Automated tool to reveal throttling and overheating - GitHub

[graemenail]

Aha! A neighbour
Yes you read that right.
They listened to what I said and I showed them my graphs. They ran their initial hardware tests and it passed. He then spoke to another Genius and he suggested running the extended cooling & performance test.
This test can take some time so I left it with them.
They phoned me the same day to say that it had failed the cooling test and they had already replaced the thermal paste twice and re-tested, which it failed.
As they weren't sure whether the logic board or the heatsink was causing the issue they were prepared to replace both.
There was a delay of a few days because they had to order the heatsink (which does not come with a new logic board - for some reason).

Since its return it has been great.

Good luck.

A neighbour indeed! I've booked it in for over the weekend, so hopefully they'll be able to sort it pretty quickly. Especially if they've ordered in spare heatsinks

Thanks!

 

[Quackers]

Good luck and let us know how it goes.

 

[graemenail]

Good luck and let us know how it goes.

First thing he did was clean out the fans. Then ran their hardware test and noted that the temperatures were high. I'd mentioned these comparisons and that you'd had an extending cooling test performed, and he agreed that's the obvious next step. Didn't even ask to see the comparisons. The high temps, after dust removal, potentially rules dust out. I'm curious as to what they'll come back with. Should be getting a call within 48 hours.

 

[Quackers]

Well that's the test that confirmed my problems.
I would have been tempted to show them your x264 graph. The max temp and the initial temp spike (immediate rise) might have given them a clue.
Anyway, I suspect their test will give them what's needed.
Best of luck and please keep us informed

 

[graemenail]

Well that's the test that confirmed my problems.
I would have been tempted to show them your x264 graph. The max temp and the initial temp spike (immediate rise) might have given them a clue.
Anyway, I suspect their test will give them what's needed.
Best of luck and please keep us informed

Just got a call, it failed the test. He basically described the process they did with yours: heat sink was the failure, going to try the thermal paste then retest and if there's still problems then it's bye bye logic board.

 

[Quackers]

So similar to mine then. Well at least they've diagnosed and repairs are proposed, so that's great.
Mine failed the re-paste twice which was when they decided to replace the logic board and the heatsink. Maybe yours will be the same. Whatever, as long as the repairs make the unit pass the test that's fine.
Another result for qnxor's little gizmo

 

[thundersteele]

Weird behavior

Here are some of my results. Late 2011 15'' MBP with 6750M GPU and 2675QM CPU. I ran the tests on a very warm day (30C), so I did expect some throttling.

Now the machine seems happy to run at 95C-100C with decent performance at 2.2 GHz, can't really expect more with these outside temps. However after a few minutes it decides that this is too warm, and it throttles down to around 2.0 GHz with a temperature around 90C (See first attached plot below).

It is not fully clear to me yet how this is triggered. The first test I ran had even worse throttling, down to 85C (second attachment). I did restart the system and reset the SMC and the situation improved. Maybe it is some other system component that heats up over time (the GPU would be my main candidate - maybe related to RadeonGate?)... hard to say. Or there is some other bottleneck?

Anyways, the reason I ran these tests is that I noticed a similar slowdown after a few minutes when gaming, and this is quite annoying. If anyone else has results of a 2011 MBP I would be very interested in seeing their results.

The remaining attachments are:
Prime95 before SMC reset
Prime95 after SMC reset
GPU test before SMC reset

 

[calviin]

Mhmm. I have a late 2013 15" rMBP with the 2.3 GHz i7 and the Nvidia GT 750M. Not as amazing as some other people's but no evidence of an unusual amount of throttling (if I'm reading these graphs correctly). Though, for some reason, the Prime95 didn't stop for me. It hit the end of the five minutes and then stopped.







Edit.
Thanks for making such a great tool qnxor! But can you make the file name use the time that it runs the test? As opposed to when the program was first open? If you run the same test over and over again, the graph will overwrite the previous graph. Just a suggestion. Thanks again!

 

[qnxor]

I see the MacOH tool may contribute to saving another laptop, @graemenail's. This is exciting. Let us know!

@calvin: good suggestion. Can you please open a github issue about it so I can keep track?

I'll try to catch up on the other posts soon (swamped with research at the moment). Keep the results coming

Bogdan.

 

[Barney63]

I have tried to remove Macoh from my rMBP, however there are 3 files in the trash in the macoh/tmp folder that cannot be deleted.

HandBrake-0.9.9-MacOSX.6_CLI_x86_64.dmg
Intel%AE Power Gadget.dmg
gle.dmg

How do I clear them from my trash?


Barney

 

[laurihoefs]

I have tried to remove Macoh from my rMBP, however there are 3 files in the trash in the macoh/tmp folder that cannot be deleted.

HandBrake-0.9.9-MacOSX.6_CLI_x86_64.dmg
Intel%AE Power Gadget.dmg
gle.dmg

How do I clear them from my trash?


Barney

Open Finder, and see if you still have them mounted. Are they visible in the Sidebar on the left? If so, press the Eject arrow next to each of them, and you should be able to empty your trash.

 

[Barney63]

No they aren't mounted in Finder, they aren't in Force Quit either.


Barney

 

[laurihoefs]

No they aren't mounted in Finder, they aren't in Force Quit either.


Barney

You can try changing permissions or taking ownership of the files.

Select one of the files, press Cmd + i, and see what the permissions on the bottom of the window look like. Click the lock in the lower right corner of the window to make any changes. You can either change permissions for everyone from 'Read only' to 'Read & Write', or you can click the + button under the permissions to add your own user to the list (if it's there already, no need to add it of course) and set is as the owner by selecting the user, clicking the gear button under the list, and selecting 'Make "(username)" the owner'.

Make either of these changes to all of the files, and you should be able to remove them.

 

[hexxed]

I just received a warranty replacement from Apple after my last machine (Late 2012 rMBP suffered logic board failure after a string of repairs) so after stumbling across this thread I thought id test mine.

Late 2013 rMBP 15 inch. 2.6ghz, 512Gb Storage, 16Gb Ram.

If I am reading the graphs right (which I may not be as I drove my maths teacher mental at school with my general ability at mathematics) I don't think I have anything to be worried about, but if anyone does have an eye for this and wouldn't mind eyeballing my results that would be great.

MacOH 1.3.1-beta. Ambient temperature: 23 degrees centigrade.

 

[qnxor]

@graemenail: I'm glad to see that you're on your way to fixing the problem. Let me know if Apple sorts it.

@calviin: not much to worry about in your graphs. Slight throttling compared to others (I wouldn't lose sleep over it though).

@Barney63: Try from the terminal using sudo (sudo rm -rf ~/macoh/tmp/*). It will ask for your password. If that doesn't work, then they are mounted. The Intel one was mounted as root (using sudo), a reboot should then allow you to remove them if not. However, if the files are actually in "Trash" then all you need is to empty your Trash.

@hexxed: The x264 test looks good enough. The GpuTest one, however, throttles more than I'd expect. The CPU is not pushed in that test, yet it throttles down to 2.6 GHz, almost surely because of the GPU overheating (they sit under the same heatpipe and influence each other's temperature, see my older posts). If you play 3D games which push both the CPU and the GPU, you may see even worse throttling (you can start Intel Power Gadget from Apps and then the game), but it's not guaranteed.

@thundersteele: Doesn't look too nice, but I have little experience with the 2nd gen Core i7-2xxx Macbook Pros. I can't say whether that's expected or not (the heatsink designed by Apple for that generation may just not perform). It would be good if others with the same laptop and CPU ran the same tests. The throttling in GpuTest is most likely due to the same reasons as @hexxed's ... common problem. If you play games, you can try the same thing I suggested above.

Bogdan.

 

[hexxed]

Thanks qnxor, I do game. I assume the nVidia chip will be used the majority of the time so should I be worried?

I also notice that mamarrash on page 4 of this thread has the same spec as me, and has almost identical results. I wonder if the 2.6 chip limits itself when using the Iris Pro.

https://forums.macrumors.com/attachment.php?attachmentid=474482&d=1401685585

 

[mamarrash]

That's correct hexxed. Almost identical results. We should see more integrated GPU tests in the rMBP BTO. When using the NVidia GPU, the CPU does reach 3.6 frequency. Perhaps is normal behavior.

 

[Barney63]

Here are my results on a late 2013 15"/2.3/16/512/gt750m

How do they look?
Is there anything out of spec?


Barney

 

[qnxor]

@hexxed: My apologies, I didn't pay attention to the graph title and I thought it was using the Nvidia GPU. Yes, it is expected for the CPU to throttle more when using the integrated GPU (Iris Pro). That's mostly because the chip max TDP is hit rather quickly. The 2.6 GHz will also throttle a bit more than the 2.3 GHz (it heats more), yet it may maintain higher clocks.

That being said, the amount of throttling in the GpuTest test with the Iris Pro is still a bit much (compare to my 2.3 GHz for instance). Try running the GpuTest test with the Nvidia GPU, and have it run for longer than 10 minutes. See if it throttles then.

@Barney63: Not much to be worried about. It throttles a bit more than mine in x264 (see my 1st post) but I wouldn't lose sleep over it.

 

[mamarrash]

@hexxed: My apologies, I didn't pay attention to the graph title and I thought it was using the Nvidia GPU. Yes, it is expected for the CPU to throttle more when using the integrated GPU (Iris Pro). That's mostly because the chip max TDP is hit rather quickly. The 2.6 GHz will also throttle a bit more than the 2.3 GHz (it heats more), yet it may maintain higher clocks.

That being said, the amount of throttling in the GpuTest test with the Iris Pro is still a bit much (compare to my 2.3 GHz for instance). Try running the GpuTest test with the Nvidia GPU, and have it run for longer than 10 minutes. See if it throttles then.

here´s the test with the Nvidia Gpu. It lasted 15 minutes. No throttling.
And then Iris Pro test. 20 minutes. Stays at 2.8 Ghz. Perhaps normal behavior?. Some other users with a rMBP 2.6 BTO post your GPU tests!

 

[qnxor]

here´s the test with the Nvidia Gpu. It lasted 15 minutes. No throttling.
And then Iris Pro test. 20 minutes. Stays at 2.8 Ghz. Perhaps normal behavior?. Some other users with a rMBP 2.6 BTO post your GPU tests!

The Nvidia one looks great, you're good to go. It's unfortunate that the Iris Pro causes throttling but the 2.6 GHz rMBP does overheat a tad more than the 2.3 GHz. 2.8 GHz is still no slouch so I wouldn't lose any sleep over it, especially since you'll be using the Nvidia GPU for 3D anyway. I am, however, curious as well to see more results for the 2.6 GHz. Maybe hexxed can run the same GpuTest test with Nvidia.

 

[mamarrash]

Yes. It would be interesting to see more GPU tests results from other users running the same specs. It does pose some questions about clocking frequency vs real performance.

 

[SVTmaniac]

What I've gotten from all this so far is that the 2.6 may not really be worth the price increase. When it matters it runs at about the same speed as a 2.3 does.

 

[hexxed]

@SVTmaniac I would completely agree. Thankfully I didn't pay a penny for my upgrade otherwise I would be a bit frustrated!

I have attached my 10 minute nVidia results. No throttling

 

[calviin]

@calvin: good suggestion. Can you please open a github issue about it so I can keep track?

Actually, I don't know how to use Github. Programming isn't exactly my forte. Thanks for the analysis and yea. I agree there's a slight bit of throttling and I can probably remedy that by reapply thermal paste but TBH, most people won't notice throttling unless they're actually using it as a pro motion; meaning 1 GB large files in PS, FCP, Premiere, etc. I feel like this is a WebMD for Macs. But works well for troubleshooting and proving to Apple wether or not a computer is underperforming.