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

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
Dear macrumors members and guests!

I'm starting this thread mainly to obtain some nice statistics on 15" Macbook Pro with Retina Display, which may help others to decide on a purchase or determine if their machine's cooling system is working within specs. Many of us uses their machines for computation-heavy tasks and knowing that cooling system is capable is very important for system stability and reliability. If the CPU runs at 105C, it'll start to throttle down to stay at that temperature, since it is TJMax for almost all Intel CPU nowadays. It will not only hit the performance, but running at such high temperatures will also reduce your computer lifetime and increase the chance of failure.


The idea is to run 8 threads from terminal that will put 100% load to all 8 virtual cores (4x real cores + hyper threading) with fans at their highest RPM. Then you need to measure CPU core temperature and enclosure/battery temperature after 8-10 min and post the results.

Here's the protocol:
1. Install iStat Menus (http://bjango.com/mac/istatmenus/)
It's free for 15 days and is actually the best monitoring program I know.
2 (optional). Set all fans to the highest RPM. Go to Sensors -> Edit Rules and create a new one. After that you can select it from menubar icon!
It's not required as both fans will fish all the way to 5500rpm once you 100% load the CPU.

3. Open terminal and execute this command:
yes > /dev/null
create a new terminal window and repeat 8 times. Now you have 8 processes that load your CPU up to 800%.

screenshot20121121at838.png



Here's what iStat menubar looks like on the 2.6Ghz model:
rmbptemp.png



I had my hands on three different RMBP machines, and here're the results:


1st:
2.6Ghz, 16Gb RAM (July 2012):

CPU: 105C
Battery / Enclosure: 30C / 29C

2nd:
2.3Ghz, 8Gb RAM (September 2012):

CPU: 83C
Battery/Enclosure: 30C / 30C

3rd:
2.6Ghz, 16Gb RAM (November 2012):

CPU: 103C
Battery/Enclosure: 30C / 30C



All measurements are takes after at least 10min with fans at full rpm.
P.S. The 1st and the 3rd are done with both stock and replaced thermal compound. 2nd (2.3Ghz model) was completely stock.
 

b0fh666

macrumors 6502a
Oct 12, 2012
954
785
south
I see no mention of room temperature in the 'methodology', not considering that makes all this testing useless :D

cheers
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
I see no mention of room temperature in the 'methodology', not considering that makes all this testing useless :D

cheers

I agree, but there's no direct "room" temperature here since not everyone has a thermometer ready to hand.

However, both battery and enclosure sensors aren't very sensitive to the heat from the radiator so they could be used as a reference.

Ideally, one need to provide Air intake t. Nevertheless, battery/enclosure t is good enough for normalising the results in unusually hot/cold environments.
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
Even though we don't have precise air temperature readings ;) it seems that the difference between 2.3Ghz and 2.6Ghz is enormous. I can confirm that in both 2.6Ghz machines I tested it was the same stepping (Stepping 9 (09h))
according to MacCPUID.

It is still to be confirmed whether there're any different steppings available around (for 2.6 model in particular). It's also unknown if 2.3Ghz is so much cooler of it's me who is unlucky!

Considering 2.7Ghz model, if there're any users of the Ultimate configuration here, I'd really appreciate some feedback!
 

Sound Evolution

macrumors 6502
Oct 30, 2007
414
0
Netherlands
Even though we don't have precise air temperature readings ;) it seems that the difference between 2.3Ghz and 2.6Ghz is enormous. I can confirm that in both 2.6Ghz machines I tested it was the same stepping (Stepping 9 (09h))
according to MacCPUID.

It is still to be confirmed whether there're any different steppings available around (for 2.6 model in particular). It's also unknown if 2.3Ghz is so much cooler of it's me who is unlucky!

Considering 2.7Ghz model, if there're any users of the Ultimate configuration here, I'd really appreciate some feedback!

Hi, I have the 2.7Ghz. model and have thermal issues, in the sense I feel the laptop get WAY to hot.

I am willing to help and do the test. However me being a noob... :D After you put all cores to work as described above, How you stop the process?
 

Blondie :)

macrumors 6502a
May 12, 2010
698
3
Prescott, AZ
I see no mention of room temperature in the 'methodology', not considering that makes all this testing useless :D

cheers

The only way that room temp actually makes a difference is by decreasing the amount of convection from the laptop to the surroundings. Unless you're talking about a massive room temp difference, the temp difference won't actually mean much.
 

b0fh666

macrumors 6502a
Oct 12, 2012
954
785
south
really?

In my experience it can get scalding hot if the ambient is hot (and boy, it DOES get hot here in the tropics), but when it is cool you barely notice heat buildup. same type of workload, of course.

the 'ambient air' measurement is currently saying 40 celsius, but that has to be inside the laptop as it is ˜22.

cheers
 

Blondie :)

macrumors 6502a
May 12, 2010
698
3
Prescott, AZ
really?

In my experience it can get scalding hot if the ambient is hot (and boy, it DOES get hot here in the tropics), but when it is cool you barely notice heat buildup. same type of workload, of course.

the 'ambient air' measurement is currently saying 40 celsius, but that has to be inside the laptop as it is ˜22.

cheers

The reason that you don't notice the buildup as much when it's cold is due to conduction. When the bottom cover of the macbook is cooler, heat transfer from the CPU to the outside will be more efficient. But, as more heat transfers through the case, the case heats up anyways. As I mentioned before, the larger source of heat transfer will come from convection (outside air running over the bottom cover of your macbook). That's why people have been investing in fans like this one: http://www.moshimonde.com/product/zefyr-2.aspx
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
really?

In my experience it can get scalding hot if the ambient is hot (and boy, it DOES get hot here in the tropics), but when it is cool you barely notice heat buildup. same type of workload, of course.

the 'ambient air' measurement is currently saying 40 celsius, but that has to be inside the laptop as it is ˜22.

cheers

That's fine, if your ambient t is 40C them you'll have corresponding enclosure/battery temperature increase which will cancel this variable.

Hi, I have the 2.7Ghz. model and have thermal issues, in the sense I feel the laptop get WAY to hot.

I am willing to help and do the test. However me being a noob... :D After you put all cores to work as described above, How you stop the process?

Hi, you can just close all windows one by one, or quit (cmd + Q) the Terminal.app

However, make sure you do the test for at least 10 min to let the temperature stabilise. Please note if your room is abnormally hot or cold, even though it will be translated into higher enclosure/battery temperature.

If you need help with iStat Menus settings, you can PM me.
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
So I will probably get a fresh 2.3GHz model to double check that difference. Surprisingly, I see a lot of threads and discussions about heat but yet no one is bothered to do a test so that we can have at least general view on this problem. C'mon people, do some tests and post the results! Everyone will benefit out of this!
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
Fans speed should be on the maximum if you loaded CPU up to 800%, maximum fans speed 6500rpm, your MacBook Pro has tons of dust inside, clean it up.

No, it's not. Those fans inside of RMBP max out at 5500/6000rpm respectively. They're different from those in cMBP
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
So if no one is interested, why are there so many "my mbp runs hot" threads?
 

maratus

macrumors 6502a
Original poster
Jun 12, 2009
701
273
Canada
So maybe there's at least one knowledgeable person on this forum who can confirm my statistics? Anyone?
 

Ploki

macrumors 601
Jan 21, 2008
4,308
1,558
I did CPU test http://www.coolbook.se/CPUTest.html and it does get smoking hot. 2.7ghz model.

103 degrees before fans start ramping

settles on 101 degrees when fans ramp up.

funny GPU die also heats up significantly when doing this test.

it doesnt seem any hotter than my old mbp, cooler even, and more quiet.
 

JronMasteR

macrumors 6502
May 4, 2011
327
126
Switzerland
I did CPU test http://www.coolbook.se/CPUTest.html and it does get smoking hot. 2.7ghz model.

103 degrees before fans start ramping

settles on 101 degrees when fans ramp up.

funny GPU die also heats up significantly when doing this test.

it doesnt seem any hotter than my old mbp, cooler even, and more quiet.

Exactly the same here. Ultimate Model, 100-103°C after 5 minutes. This CPU really gets very hot. Already thought that this was a bad machine, but it seems normal.

Edit: Forgot to say, mine is with stock thermal compound.
 

Schranke

macrumors 6502a
Apr 3, 2010
974
1,072
Copenhagen, Denmark
I did CPU test http://www.coolbook.se/CPUTest.html and it does get smoking hot. 2.7ghz model.

103 degrees before fans start ramping

settles on 101 degrees when fans ramp up.

funny GPU die also heats up significantly when doing this test.

it doesnt seem any hotter than my old mbp, cooler even, and more quiet.

What settings did you use for testing.
Have a 2.7GHz model myself and would like to throw in data as well
 

w00t951

macrumors 68000
Jan 6, 2009
1,834
53
Pittsburgh, PA
1) Unusual, because Intel chips usually throttle themselves around 98C.
2) The GPU can be explained by two factors: First, the heat sink for the CPU and GPU is connected - which defies logic. They're probably sharing some heat. Second, some tasks can be offloaded onto the GPU by OS X to take advantage of a video processor's superior parallel processing capacities.
 

Ploki

macrumors 601
Jan 21, 2008
4,308
1,558
What settings did you use for testing.
Have a 2.7GHz model myself and would like to throw in data as well
Big test, 1 repetitions, 8 instances.
enough to get it hot and for fans to ramp up.


1) Unusual, because Intel chips usually throttle themselves around 98C.
2) The GPU can be explained by two factors: First, the heat sink for the CPU and GPU is connected - which defies logic. They're probably sharing some heat. Second, some tasks can be offloaded onto the GPU by OS X to take advantage of a video processor's superior parallel processing capacities.
Makes sense. If CPU gets smoking hot before fans kick in the whole cooling system is already at 70 degrees, and depending where the sensor is some heat gets to the GPU sensor.
 

Schranke

macrumors 6502a
Apr 3, 2010
974
1,072
Copenhagen, Denmark
Just made the test with
Big test, 1 repetitions, 8 instances.
enough to get it hot and for fans to ramp up.

Did the test after a restart and reseting of the SMC
Had intel power gadget running to see power usage and CPU frequency.

Started from idle:
Temp were 48°C
CPU 1.20-1.30 GHz
Power usage around 3.60W

Here are the numbers i got out of it:
Peak: (2nd run)
107°C
3.80 GHz
and a power usage of 50W

1st run. (after the peak the 2nd run looked like this to)
temperature at 103°C
a steady 2.70 GHz
and a power usage of around 30W

The room is about 20°C.

(did a 2nd run 5 min after the first just to see how it went, the peaks are from that run.)
 

Ploki

macrumors 601
Jan 21, 2008
4,308
1,558
Just made the test with


Did the test after a restart and reseting of the SMC
Had intel power gadget running to see power usage and CPU frequency.

Started from idle:


Here are the numbers i got out of it:


The room is about 20°C.

(did a 2nd run 5 min after the first just to see how it went, the peaks are from that run.)

what do you use to see CPU clock?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.