2nd MBP nTB with display issues--what to do?

[ghanwani]

I got my first one at the end of November. It had display issues. I ordered an identical one, and when it arrived, I went into the store and did a return of the old one and picked up the new one. The new seemed OK for a day (it was running 10.12.1) and then last night I updated to 10.12.2 and I found the same problem happen again this morning.

What do I do? Should I just order another and hope that one won't have the problem? How many such returns will Apple let me do? Anyone here have this problem with multiple machines? What are you doing about it?

This is the problem I'm talking about. It happens after the machine has been sleeping for several hours and is opened. It resolves itself if I shut the lid and reopen it.
https://dl.dropboxusercontent.com/u/61646136/IMG_1366.MOV

(With the last swap my credit card got triggered for fraud alert repeatedly and that was a mess. I ended up having to use a different credit card.)

 

[OS X Dude]

Take it to Apple for service. Chances are it'll be getting a new logic board but I can't say without diagnosing it in person

 

[skiffx]

screw new logic board get a brand new system AGAIN and let them eat the cost of 2 returned machines.

 

[kb3uru]

Im having similar issues with my ntb when hooked up to an external monitor. It either won't wake up, or I get a snowy variation of what you have, on both the external monitor and the MacBook display. Contacted apple a week ago, been given the run around, no response from engineering and I'm out of the return period from whom I bought it from..

 

[xraydoc]

Pretty sure this is a software bug.

 

[ghanwani]

If it's software, I'm surprised there aren't more folks reporting it and also that it is present in both 10.12.1 and 10.12.2.

If it's hardware, I'm surprised I have had the problem with both machines.

Perhaps I'm just jinxed.

 

[Balaamsdonkey]

Take it to Apple for service. Chances are it'll be getting a new logic board but I can't say without diagnosing it in person

They would capture it and give op a new one before replacing a board. They like capturing faulty devices to find what the problem(s) are.

 

[coolbreeze]

Just turn on your lights if it acts up again, problem solved.

(sorry, couldn't resist)

Yeah, replacement time.
[doublepost=1481850373][/doublepost]

Pretty sure this is a software bug.

We are all running the same software, so it should happen to everyone, right? Genuinely curious.

 

[xraydoc]

Just turn on your lights if it acts up again, problem solved.

(sorry, couldn't resist)

Yeah, replacement time.
[doublepost=1481850373][/doublepost]
We are all running the same software, so it should happen to everyone, right? Genuinely curious.

The 10.2.2 update was in part to correct some graphical glitches. And we didn't all experience it. I believe it to be partly related to Adobe CC software and other software that installs color profiles.

 

[RCElectricFlyer]

I have this exact same problem on recovery from a long sleep. 10.12.2 does not fix it. It has nothing to do with Adobe software.

 

[thekev]

The 10.2.2 update was in part to correct some graphical glitches. And we didn't all experience it. I believe it to be partly related to Adobe CC software and other software that installs color profiles.

What's your reasoning for that? Adobe doesn't install anything that writes to your display's framebuffer. Internally it's using ACE. Your display is getting its info through colorsync, which like other cmms likely clamps anything that falls out of bounds in its profile connection space prior to mapping that into the destination space. If a bad profile tag or something similar causes this kind of glitch, that's an Apple problem, not an Adobe problem.

 

[xraydoc]

What's your reasoning for that? Adobe doesn't install anything that writes to your display's framebuffer. Internally it's using ACE. Your display is getting its info through colorsync, which like other cmms likely clamps anything that falls out of bounds in its profile connection space prior to mapping that into the destination space. If a bad profile tag or something similar causes this kind of glitch, that's an Apple problem, not an Adobe problem.

My reasoning is that most people with glitches have Adobe CC installed. And I'm not blaming Adobe per se; it's Apple's bug but Adobe software seems to be a common (though not only) thread.

 

[RCElectricFlyer]

My reasoning is that most people with glitches have Adobe CC installed. And I'm not blaming Adobe per se; it's Apple's bug but Adobe software seems to be a common (though not only) thread.

For this particular problem, Adobe software is not relevant. I have none installed.

 

[thekev]

My reasoning is that most people with glitches have Adobe CC installed. And I'm not blaming Adobe per se; it's Apple's bug but Adobe software seems to be a common (though not only) thread.

It may not be exclusive to them. My concern is that since these problems began, people seemed to look to third parties, in part due to bad reporting on the subject. Apple significantly restricts low level access, of the type that can actually bring down your machine like this, so they'll have to be the ones to fix it. For the most part third parties can tell you if their service works as intended on the current iteration of OSX, but they can't typically say more than that.

You also mentioned the installation of color profiles. It would be really weird if they caused such a problem. You could of course test this well enough by removing certain ones. I would have to check whether Adobe reinstalls them at startup or simply retrieves a list of which ones are installed. Profile tags shouldn't cause a problem as I mentioned. They're mostly there to provide contextual information. Apple mentions updates to profile tags somewhere on this page. Just search for "New optional profile tags" if you're interested in changes. There isn't a lot of room for that kind of thing in terms of remapping, because they're basically normalized twice, once at the profile connection space and again at the destination space. It's because earlier ICC specs didn't provide for handling of signed xyz values.

I guess you could still be right, but it would be a very strange place for this kind of problem to pop up.

 

[xraydoc]

It may not be exclusive to them. My concern is that since these problems began, people seemed to look to third parties, in part due to bad reporting on the subject. Apple significantly restricts low level access, of the type that can actually bring down your machine like this, so they'll have to be the ones to fix it. For the most part third parties can tell you if their service works as intended on the current iteration of OSX, but they can't typically say more than that.

You also mentioned the installation of color profiles. It would be really weird if they caused such a problem. You could of course test this well enough by removing certain ones. I would have to check whether Adobe reinstalls them at startup or simply retrieves a list of which ones are installed. Profile tags shouldn't cause a problem as I mentioned. They're mostly there to provide contextual information. Apple mentions updates to profile tags somewhere on this page. Just search for "New optional profile tags" if you're interested in changes. There isn't a lot of room for that kind of thing in terms of remapping, because they're basically normalized twice, once at the profile connection space and again at the destination space. It's because earlier ICC specs didn't provide for handling of signed xyz values.

I guess you could still be right, but it would be a very strange place for this kind of problem to pop up.

To be honest, I don't really know much about how color space management/ColorSync works on Mac OS. I don't use Photoshop/Lightroom/Creative Cloud, etc.

But what I have seen is that owners of every model of new 2016 MBP have seen these types of graphical glitches. Thing is, there's two different processors available for the nTB model, three for the 13" TB model, three for the 15" model plus three different dGPUs for the 15" model. Issues involve chips using Intel's 540 iGPU, 550 iGPU, 530 iGPU and all three versions of the AMD Radeon (the 450, 455 and 460).

That's, what, 11 different combinations of CPU+GPU combined. And graphics issues have been reported on all of them. I don't think it's an Intel Skylake issue - the processor has been out for ~1 year already and no such issues have been reported on PCs. It's not an AMD Radeon issue since the 13" machines are reported to also be affected. Could be a custom IC that Apple is using on all their 2016 machines I suppose, but why isn't every machine affected then?

There have been reports of errors starting moments after installing certain Adobe software. Not saying Adobe software is the cause, but something it does is triggering the issue. Not everyone affected is using Adobe software, but I wonder if there's not some common software function that's triggering the bug.

Apple has said publicly that the 10.12.2 update was supposed to fix those issues. Seems to have done so for many, but apparently not all. Perhaps another point update is needed to fix a more rare situation.

I don't use any Adobe software other than Adobe Reader and on two different machines (a non-TB 13" and a 15" 2.7/455) I've not experienced the graphics issues described on these threads.

Again, not suggesting people's problems aren't real. But evidence strongly suggests a software problem versus a hardware problem.

 

[thekev]

Again, not suggesting people's problems aren't real. But evidence strongly suggests a software problem versus a hardware problem.

I actually agree with you, and I suspect it's low level software. Adobe and certain others may use low level APIs which expose this particular bug. Other people mentioned some backup issue with external drives. My broader is that stable implementations of draw protocols and opaque APIs would make it nearly impossible for a third party to even produce something like this without explicitly drawing it in full screen mode (which sounds like a very mean prank).

I will say this, I feel sorry for whoever has to track that down that particular issue.

 

[ghanwani]

I have this exact same problem on recovery from a long sleep. 10.12.2 does not fix it. It has nothing to do with Adobe software.

Thanks -- you're the first one that has validated this specific problem. What are you doing or planning to do about it?

 

[RCElectricFlyer]

Two questions for everyone experiencing this problem:

1. Does the problem completely disappear after log in?

For me it does, except for artifacts that may sometimes remain on the screen temporarily until the screen is refreshed.

2. Does everyone with this particular problem have a non-touchbar MBP?

I do.

If this is the nature of the problem for everyone, then it is a relatively minor problem as it affects a limited number of users and doesn't affect normal use. As such, it would receive lower priority for a fix.

It sure looks like a driver issue to me. I hope I'm right.
[doublepost=1481902899][/doublepost]

Thanks -- you're the first one that has validated this specific problem. What are you doing or planning to do about it?

In the short term, research the issue and hope for a software fix

In the longer term, talk to Apple

I really don't want to jump the gun as I have a lot invested in this laptop now - i.e. BootCamp.

 

[ghanwani]

Two questions for everyone experiencing this problem:

1. Does the problem completely disappear after log in?

For me it does, except for artifacts that may sometimes remain on the screen temporarily until the screen is refreshed.

Sometimes it does, and sometimes it doesn't. The more common case is where I have to wait until the screen blanks out due to inactivity, or force it to blank out again by shutting the lid. When it wakes up again, the problem is gone.

 

[RCElectricFlyer]

Sometimes it does, and sometimes it doesn't. The more common case is where I have to wait until the screen blanks out due to inactivity, or force it to blank out again by shutting the lid. When it wakes up again, the problem is gone.

Interesting. That is slightly different than my experience.

What is your system config? I have the base processor, 16GB and 512GB.

 

[ghanwani]

Interesting. That is slightly different than my experience.

What is your system config? I have the base processor, 16GB and 512GB.

Same!

 

[RCElectricFlyer]

Same!

Hmm. Is the problem unique to this configuration? The plot thickens...
[doublepost=1481903939][/doublepost]

Hmm. Is the problem unique to this configuration? The plot thickens...

Wild guess. Does this mean video memory management could be the issue (the on board accelerator uses RAM doesn't it?)?

 

[ghanwani]

Found another one here, but there is no information about their configuration.

They are seeing the same things as in this thread -- happens with multiple machines and doesn't get addressed by 10.2.2.

 

[MrPepe]

Hmm. Is the problem unique to this configuration? The plot thickens...

I'm also having the same issue as are others from different threads. I have until Jan 31, 2017 to return or exchange (purchased from Newegg), so I'm hoping this gets resolved before then.


PS - 13" non touchbar, base CPU, 8GB, 256GB

 

[RCElectricFlyer]

Found another one here, but there is no information about their configuration.

They are seeing the same things as in this thread -- happens with multiple machines and doesn't get addressed by 10.2.2.

Why is it blinking - and why at this rate? Mine looks like this too. This is another thing that points to software for me. I don't think it would manifest like this if it was a hardware problem.