stated with these 12-cores I found out for the first time the NB made all the difference. Its just that most will never push the MP to its limits, I know I never did!

I know for 100% the thermal paste was what made the difference!

If the NB heatsink is loose, and the thermal paste is totally dried. I don't think the temperature can maintain below 70C.

Well, think again, I know I found something new. Here's the iStatMenus, just before the crash(es).
BTW, Cinema 4D was used, rendering 100% on the CPU's.
12core shut down.png
 
Again, I am not saying you lie, but just may be we are in the wrong direction. Re-paste seems doing nothing actually (assume NB is not the root cause), however, if you did reseat the CPU / RAM etc during the re-paste process, that may be the action that actually fix the problem (re-seat CPU sometimes did the tricks).

All we can tell at this moment is you fix 2 crash by the "re-paste process", but cannot conclude that it's actually the NB cause the trouble. Do you know what I mean? We simply don't have enough data to conclude that. Especially the NB temperature and the NB heatsink temperature both looks rational.

If the thermal paste is not working at all. With that 25xx RPM, I will expect the heatsink is much cooler than the NB chip (because all heat are trapped inside the chip but cannot transfer to the heatsink effectively). Also, the NB can reach 100C (or over).

I never touch my NB (my Mac works fine, I don't want to over do the maintenance and kill it), but I Am really interested in the difference before and after the re-paste. Could you mind share the temperature that's after the re-paste? And for better comparison, that's better also under 100% stress with system ambient around 31C.

And which solution you use to replace the original rivet?

Just some sharing.

I have a 7950 and a R9 280 inside my 4,1. The R9 280 shows some artifact occationally about 2 months ago (randomly a small block of square on the screen in different colour). That's looks like a graphic card failure.

And what I did is pull out the card, blow the dust from the exhaust (pushing them backward), blow the slot. And reinstall the card.

After this process, no more artifact for more than a month now. Can I conclude that I blow the slot fix the issue? Or I blow the card fix the issue? I cannot, in fact, I re-seat the card as well, and also including a unintentional SMC reset. All these may contribute to the fix, but I don't know which part is the key.

The card never shows overheat, it never works above 70C, however, this "blowing out the dust from the card" process fix the issue. I can only say this process fix the issue, but not any particular part of the process fix the issue.

To others, I remember that when I bought my HIS IceQ R9 280 last year. Someone said HIS card usually is a trouble maker. I am happy to say that so far I only has 2 problems.

1) the artifact mentioned in above. And that's fixed. I suspected that when I opened the case and take out the PSU for cleaning few months ago (really dusty after almost 7 years of 24/7 usage in a higher air polluted city), I actually leave some dust in the slot, and that cause the occational artifact. Of course, may be the dust inside the card really affect some electronics. I can't find out the actually cause, but anyway that's completely fixed now.

2) despite it has the standard port layout. If I boot from the Mac EFI ROM (either only the 7950 Mac Edition with EFI or the R9 280 also boot from the EFI ROM), the HDMI port on the R9 280 won't work (but if both card boot from PC BIOS, then the HDMI port is fine). That's no big deal for me. I have 2 cards and plenty of port. Also, all I need to do is just flip the switch. In fact, now I connect my TV to that port, and intentionally boot with EFI enabled to disable that port when I don't need the TV. Because even though the TV is off, OSX still consider the screen is there, sometimes my mouse cursor will move to that offed screen which is not very handy. And now I can control it via a switch, but no need to touch the cables (I know I can use software to disable the screen, but that's not a good solution for me, because the setting will be reset after profiles switching, and I switch between profiles multiple times a day).
 
I have upgraded ±20 MP's the in last 6 months, its not that I'm doing this for the first time, you know.

I always run iStat before upgrading, to know what temperatures where there before the procedure! Very important!!

The spread between CPU A and B being 'always' 10-14c with ALL 0cto's I upgraded, made me aware about the NB.
Nobody really wants to tackle the NB if not really neccesary (broken rivet),...neither do/did I!!
When the temperature is staying below 80c why bother??

So finally a few weeks ago, when I took of the heatsink of CPU A, I saw a rivet snap...so I had no choice and had to fix it. Also this MP had a 12c spread between CPU A and B. After the fix the spread was only 5c.

I use Gelid GC Extreme, which looks identic to the paste that Apple uses.
When I remove a recently placed CPU, I have to twist the CPU to let go from the heatsink,...I explain this in detail to tell you I know what I'm doing!

So I can tell you 100% sure it is NOT an issue applying the paste!!
No accidental fix!!

The temperatures after the NB fix stay the same, just the Mac does not crash anymore!

The second 12-core I built for them, a week later, crashed in exactly the same way.
We swapped trays to see if it is a hardware issue,...nope...same problem.

The screenshot was taken ±2 minutes before the crash,...all is the same...but the CPU A can reach 76c now, without the Mac crashing.

I know this is NEW for everybody....the NB fix is also not hard to do, there has been written enough about that. I used the Chinese rivets.

PS, you are wrong stating that the CPU can reach 80c,...78c is shut-off for the X56...
 
carpsafari said:
I have upgraded ±20 MP's the in last 6 months, its not that I'm doing this for the first time, you know.

I always run iStat before upgrading, to know what temperatures where there before the procedure! Very important!!

The spread between CPU A and B being 'always' 10-14c with ALL 0cto's I upgraded, made me aware about the NB.
Nobody really wants to tackle the NB if not really neccesary (broken rivet),...neither do/did I!!
When the temperature is staying below 80c why bother??

So finally a few weeks ago, when I took of the heatsink of CPU A, I saw a rivet snap...so I had no choice and had to fix it. Also this MP had a 12c spread between CPU A and B. After the fix the spread was only 5c.

I use Gelid GC Extreme, which looks identic to the paste that Apple uses.
When I remove a recently placed CPU, I have to twist the CPU to let go from the heatsink,...I explain this in detail to tell you I know what I'm doing!

So I can tell you 100% sure it is NOT an issue applying the paste!!
No accidental fix!!

The temperatures after the NB fix stay the same, just the Mac does not crash anymore!

The second 12-core I built for them, a week later, crashed in exactly the same way.
We swapped trays to see if it is a hardware issue,...nope...same problem.

The screenshot was taken ±2 minutes before the crash,...all is the same...but the CPU A can reach 76c now, without the Mac crashing.

I know this is NEW for everybody....the NB fix is also not hard to do, there has been written enough about that. I used the Chinese rivets.

PS, you are wrong stating that the CPU can reach 80c,...78c is shut-off for the X56...
Click to expand...

No, no, no, I am not saying the thermal paste on the CPU is the root cause, but since you are forced to reseat the CPU on every NB re-paste, the "reseat" may accidentally fix the issue (e.g. the pin has bad contact on the previous installation).

May I confirm that the NB temperature are rough the same before ad after the re-paste? If yes, then it means the old paste was actually working very well. (Since we are talking about re-paste the NB, I was asking the NB temperature, but not the CPU temperature. Sorry I didn't make it clear on the last post).

And I can sure you are 100% wrong in the CPU temperature. X56xx won't even throttling at 80C, they won't shut down at 78C.

84C on this one without shut down (In fact, OP confirms the system 100% stable on his last post)
https://forums.macrumors.com/threads/xeon-x5690-temperatures-and-fan-speed.1958060/

98C on this one without shut down (he then fix the heatsink and re-paste, the CPU run at 81 without shutdown)
https://forums.macrumors.com/thread...pgrade-recommendations.1888116/#post-22571200

92C without shutdown
https://forums.macrumors.com/threads/mac-pro-2010-from-qc-to-6-core….1904232/

88C and 86C without shutdown (with screen capture)
https://forums.macrumors.com/thread...fference-of-10-degrees.1767463/#post-20556695

82C on my own W3690 without shutdown (W3690's T-case max is 10C lower than X5690, only 68C)
https://forums.macrumors.com/thread...fference-of-10-degrees.1767463/#post-20556926

May be you mixed up T-case temperature with the T-diode temperature. I explain that a little bit in this post.
https://forums.macrumors.com/threads/xeon-x5690-temperatures-and-fan-speed.1958060/#post-22605468

Anyway, it's not about how many systems you've upgraded. But how much you study and know about the system / terms, etc. I only own one Mac Pro and only upgrade my own Mac Pro, but I know the X5690 won't shutdown when T-diode reach 78C, and can easily run stably at 80C (proved by the above post).

P.S. I am not trying to argue with you, or I want to win. But I want to find out / let the others know the truth via debate.
 
Last edited:
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back