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

m3rob

macrumors member
Original poster
Dec 28, 2010
41
7
Hi guys,

Stupid question... Is it cooler to run my Mac Pro 4,1 with the case cover off or on? Super hot here and rendering a ton lately with it and I have to use Macs Fan Control to keep it rendering my big jobs. I upgraded my dual cpus to 2 x 3.46 6 cores. Just wondering if its any cooler with the case cover off or if the case cover helps the fans work more efficiently. Sorry I couldn't find the answer anywhere probably because its a silly question. :)

Rob
 
One would like to think that the Apple designers engineered for ultimate air flow with the case in it's normal condition, closed!
Having the side door off totally defeats the front (inlet) to back (outlet) air flow design.
This is true, however with side panel off, hot air would have a faster way of escaping the inside of the case keeping the components cooler without taxing the fans too much.
 
Running with the side panel makes heat dissipation slightly worse (see this thread post #22). One could hypothesize that it has to do with the fans having a less directed path to push hot air out of the case.
 
I would have said the side panel is best left on, purely for the reason the intake and exhaust fan draws air from front to back across all components, with the side case off intake air takes shortest route to exit out side panel, and exhaust fan draws air from side of case with panel off as its the easy path for air to enter exhaust fan. which means no air flow drawn from front to back over components on logic board or CPU board.
 
With an active cooling profile that links macs fan control to the temps of the RAM slot diodes, ddr3 thermal ram throttling techniques implemented with the x58 and 5500 chipsets can be avoided.

Apples default profile keeps the cpu bay within tolerances, but warm enough to penalize maximum bandwidth. When it comes to ram temps, running the system without the side panel makes a bad situation worse.

YMMV.
 
With an active cooling profile that links macs fan control to the temps of the RAM slot diodes, ddr3 thermal ram throttling techniques implemented with the x58 and 5500 chipsets can be avoided.

Apples default profile keeps the cpu bay within tolerances, but warm enough to penalize maximum bandwidth. When it comes to ram temps, running the system without the side panel makes a bad situation worse.

YMMV.


leave your side panel on, speed up your intake to 1150 rpm speed up your exhaust to 1150 rpm, this will not turn your Mac into a vacuum cleaner. if you don't have heat spreaders on your ram, try some, they really can help if fitted properly. this should ensure you get maximum bandwidth and not temp controlled bandwidth.

Air flow through the case is important to cool everything, the Mac Pro 4.1 and 5.1 dual CPU with an extra 4 ram slots and 2 CPU's to get full potential need to be cooled, even using high grade thermal paste like thermal grizzly also helps for CPU temps and keeping that Northbridge cooler. after all we like speedy Mac pro's
 
Apple never put 2 x 130W TDP XEONs in cMP though. They used 2 x 95W max. They work, but get ready for some extra noise.

One would like to think that the Apple designers engineered for ultimate air flow with the case in it's normal condition, closed!
Having the side door off totally defeats the front (inlet) to back (outlet) air flow design.
 
leave your side panel on, speed up your intake to 1150 rpm speed up your exhaust to 1150 rpm, this will not turn your Mac into a vacuum cleaner. if you don't have heat spreaders on your ram, try some, they really can help if fitted properly. this should ensure you get maximum bandwidth and not temp controlled bandwidth.

Air flow through the case is important to cool everything, the Mac Pro 4.1 and 5.1 dual CPU with an extra 4 ram slots and 2 CPU's to get full potential need to be cooled, even using high grade thermal paste like thermal grizzly also helps for CPU temps and keeping that Northbridge cooler. after all we like speedy Mac pro's

Absolutely.
Apple never put 2 x 130W TDP XEONs in cMP though. They used 2 x 95W max. They work, but get ready for some extra noise.

Although Apple never shipped with high-end CPU's in the cMP, perhaps it was partly due to keeping the cMP slower than the nMP. A small increase to the speed of the Intake and Exhaust fans should be sufficient to keep dual 130w CPU's supplied with a sufficient level of cool air.

I'm in the process of upgrading a 5,1 from a single w3680 with 48GB to a dual x5677 with 96GB on Monday/Tuesday. Running in a 2009 single core, the x5677 idles at 32c and maxes out at 45c under 100% load. I'm confident that dual chips will continue to run cool with a minimal increase in noise in a dual setup.
 
Absolutely.


Although Apple never shipped with high-end CPU's in the cMP, perhaps it was partly due to keeping the cMP slower than the nMP. A small increase to the speed of the Intake and Exhaust fans should be sufficient to keep dual 130w CPU's supplied with a sufficient level of cool air.

I'm in the process of upgrading a 5,1 from a single w3680 with 48GB to a dual x5677 with 96GB on Monday/Tuesday. Running in a 2009 single core, the x5677 idles at 32c and maxes out at 45c under 100% load. I'm confident that dual chips will continue to run cool with a minimal increase in noise in a dual setup.

IMO that would be conspiracy theory to say that. They never used single W3690 either. Maybe for the insane price jump, maybe for the higher TDP, maybe both.

How come does your x5677 goes 55Cmax? Do you control fans?

X5677 is a good choice. Tho it has 130W TDP too, will work colder than X5690 in a dual CPU setup, because of the lack of two cores. Will be a safe and nice upgrade.
 
  • Like
Reactions: handheldgames
IMO that would be conspiracy theory to say that. They never used single W3690 either. Maybe for the insane price jump, maybe for the higher TDP, maybe both.

How come does your x5677 goes 55Cmax? Do you control fans?

X5677 is a good choice. Tho it has 130W TDP too, will work colder than X5690 in a dual CPU setup, because of the lack of two cores. Will be a safe and nice upgrade.

Thanks, I'm hoping for a successful upgrade. To clarify, my x5677 hits a max Diode temp of 45c under load.

I control the fans with MacsFanControl. Intake/Exhaust is controlled by RAM temps, running around 900 rpm.
BoostA is controlled by CPU Temps around 1100 rpm. The profile is louder at startup due to the default profiles higher temps, quickly quieting down when the temps lower.

Switching from Thermal Paste to Innovation Cooling graphite thermal interface material I'm achieving a a 1c variance between Diode and Heatsink at idle 32c/31c and a 5c variance at load 45c/50c in a single core install. I'm looking forward to see how it works in a dual x5677 CPU setup.
 
They will work perfectly. Feel free to use auto fan speed Apple engies set and enjoy the silence. These devices have been used under constant heavy load for long years.

On the TDP subject again. If the small heatsinks of dual CPU trays were good or silent enough for 130W monsters like W3680 or W3690, Apple would not spend any more money for bigger heatsinks in single CPU trays.

Thanks, I'm hoping for a successful upgrade. To clarify, my x5677 hits a max Diode temp of 45c under load.

I control the fans with MacsFanControl. Intake/Exhaust is controlled by RAM temps, running around 900 rpm.
BoostA is controlled by CPU Temps around 1100 rpm. The profile is louder at startup due to the default profiles higher temps, quickly quieting down when the temps lower.

Switching from Thermal Paste to Innovation Cooling graphite thermal interface material I'm achieving a a 1c variance between Diode and Heatsink at idle 32c/31c and a 5c variance at load 45c/50c in a single core install. I'm looking forward to see how it works in a dual x5677 CPU setup.
 
  • Like
Reactions: handheldgames
They will work perfectly. Feel free to use auto fan speed Apple engies set and enjoy the silence. These devices have been used under constant heavy load for long years.

On the TDP subject again. If the small heatsinks of dual CPU trays were good or silent enough for 130W monsters like W3680 or W3690, Apple would not spend any more money for bigger heatsinks in single CPU trays.

I would use Apple's default profiles IF the additional heat did not slow down RAM performance. FWIW... I'm also running a 2009 with a 990x that idles at 32c. The existing single core heatsink is more than adequate. :)
 
I would use Apple's default profiles IF the additional heat did not slow down RAM performance. FWIW... I'm also running a 2009 with a 990x that idles at 32c. The existing single core heatsink is more than adequate. :)

They will be fine on temps, I am running dual X5690's and with a tweak on the intake and exhaust fan profile in 23c ambient I idle 28c and the smaller heatsinks are no issue at all. keeping the ram cool is also a top priority, sure can get toasty that ram:)
 
  • Like
Reactions: handheldgames
My house has been over 36 degrees Celsius for the last few weeks, my cMP hasn't skipped a beat...leave the door closed.
Definitely.

Unless you have a room fan blowing directly into the open case - trust the Apple engineers to move the air with the cover on.

Do, however, make sure that the air at the front of the cabinet is as cool as possible. Don't let it recycle hot air from the back of the system.
 
  • Like
Reactions: handheldgames
They will be fine on temps, I am running dual X5690's and with a tweak on the intake and exhaust fan profile in 23c ambient I idle 28c and the smaller heatsinks are no issue at all. keeping the ram cool is also a top priority, sure can get toasty that ram:)

Impressive. Now that's what I call a well-performing cooling setup.
Can you share your fan profile? It would be interesting to see your approach.

For comparison, this is the 990x. More or less the same as one of your x5690's. Running with 56GB 16x16x16x8 RDIMMS.

Intake/exhaust is based on DIMM3. 30c / 50c
BoostA, based on CPU Diode 30c / 55c
PS, based on PS1 34c / 45c

window 8-11-185.07 PM.png
 
Here is my Mac fan control panel settings, it stays nice and cool this way and don't have any issues at all with heat.



Screen%20Shot%202018-08-12%20at%2007.33.52_zpsjkzngnko.png
 

Attachments

  • Screen Shot 2018-08-12 at 07.33.52.png
    Screen Shot 2018-08-12 at 07.33.52.png
    178.6 KB · Views: 152
Last edited:
  • Like
Reactions: handheldgames
Impressive. Now that's what I call a well-performing cooling setup.
Can you share your fan profile? It would be interesting to see your approach.

For comparison, this is the 990x. More or less the same as one of your x5690's. Running with 56GB 16x16x16x8 RDIMMS.

Intake/exhaust is based on DIMM3. 30c / 50c
BoostA, based on CPU Diode 30c / 55c
PS, based on PS1 34c / 45c

View attachment 775475

I think PSU1 is the PSU ambient temperature, and PSU2 is the PSU components temperature. If you want better cooling for the PSU, base on the components temperature may make more sense.
 
Since your on the subject of heat I have a related question.

How impactful are heatsinks on RAM?

I've got 3 sticks of 16GB ECC Hynix in my system. They have no heatsinks or additional cooling. I didn't think anything of it when purchasing/installing. I don't recall the Apple Installed memory having heatsinks either.

Hynix DDR3-1333 16GB/ 2Gx72 ECC/ REG Hynix Chip Server Memory
 
Since your on the subject of heat I have a related question.

How impactful are heatsinks on RAM?

I've got 3 sticks of 16GB ECC Hynix in my system. They have no heatsinks or additional cooling. I didn't think anything of it when purchasing/installing. I don't recall the Apple Installed memory having heatsinks either.

Hynix DDR3-1333 16GB/ 2Gx72 ECC/ REG Hynix Chip Server Memory


Keeping the memory cool means your bandwidth won't suffer by decreasing with heat, Heat spreaders do function well as long as they are fitted properly using either a non conductive thermal paste or thermal pads. The idea is to take heat away from the memory chips. My heat spreaders get quite warm so they are dissipating heat nicely.

most high end memory is fitted with heat spreaders of varying designs these days, for cometic effect as well as thermal efficiency. the worst fitment of heat spreaders I have seen was on OWC ram. terrible fitment using thermal paste that wasn't even fully over some chips.

I am sure you could fit some easily to your own ram.
 
  • Like
Reactions: crjackson2134
Since your on the subject of heat I have a related question.

How impactful are heatsinks on RAM?

I've got 3 sticks of 16GB ECC Hynix in my system. They have no heatsinks or additional cooling. I didn't think anything of it when purchasing/installing. I don't recall the Apple Installed memory having heatsinks either.

Hynix DDR3-1333 16GB/ 2Gx72 ECC/ REG Hynix Chip Server Memory

I'm also running used samsung server pulls from ebay, averaging $40 each. I haven't seen any ddr3 16gb rdimms with heatsinks.

With the side door installed, your ram without heatsinks should have enough cooling. Increasing the speed of intake / exhaust can help keep your server memory a bit cooler.

You can try slightly Increasing the intake/exhaust speeds from the default 600 rpms, finding the sweet spot of increased cooling without additional noise. Perhaps @ 850 rpm
 
  • Like
Reactions: crjackson2134
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.