Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
So I have a refurb that has this problem and the GPU panics. Would Apple bite if I propose them to return this one if I buy a similarly spec'ed machine new with a 5600? The GPU panics appeared after my 14 day return period but I've actively pursued them ever since.
 
Keep it or take it back? That's a tough decision.

Based on history....

These all had GPU problems:

The MacBook Pro (15-inch, Early 2011)
MacBook Pro (15-inch, Late 2011)
MacBook Pro (17-inch, Early 2011)
MacBook Pro (17-inch, Late 2011)
MacBook Pro (Retina, 15-inch, Mid 2012)
MacBook Pro (Retina, 15 inch, Early 2013)

A class action lawsuit in 2014 resulted in Apple offering to fix "a small percentage of MacBook Pro systems"

So if it is indeed a defect, that may go into 2024 before (if) it gets "fixed"

Now with the engineering quotes of "performing as expected", that sounds like legalese for "it works, quit griping". They're probably betting it will last 3 years or more, exceeding any warranty or applecare period.

My 16" MBP is a 2.4GHz I9 with a 5500M GPU. I use it regularly with the lid open and have a Dell U2717 connected to it using a Cables To Go USB-C to displayport cable. The temperatures are consistently 63-64C range, and the GPU consistently draws 19W, based on the last 30 days of istat data. Fans typically at 24-2500RPM.

I ended up installing Turbo Boost Switcher just to keep it cooler under load. I'll usually have 2 VM's running plus chrome/firefox/safari and Outlook. If I remember correctly, that cut temperatures about 10-12C. With Turbo off, the CPU is enough to do what I need most of the time. With no external monitor, temps drop down to the 43-45C range.

Is the 5600M a fix? I don't know. I personally don't think its worth another $700 to "fix" something on what is supposed to be a premium laptop.

Your choices are basically to return it, to live with it, or upgrade and hope the 5600 fixes it. I'd make sure to have apple care on it for sure, regardless.
 
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
So I have a refurb that has this problem and the GPU panics. Would Apple bite if I propose them to return this one if I buy a similarly spec'ed machine new with a 5600? The GPU panics appeared after my 14 day return period but I've actively pursued them ever since.
Let us know if it works even though I don't like this idea because ultimately apple is making more money instead of getting their **** together.
 
So I have a refurb that has this problem and the GPU panics. Would Apple bite if I propose them to return this one if I buy a similarly spec'ed machine new with a 5600? The GPU panics appeared after my 14 day return period but I've actively pursued them ever since.

It's worth a shot. Worst case you'll get your repair ticket started.
 
I suppose we’ll have to wait for ARM to fix this for us, or spend an extra 800 dollars on a dGPU most of us will never need :(.

This seems to be what’s happening here:




So, this is probably how it’s suppose to work by design, although some people seem to have it fixed:

i had this problem with a R9 290X and 3 monitors attached since the crimson drivers, they tried to tell me this was normal.



BUT, LO and BEHOLD, the new crimson driver 17.4.1 suddenly manages to bring down the memclockspeeds at idle back to 150 MHz instead of the hot 1250 it had been running all the while.

So the excuse of it being normal with multiple monitor setup has been debunk for good.



Anyway, thanks amd for fixing my idle hot running. Now to keep it this way plz.

Also, This would explain why some people are not having issues when selecting standard refresh rates:

this is the expected behaviour, it is a workaround to prevent graphical corruption/screen flickering,

From my readings ,essentially changes in memory clock must occur during the Vblank period when the monitor is not displaying anything, currently only 60/120hz have the proper sync for downclocking to work, this is sort of fixed in newer drivers for other refresh rates, but many monitors do not use the a standard vblank time for frequencies like 75hz, 144hz, 90hz (often reduced vblank, or other), and so it is simply forced to the maximum clock to avoid desync (which would cause flickering) . there is more information about the issue on linux here https://bugs.freedesktop.org/show_bug.cgi?id=102646


Also:


Just Google “amd maximum clock 2 monitors” and you’ll see tons of reports.




If you have dual or multi monitor, you need to run the exact same refresh rate AND color bit AND black levels/rgb type to have 300mhz mem.

If you dont do this, mem will stay at highest clock.

This been happening forever on amd cards.

I have one monitor supporting 12bit color and the other 8 bit. If i dont set both to 8, it wont stay at 300mhz.

This idle clock issue is a very old one with multimonitor. In order to keep the clocks low, the monitors need to match settings: refresh rate, connection type, etc. If you connect 2 monitors that are identical, but connect one via DP and another HDMI, it still won't drop. I was running 3 Dell 1080p monitors off my 390, and I had to invest in getting DP to HDMI active adapters so that I can have my idle clocks. (These monitors predate DisplayPort)

It seems that some people have it fixed by either changing some power savings options for the GPU or installing different/updated drivers.
 
Last edited:
I understand that it is expected from their perspective as they know that the design is faulty and this is how it works, but it doesn't change the fact that it isn't reasonable behaviour from a pro/top/best notebook or acceptable in any way.

It's just the usual blah-blah. In really only a few cases Apple said, hey yeah, it was stupid of us. With the release of the 16", the keyboard is exchanged quietly, everybody knows why, but only Apple doesn't say anything about it. With the Antennagate of the iPhone 4, mobile phones from other manufacturers that had at least as bad signals are shown at the keynote for minutes. At first, there is a long silence and suddenly (sometimes) there are longer repair/warranty periods for some products. It did all never damage to Apples reputation. I am also annoyed but I shop there diligently. In the Apple universe there is only Apple.

What bothers me so much about this desperate discussion is, why have I been buying macs for 20 years? So that I don't have to think about this fan nonsense, underclocking or overclocking and whatever. And now I have to start this ******** with my Macbook?
 
Last edited:
  • Like
Reactions: bomby0
I have the base model and the SwitchResX doesn't do anything with lid open and 2K external screen.
Radeon is still pushing 18w+. CPU seems normal depending workflow but it idles at around 5-6w.

I'm well aware the SwitchResX "fix" is for clamshell mode only.. I'm reminded everyday I turn on my machine.
 
It's just the usual blah-blah. In really only a few cases Apple said, hey yeah, it was stupid of us. With the release of the 16", the keyboard is exchanged quietly, everybody knows why, but only Apple doesn't say anything about it. With the Antennagate of the iPhone 4, mobile phones from other manufacturers that had at least as bad signals are shown at the keynote for minutes. At first, there is a long silence and suddenly (sometimes) there are longer repair/warranty periods for some products. It did all never damage to Apples reputation. I am also annoyed but I shop there diligently. In the Apple universe there is only Apple.

What bothers me so much about this desperate discussion is, why have I been buying macs for 20 years? So that I don't have to think about this fan nonsense, underclocking or overclocking and whatever. And now I have to start this ******** with my Macbook?

Because nothing is as good as Apple, besides microsoft.
Dell is crap, their reddit page is filled with 2020 dell quality issues, CS nightmare stories.
[automerge]1593465149[/automerge]
I suppose we’ll have to wait for ARM to fix this for us, or spend an extra 800 dollars on a dGPU most of us will never need :(.

This seems to be what’s happening here:




So, this is probably how it’s suppose to work by design, although some people seem to have it fixed:



Also, This would explain why some people are not having issues when selecting standard refresh rates:




Also:


Just Google “amd maximum clock 2 monitors” and you’ll see tons of reports.

We need to pressure APPLE to update our drivers, or address this!
 
I don't think the clock is sticking - I'm able to get the clock to drop in clamshell if i have the right settings.

I reported earlier that running different rotations with dual displays seems to trick the power usage to drop (one normal, one at 90 degrees). I thought I'd try something... if I run 2 x 4K displays both in normal rotation I'm around 21W. If I flip one monitor upside down then it drops to 7w. Given I've VESA mounted the monitors I can run one upside down by physically inverting the monitor.. Going to give this a try for a while.

A ridiculous solution to the problem... but (hopefully a temporary) workaround.

Screen Shot 2020-06-30 at 8.25.37 am.png


Also for those of you trying to run a single monitor not in clamshell.. does tweaking the internal display frequency help:
Change the refresh rate on your 16-inch MacBook Pro or ...support.apple.com › en-us

Didn't seem to help me .. but could be worth a try???
 
  • Like
Reactions: IgorK
If that is the case then the 5300 and 5500 are a design flaw because I am unable to work with a machine that heats up to 100 Celsius on basically any "professional" task which then slows the performance in order no to melt itself.
Have you tried touching the keyboard top cover after the cpu runs at this temps for a bit?
I totally agree with you on this point. I have touched it, and it is very hot (not just "warm"). While I can force myself to leave my fingers there without burning them, I would certainly not let my kids touch it.

I would urge anyone who is dissatisfied to return it if they can, and if they are outside their return window to ask for a full refund so they can purchase a different model (e.g. 13" MBP, or 16" MBP with 5600m, or Apple Silicon MB/MBP), because:

- It is a design flaw and not a manufacturing defect or software bug. But Apple support will always classify this issue as either a manufacturing defect or software bug.
- It will not be fixed no matter how long we wait, and Apple will never provide an ETA for a software fix even if they could.
- It will not be fixed no matter how many times Apple attempts to repair (replace logic board) or replace the machine (with a brand new 16" MBP with 5300m or 5500m graphics).

Be prepared for a long fight, where you will have to politely jump through every hoop they throw your way and then escalate the issue by email to tcook@apple.com when they eventually say "works as designed, no refund".

What I've been told by Apple support is that if you want a replacement (same model/line), they can do that if there is a "major fault" and at that time you can choose to upgrade to 5600m. Senior advisors can approve this. But if you want a different model/line, you must request a refund and then purchase the replacement yourself, which requires a whole other chain of approvals. The bar is already very high for a replacement, and even higher still for a refund.

Apple offered to replace my 2019 15" MBP with a new 16" MBP, but I wanted a 13" or iMac so I asked for a refund. That extended the process by another 5 weeks (to 9 weeks total) and took up 40+ hours of my time running tests and on the phone every 2-3 days waiting on engineering response and various approvals and multiple attempts to repair.

At the end of it all, they denied my request AND WITHDREW their earlier offer of a replacement. I emailed tcook@apple.com explaining the whole chain of events in detail and expressing my deep frustration with the support process.

After another week waiting they agreed to replace the machine, while explicitly noting that they "don't normally only replace for a major fault, and we do not consider this issue to be a major fault, but this is not the support experience we want our customers to have".

If enough people complain, they may be more inclined to agree to "keep you happy" without admitting any fault or defect. Or we may see an extended warranty program (similar to the butterfly keyboards) in 3 years time, which doesn't help anyone who actually wants to use the machine now.
[automerge]1593473856[/automerge]
Once travel opens up again i'll be using the machine more in laptop mode than connected to a display. I know that i shouldn't have too do this but keeping it in clamshell mode when at my desk isn't a total deal breaker. Im able to keep the temp around 50 to 55 Celsius when in clamshell. My concern is the longevity of the machine and if these temps are acceptable long term.

I do agree that anyone who uses their machine at a desk shouldn't even remotely have to be forced to keep it in clamshell mode to get a desired temperature and lower fan noise.

Also Second update:

Spoke with a senior engineer and he essentially said that again this is expected behavior. 🙃 Can't make this up lol
If you prefer clamshell mode, this is an amazing machine. If you can afford the 5600m model, this is an amazing machine. If you can afford an eGPU, it's a pretty great machine. If you can't afford the 5600m model or an eGPU and you want to use it with the lid open and an external display (even if you can "live with" it in clamshell mode), return it. Get a 13" instead. It'll still be plenty fast, just not quite as fast. It'll still have a great internal display, just not as big. Or wait for Apple Silicon MBP, if you can.

For myself, I'm keeping the machine because I already spent 9 weeks and 40+ hours on the phone battling Apple and I was not given any option to switch to a different model (13" or iMac). So I bought an eGPU (before 5600m was an option). The eGPU quirks are annoying (quit apps before unplug), but the machine is incredibly fast and basically perfect in every other way.
 
Last edited:
From my understanding, you’re not satisfied with the thermals and cooling of the machine, is that correct?

I certainly understand the frustration if that’s the case however I don’t see escalation getting you the exchange or refund you want.

If you do end up combining, try to be civil and understanding to whoever you’re speaking to. They’re a person too. I used to work for Apple retail and they do try to lean on the side of the consumer. However, if you’re going to act ignorant and demanding then they’ll just tell you “tough cookies”.

Just my two cents. :)

Just for the record, I am always polite and civil to whoever I talk to, including on the calls I’ve had so far with Apple support on this issue. Being ignorant and insulting is always counter productive, in my opinion, and just isn’t me anyway. So, I agree with you in this respect.

Nonetheless I will continue to complain, escalate and explain that this “expected behaviour“ position is not on. This isn’t ignorant or demanding, it’s protecting my consumer rights. And I’m a big Apple fan, advocate and spender, have had multiple Macs over the years, have always used iPad and iPhone (as do my entire family), subscribe to Apple Music family plan, iCloud storage 2TB, have multiple HomePods and Apple TV’s around the house etc. I’m firmly on Apple’s “side” in the grand scheme of things, that doesn’t mean I shouldn’t escalate something I find entirely unacceptable. But I’ll do it with politeness and respect as always.
 
Well I guess it does work like a normal computer when it is in clamshell mode.
Only when rebooted or started from cold with the lid closed.
Even with turbo boost switcher on, there's not much I would complain about if this thing worked like this in dual screen.
Now here's the thing where I believe this is a software bug. iGPU on it's own draws about 0.8w on idle for me while the 5300 draws about 5w. There's no way that when they run both all of a sudden 5300 needs all that extra power. Not buying it. It's a driver issue same as gpu crashing for random people when it goes to sleep while connected to external screen.

Glad I'm getting an ultrawide but still won't stop pushing apple for real solution.
For now this will do.

**forgot to add that this is using a usb-c ---> display port cable.


clamshell.png
 
Last edited:
Now here's the thing where I believe this is a software bug. iGPU on it's own draws about 0.8w on idle for me while the 5300 draws about 5w. There's no way that when they run both all of a sudden 5300 needs all that extra power. Not buying it. It's a driver issue same as gpu crashing for random people when it goes to sleep while connected to external screen.
I don't think the iGPU has anything to do with it. The problem does not occur because the iGPU (internal) and dGPU (external) are active. The problem occurs when the dGPU drives two displays (BOTH internal + external) at different resolutions and/or refresh rates. The dGPU memory then runs at full speed, using more power and generating more heat. That behaviour is reportedly by design to avoid glitches because memory clock changes can only happen during v-blank, which can't be synchronised with multiple displays in many cases in clamshell, and in all cases with lid open.

Has anyone tried adjusting the refresh rate of the internal display, and then adjust the external to match each of those options? See: https://support.apple.com/en-us/HT210742
 
The problem occurs when the dGPU drives two displays (BOTH internal + external) at different resolutions and/or refresh rates.

Not quite right. See my post above. I've had high power usage with two of exactly the same monitor with exact same connection type. Setting the rotation on one display to something other than Standard causes the power usage to drop?!

Now been running all morning with one monitor physically upside down (and set to 180 deg. rotation). Machine is noticeably cooler and less fan noise.

Has anyone tried adjusting the refresh rate of the internal display, and then adjust the external to match each of those options? See: https://support.apple.com/en-us/HT210742

Yes - as per my post above - didn't help - running internal and one external display at exactly 60Hz
 
...if I run 2 x 4K displays both in normal rotation I'm around 21W. If I flip one monitor upside down then it drops to 7w. Given I've VESA mounted the monitors I can run one upside down by physically inverting the monitor.. Going to give this a try for a while.

That is absolutely absurd and hilarious, but fascinating (and potentially a useful workaround).

I have a 16" on its way, and I'm a bit nervous about this heat issue, but I'll definitely be adding the monitor-flip to the troubleshooting list :)
 
  • Like
Reactions: sco
welp, I can't believe it, but I organized to return the 13" and I'm going to give this 5600M a shot. I reeeeeeally hope this works like the other poster said it does. I like the 13" - a lot - but I know that if I don't give the 5600M a shot it will bother me every single time I'm rendering, or compiling, or editing, or or or. I wish there was more data than just the one poster in this thread though :(

Really looking forward to running this 16 inch without any hacks at a cool 8-12W.
 
So this morning, MBP on for about 20 minutes and this is the current state of my machine. Temp over 90c and fans running at 5,599 and 5,205. No video running, just MS Office apps, Teams, Chrome, Safari and App Store. Expected behaviour apparently 🤨

IMG_2008.jpeg
 
My report (MBP 16", i9 + 5500m):

- Clamshell, always.

  1. 1x 1440p@120hz connected via USB-C to DP cable:

    Stays around 5-8w, stays quiet with light workload, gets loud when I start to actually do stuff (VSCode, Docker Desktop, Chrome, Safari, Teams):

    Screenshot 2020-06-30 at 14.07.31.png

    stats shown with only Safari and a few system services running

  2. 1x 1440p@120hz display connected via USB-C to DP cable.
    1x 1440p@60hz display connected via USB-C to HDMI dongle (not Apple).

    Always loud, always hot, even with no apps running:

    Screenshot 2020-06-30 at 14.14.11.png

    stats shown with only Safari and a few system services running
 
Last edited:
My report (MBP 16", i9 + 5500m):

- Clamshell, always.

  1. 1x 1440p@120hz connected via USB-C to DP cable:

    Stays around 5-8w, stays quiet with light workload, gets loud when I start to actually do stuff (VSCode, Docker Desktop, Chrome, Safari, Teams).

  2. 1x 1440p@120hz display connected via USB-C to DP cable.
    1x 1440p@60hz display connected via USB-C to HDMI dongle (not Apple).

    Always loud, always hot, with no apps running.
Loud? What’s your temps and rpms in both options?
 
I don't think the clock is sticking - I'm able to get the clock to drop in clamshell if i have the right settings.

I reported earlier that running different rotations with dual displays seems to trick the power usage to drop (one normal, one at 90 degrees). I thought I'd try something... if I run 2 x 4K displays both in normal rotation I'm around 21W. If I flip one monitor upside down then it drops to 7w. Given I've VESA mounted the monitors I can run one upside down by physically inverting the monitor.. Going to give this a try for a while.

A ridiculous solution to the problem... but (hopefully a temporary) workaround.
I love your solution :D It must be a software bug. I think I've heard somewhere that Apple has delegated GPU Bootcamp driver development to AMD. Maybe it's the same for macOS? From what I know AMD has not released any driver updates since 16"s release and their track record with drivers is.. well.. unsatisfactory. Could this explain the situation?
 
Got a reply from the Apple rep after he spoke with the "engineers" and apparently there is nothing wrong with the computer and it is working as intended.

I bought my mbp 16" in December and have used it without a external screen since then.
Last week I bought a BenQ PD3220U, plugged it in, and mere seconds later my joy turned to horror.

I work a lot with audio and its impossible to work with this setup.

I've been in contact with Apple in Sweden and the senior rep I spoke to last week recorded some logs to send to the "engineers".
Today I read up on this thread and heard about the updated 16" with the 5600M. I called them and asked if a trade (with money in between) would be possible. Didn't get much of an answer except they have to wait for a reply from the "engineers" but I doubt it will be.

I also asked if the rep knew about this issue and he was reluctant to answer but said that he could see a few similar cases.

He said this would likely be fixed in an upcoming update.
But from a customer point of view this is a disappointing answer. Who knows how long that will take. I need to use this setup for work now.

In istat pro I see the high GPU drawing 19w without load. If I clamshell it goes down but its still hot and noisy.
I tried turning off the turbo boost and it helped a bit but not enough for me to have it turned off.
Simple browsing and YouTube puts the temps up to 80C and the fans runs at almost max. Keyboard also gets very hot to the touch.

I researched this computer very intensively before buying it and I haven't seen a single review or article mentioning this issue. But if you put "MacBook Pro 16" heat" in google there are so many hits.
Very odd.
 
So I currently don't have a personal computer (just the work laptop - on which I really don't want to do personal things). I originally wanted to grab a 16" for the performance.... but, as I don't really need advanced graphics and hearing these nightmares with the dGPU, I'm seriously reconsidering getting a 13".... or even worse, waiting for the ARM versions... though I really don't want to have to do without a personal computer for... probably a year.

Do the 13" MBPs function properly? I'm looking at grabbing the i7, 32GB, 1TB version. This'll be used mostly for iOS development, medium heavy music production, and the standard netflix/email/webbrowsing that a toaster can handle. I would be using it with an external monitor (DELL U2715H) plus the internal monitor.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.