Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I might have a similar issue with my lg ultra wide monitor as per this video (towards the end)

My solution is to flick the switch at the back on / off otherwise you have to wait for the grey 'no input' screen then wait for the standby light to flash THEN press the one button to wake it.

The 31MU972-B has been working fine on El Cap, all day yesterday and so far this morning. I really think it's a Sierra issue.

On Sierra if I powered off completely, then rebooted but didn't login, then let the Mac sleep, then tried to wake it - the screen would never wake w/out pulling the plug on the Mac and restarting it, no matter what monitor on/off or thunderbolt port dance I did.

This worked the first time I tried it in El Cap. I don't think a day is enough of a sample test for the other random times the screen wouldn't wake up in Sierra but I'll let the thread know if it happens again.

Now I just need to figure out a contingency plan should this system fail that doesn't rely on Apple.
 
The 31MU972-B has been working fine on El Cap, all day yesterday and so far this morning. I really think it's a Sierra issue.

On Sierra if I powered off completely, then rebooted but didn't login, then let the Mac sleep, then tried to wake it - the screen would never wake w/out pulling the plug on the Mac and restarting it, no matter what monitor on/off or thunderbolt port dance I did.

This worked the first time I tried it in El Cap. I don't think a day is enough of a sample test for the other random times the screen wouldn't wake up in Sierra but I'll let the thread know if it happens again.

Now I just need to figure out a contingency plan should this system fail that doesn't rely on Apple.

Surely the support guy you've talked to can't dispute this as evidence you don't need a GPU replacement, but that he just needs to pass this on to engineering... Could he?
 
Surely the support guy you've talked to can't dispute this as evidence you don't need a GPU replacement, but that he just needs to pass this on to engineering... Could he?

When I brought that up during the first time he mentioned that he wanted to replace the cards his explanation was something to the effect of that Sierra might use the GPUs differently than El Cap and it's exposing some kind of flaw...
 
Just take the SSD out and let them fix the issues.

If they question it, it's like when I always removed the HDDs in my old Mac Pro before service.
 
When I brought that up during the first time he mentioned that he wanted to replace the cards his explanation was something to the effect of that Sierra might use the GPUs differently than El Cap and it's exposing some kind of flaw...

… Yes, the OS renders elements differently with the GPU under Sierra compared to El Cap (in particular more elements are written with Metal and Swift, an example being the Dock). Will this expose a problem with the GPU that you wouldn't notice otherwise? Very, very, very unlikely. Especially if you actually use graphics for anything at all. It's very hard to believe that this issue occurs because some particular call is being made to the GPU that was never made before, by any application. I don't put much faith into the technical knowledge of the support guy.
 
… Yes, the OS renders elements differently with the GPU under Sierra compared to El Cap (in particular more elements are written with Metal and Swift, an example being the Dock). Will this expose a problem with the GPU that you wouldn't notice otherwise? Very, very, very unlikely. Especially if you actually use graphics for anything at all. It's very hard to believe that this issue occurs because some particular call is being made to the GPU that was never made before, by any application. I don't put much faith into the technical knowledge of the support guy.
As a software engineer, I've seen many cases where changes to software expose bugs and problems that were never seen before.

I would consider it very possible that a shift to Metal and Swift can expose GPU problems that were never seen before.
 
As a software engineer, I've seen many cases where changes to software expose bugs and problems that were never seen before.

I would consider it very possible that a shift to Metal and Swift can expose GPU problems that were never seen before.

I will not at all dispute you, but would this not only be the case if not similar Metal/Swift calls were made to the GPU on El Cap? Now considering that Safari, iMovie, Photos, and many many more used Swift and Metal on El Cap, and I can't see which operations the Dock could need that the other applications didn't use, I find it unlikely, unless of course our OP didn't actually use any Metal/Swift apps on El Cap. Would you agree with this statement?
 
This is a known problem with the LG Thunderbolt 1440P monitor and I don't think there is anything you can do on the Apple side to fix it, so I'm curious if you have that particular model.
This is a known problem with a lot of monitors with Apple. I have it with my 4k Dell. I've had it with my 1440p Crossover. Have had it with DLP projectors... It's been a problem for years. Never happens when I'm in Windows though so pretty sure it's software.
 
I will not at all dispute you, but would this not only be the case if not similar Metal/Swift calls were made to the GPU on El Cap? Now considering that Safari, iMovie, Photos, and many many more used Swift and Metal on El Cap, and I can't see which operations the Dock could need that the other applications didn't use, I find it unlikely, unless of course our OP didn't actually use any Metal/Swift apps on El Cap. Would you agree with this statement?
Anytime you make changes to the patterns of calls to lower level APIs, you can expose problems. Anytime. Perhaps the new code has bugs, or perhaps in the previous calls bugs were masked.

When I was involved in operating system releases, after field test you couldn't even fix typos in messages. The new message might be shorter or longer, and expose problems. (You could fix the typos in the .1 release, since that will go through a full testing cycle.)

"Similar calls" don't mean anything. "Exactly the same calls, in the same sequence, with the same arguments and the same timing" means little. Computers should be deterministic, but it's almost impossible to comprehend all the factors in play, so in practice they seem to be non-deterministic. For example, a kernel component might have a "use after free" bug - some memory is read or written *after* it has been returned to the OS. This might not be noticed for years - or only noticed when a system is low on memory and the freed memory is quickly reused. (You'll sometimes see posts about people saying that memory pressure makes a system unstable. That's pure bunk for the most part. What is usually going on is that running low on memory exposes "use after free" bugs. Running low on memory should cause a computer to run more slowly, it should not cause the system or apps to fail.)

All of this is why you hear people say "I never install 10.x - I wait for at least 10.x.2" or "I'll wait for Service Pack 1". Changes to the OS components expose hidden bugs, and it takes a while to get things stable again.
 
Last edited:
  • Like
Reactions: casperes1996
Am having the same issue with a 2013 Pro & dell UP3216q, it seems to skip OS upgrades and every second or third upgrade the problem rears its head. Would be interested to hear if you ever get it fixed.
Not specifically to your answer, but regarding thunderbolt + external stuff issues. It seems that there is also a reoccuring issue that is OS dependant with Firewire800 and thunderbolt concerning some external hard drive enclosures https://discussions.apple.com/thread/5486167?tstart=0 as well.

Maybe another hint at software... (or software + chipset of the external device).
 
I really am curious about what they use inside Apple when they work on their own apps/OSes, etc. I wonder what their workflow for problems is as well, do they just call IT and someone comes up with a new Mac?

They use iMacs. Or at least that's what my contact uses/used. Note though that this is someone who works on low level OS stuff, not application UI.
 
Anytime you make changes to the patterns of calls to lower level APIs, you can expose problems. Anytime. Perhaps the new code has bugs, or perhaps in the previous calls bugs were masked.

When I was involved in operating system releases, after field test you couldn't even fix typos in messages. The new message might be shorter or longer, and expose problems. (You could fix the typos in the .1 release, since that will go through a full testing cycle.)

"Similar calls" don't mean anything. "Exactly the same calls, in the same sequence, with the same arguments and the same timing" means little. Computers should be deterministic, but it's almost impossible to comprehend all the factors in play, so in practice they seem to be non-deterministic. For example, a kernel component might have a "use after free" bug - some memory is read or written *after* it has been returned to the OS. This might not be noticed for years - or only noticed when a system is low on memory and the freed memory is quickly reused. (You'll sometimes see posts about people saying that memory pressure makes a system unstable. That's pure bunk for the most part. What is usually going on is that running low on memory exposes "use after free" bugs. Running low on memory should cause a computer to run more slowly, it should not cause the system or apps to fail.)

All of this is why you hear people say "I never install 10.x - I wait for at least 10.x.2" or "I'll wait for Service Pack 1". Changes to the OS components expose hidden bugs, and it takes a while to get things stable again.


Hadn't thought of it like that. You deserve a cookie for being informative and nice on this forum.
 
Just take the SSD out and let them fix the issues.

If they question it, it's like when I always removed the HDDs in my old Mac Pro before service.

I asked to do this specifically and was told no because it's not a user replaceable part.

… Yes, the OS renders elements differently with the GPU under Sierra compared to El Cap (in particular more elements are written with Metal and Swift, an example being the Dock). Will this expose a problem with the GPU that you wouldn't notice otherwise? Very, very, very unlikely. Especially if you actually use graphics for anything at all. It's very hard to believe that this issue occurs because some particular call is being made to the GPU that was never made before, by any application. I don't put much faith into the technical knowledge of the support guy.

I get what you're saying and the conversation between OS updates, at some point the hardware is running the code natively though - that's one thing that usually doesn't change between OS releases.

They use iMacs. Or at least that's what my contact uses/used. Note though that this is someone who works on low level OS stuff, not application UI.

I really do wonder if I made the right choice...
 
I get what you're saying and the conversation between OS updates, at some point the hardware is running the code natively though - that's one thing that usually doesn't change between OS releases.

Of course, but as someone else pointed out, there can be a lot of things at play at once, making it difficult to determine if the software change really could alter the way the GPU handles the output enough to bring issues to light like this. The frame buffer implementation could be differently implemented for instance, which could easily lead to clunky issues if the hardware was in some way defective.

I still doubt it, but I think getting a new GPU installed just to check it, and at least getting a new GPU if it's not the problem anyway, is a good idea. That way you could probably also have an easier time getting the case pushed further. I get you have work to do, and being without the machine until Apple does the replacement is a hassle, but I think you should seriously consider if you in some way can use something else while the replacement is going on to make it possible to live without.
 
Of course, but as someone else pointed out, there can be a lot of things at play at once, making it difficult to determine if the software change really could alter the way the GPU handles the output enough to bring issues to light like this. The frame buffer implementation could be differently implemented for instance, which could easily lead to clunky issues if the hardware was in some way defective.

I still doubt it, but I think getting a new GPU installed just to check it, and at least getting a new GPU if it's not the problem anyway, is a good idea. That way you could probably also have an easier time getting the case pushed further. I get you have work to do, and being without the machine until Apple does the replacement is a hassle, but I think you should seriously consider if you in some way can use something else while the replacement is going on to make it possible to live without.

I will eventually, I've been on El Cap with no issue since Friday so I'm not really inclined to try and set this up at the moment.

If it ain't broke...
 
  • Like
Reactions: ActionableMango
If it ain't broke...

"We believe if it ain't broke, don't fix it. Engineers believe that if it ain't broke, it doesn't have enough features yet"
-Scott Adams

But yeah. As long as your workflow works on El Cap, and you're still well within warranty, who really cares? – Except of course for the service in general being something to care about
 
  • Like
Reactions: thefredelement
"We believe if it ain't broke, don't fix it. Engineers believe that if it ain't broke, it doesn't have enough features yet"
-Scott Adams

But yeah. As long as your workflow works on El Cap, and you're still well within warranty, who really cares? – Except of course for the service in general being something to care about

Exactly, I've got another 2 years of Apple Care.. I'm more stuck on getting a contingency plan. I'm ranging between a Hack or 5k iMac (especially if I can figure out how to get the iMac to be the machine I use all the time but the Pro does the compilation), I'm not sure though and need to do more research. I'm also not thrilled about buying more Apple stuff currently - though I could get into that JointVenture program with the purchase which may also help to alleviate the amount of time of any future technical issues.
 
Exactly, I've got another 2 years of Apple Care.. I'm more stuck on getting a contingency plan. I'm ranging between a Hack or 5k iMac (especially if I can figure out how to get the iMac to be the machine I use all the time but the Pro does the compilation), I'm not sure though and need to do more research. I'm also not thrilled about buying more Apple stuff currently - though I could get into that JointVenture program with the purchase which may also help to alleviate the amount of time of any future technical issues.

Well, a hack will obviously be more flexible, but I've got to say that I absolutely love my 5k, and hacks could bring problems down the line if Apple decides to change something. It'll most likely be fixed, but you will again have a problem if Xcode for some reason doesn't work with your hardware for a month or something. The 5k display on the iMac is also super brillant, and text is just so much more pleasing to look at. I may however be partial to the machine as it's what I use day to day for my film work.
Regarding doing compilation on the MacPro, and take into consideration I haven't tried, couldn't you just ssh into the MacPro from the iMac and run compilation from the Terminal?
 
Just got off the phone w/the person at Apple. There is no supervisor, he's at the 'top of the food chain' as he put it, so I'm stuck in that regard.…

It's not just this issue now, what if this Mac is fried and my work just stopped dead in it's tracks? I need a company that has my back.
Your experience is troubling. It reminds me how much better SquareTrade warranties are for some things. They are cheaper than AppleCare and have a five-day-repair-or-replace guarantee. In practice, that window might be extended if a serviceman arrives within 5 days and must order a part. Other advantages are on-site service and SquareTrade makes the mind-numbing phone calls to the technician. (I have to wonder how that would work out with an issue as complex as this.)

Hopefully, your issue will be resolved soon. I am always saddened when I hear Apple losing its luster.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.