Q: 2011 MacBook Pro and Discrete Graphics Card

Yes John I bet your pleased, it's your Company!

Perhaps Mavericks make graphics card run at higher temperature due graphics improvements, and this could be the reason that a lot of machines crash at the same time.

That's definitely not it. For one hadn't upgraded before my GPU broke, secondly the GPU is normally not used at all (unless you connect a second monitor or put on some GFX load) and certainly not always under full load and thirdly I just recently upgraded from ML to Mavericks and even if I force the dGPU on the power use and temperatures are  even slightly lower than under ML under about the same workload.

You can believe Apple is "more premium" or not, but please at least use a better example. IBM T42 in 2011? IBM hasn't made laptops in almost 10 years and the T42 is from 2004. The ThinkLight also had many other great purposes, like reading documents and books. You could even enable it with the computer off. MacBook Pro backlit keyboards don't help you do that, and having owned a T42, it is something I miss on all my computers today.

I'm one of the really lucky ones. I had repeated GPU (6750m) failures in 2011-2012 so they replaced mine with a low-end 2012 MBP (512MB 650M, runs out of memory all the time). My use probably made the issue much worse as mine goes through a ton of heat cycles. When I am plugged in, my computation software always uses ~25% CPU, leading the temps to 90-95C all the time because Apple's fan policy is "oh we are at almost TJMax, lets turn the fans on a bit". Then, when I'm not doing that, I'm using the computer to take notes in class, which is low stress and temperatures.

2012 MBP has been good except for mad flickering only on the Intel GPU, but how it will fare in 1-2 years remains to be seen. There are sometimes artifacts and flickering with my external monitor plugged in, but I am attributing that to Mavericks bugs.

I will upgrade when Broadwell drops, but it probably won't be Mac.  For example, the Dell M4800. 2 fans like MBP, but with Prime95 + FurMark (100% unrealistic situation) both chips are under 90C without throttling. In fact, it is built with the redundancy that if one fan fails, you can still run both chips without throttling (obviously closer to TJMax though). It may be a huge brick, but for me it is worth escaping these issues.

I have not had this issue yet. Has anyone with the anti-glare (non-glass) screen had this issue?

Yes, Mine's a BTO 17" with non-glare screen.

feel free to redistribute, I waive any copyright on the post below :

I would like to raise a principal issue on how Apple handles these problems having gone through the MBP 2011 GPU issue full cycle twice:

in a synopsis I have reason to believe that wide-spread issues such apparently this one are being systematically:

(a) not acknowledged,

(b) not communicated in symtoms and resultion within the service/support organization,

(c) only handled reactively by design and

(d) only taken care off in a transparent and proactive way once pressure through too many publicly reported incidences has reached significant proportions as for prior instances such as the buzzing first Intel motherboards that I experienced first hand. 

Here are my experiences in 4 acts substantiating this proposition through observations of how my 2 rMBP 2011 cases were handled:

Act 1: multi-month "diagnosis" phase

- when a rMBP 2011 started to throw graphics-related issues -- first symptoms being frequent _reproducible multi-monitor, hot-plugging triggered crashes October 2012 I:

* searched public information -> no obvious pattern through the google index

* suspecting software then I filed a diligently described bug report against the OS in http://bugreport.apple.com

* a week later the bug screeners asked for post-crash sysdiagnose output

* I could provide no sydiagnose dumps though as triggering them post-crash with Shift-Control-Option-Command-Period was not processed by the machine any more. Looking through telling log entries in /var/log/system.log (I am unix-macosx fluent) did not yield cause candidates over weeks.

Act 2: Trying to diagnose the issue myself for 4 Months

The next 4 months I tried to diagnose the computer's crashes diligently and in the process spending significant time on it.With sometimes several crashes per day with reproducible crashes on monitor hot-plugs at a rate of around 7 out of 10 tries there was plenty of opportunity to try things. Still, I suspected software considering there still was no obvious problem pattern visible through the google index. My interaction with the SW-Dev side of the Apple business did also not hint on hardware as a cause candidate. So besides the obvious of hoping for each OS service dot release to include a fix, I continued trying to find kernel module driver culprits, running HW diagnostics (yes you never know), watching the trigger pattern and for the most part the OS logs. In hindsight there were cues to a GPU issue as freezes appeared more often with GPU-related operations, but hot-plugging remained the only reproducible fault. Fast forward 4 months to the 3rd act.

Act 3: 1st motherboard swap

Having thoroughly exhausted my options on how to handle sometimes several crashes per day, I at last contacted Apple support. Not wanting to appear that I am looking for quick uninformed chancy HW swap I informed phone support on the issue history including the bug report# logged against the OS. I of course made note of the amount of time I had spent on it already.

* asked twice if there is anything in the customer support database(s) on a related issue - negative I heard.

The calls left off with the token on me to reinstall again and re-contact if the issue persists. Still, I would rate the handling on part of phone support perfectly right as after all I had just contacted them and the person on the other end of the line clearly did not have more information than what the google index provided to me (or to Apple support maybe as there was a hint that they look for things there too).

About 5 weeks later I finally gave up on the progress of getting the notebook's crash proneness resolved and contacted Apple phone support on feeling that pressing a hardware resolution is sensible having _really done everything possible from my side. The again dedicated!! 1st line phone service person on the ticket (where I would rate the phone service response again highly and this is not meant sarcastically) promised to talk to the repair shop if necessary. He even followed-up by email how things are going.

Getting a motherboard swap from the repair shop was no self-starter. I had talked the service counter through the multi-month ordeal. Yet, I got the sense that I had to be grateful for them swapping the mother board as the evidence of a HW issue was not water-tight. So I received the rMBP back with a new motherboard - the serial stayed the same, checked/asked - seems to be standard procedure to flash a new board with the previous one. Following the repair graphics related issues persisted, although slightly different, going to act #4. 

Act 4: Towards the 2nd motherboard swap

The machine showed _fewer (then how do you catgorize the repair?) monitor HDMI/DisplayPort/VGA over DisplayPort hot-swap issues then before, at a rate of maybe one crash every 3 days-they persisted though. I logically? concluded that the issue was software after all. Feeling bad having caused an "unnecessary" mother board swap with an OK board to land in electronics scrap I kept quiet an endured. The sure emergence of a next major release 10.9 made enduring with a hoped for fix bearable. Desolation set in when 10.9 did not resolve the crashes. I am not sure how this would have gone on had the HW not died conscutively.

Attaching a 30" Apple Cinema Display with an Apple DisplayPort adapter which had always worked not counting visible HF noise on the dual-link DVI started to fail. First it would not sync every so and so, a few weeks later attaching it and getting a picture failed completely. I cross-tested with other Macbooks and a Mac Pro, with no issue there. Days following loss of the ability to attach the Apple Display the rMBP started to show fails on booting with the built-in display only; it got stuck in the boot processes with too little verbosity in the console messages to be able to tell. Yet again, despite being exceedingly unlikely at that stage I wanted to rule out software and spent time (...) trying to find which was the last OS startup initilization.

Act 5: the 2nd swap

* Finally, another new symptom showed which made it utterly clear that it is a graphics-path HW issue; booting a rescue system off external media (conclusively ruling out any software cause) brought an audible sound in late boot as for an ok system being rescue booted? but no screen. I brought it to repairs, explaining the clarity of the case and how easy it was to reproduce, and expected a quick turn-around without issues. However, a few days into the repair I was called and asked to provided the purchase receipt. Provided that it was purchased abroad by a relative that was not an easy undertaking. A week into the repair I could provide a level of purchase documentation and the repair shop decided to ask Apple for a motherboard swap authorization (at t+7 calendar days into the repair). I of course provided information about the _long history of problems, the prior swap and for simplicity the repair was logged with the same shop that done the first repair. The Apple response was negative though.

* Issue now was that the system was already more than 1 year old. While 2 years extended warranty is statutory law in the EU, the warranty right is against the seller and not the manufacturer. The machine had been bought in Japan.

* I recontacted my Apple Service person from the 1st swap. Unfortunately, email bounced and calling in showed that he was not with the company any more. I then went through 1st line phone support again, explaining -- this time with hard facts that it _IS a HW issue. Through the serial number it was quickly triaged (...) to be out of the 1yr manufacturer warranty window. I explained it needing repairs, the reasonableness of the request given the history of the case and so on. 1st level handed it to a 2nd level manager who clearly was authorized to grant repairs which de-jure do not have coverage. Again, the handling was professional. Nevertheless, I was in the pleading role and that when it appeared edgy to go the wrong way for a net of 30 minutes talking to the 1st line staffer and the manager on the phone. When 2nd level hinted she would probably approve, but "first needs to talk to the repair shop just to be correct" I knew I was close to a 2nd repair without having the de-facto total loss of the machine due to replacing a broken motherboard probably not being economical. Following such a path or procedure (?) was perfectly correct. The motherboard was finally swapped for "GPU problems" shortly thereafter (total time for the repair approx. 9 calendar days). The experience left me with a manufacturer to-be-perceived notion that I should be grateful for the free repair. While I stayed cordial in the whole process, the gratitude stayed within the dry limit of having safed a 1.5 yr. old machine from scrapping had the motherboard not been replaced for free. 

Knowing now that many people are affected and surmising that there is an organization within Apple, that having the data -- see Epilogue -- knew all along poses a question. Why were at least the service 1st-and 2nd line, let alone external repair left in the cold information-wise for this problem. I leave it partially to the reader what to make of it.

Epilogue

Again, I want to underline that the cases were handled very well in the scope of the involved people's roles (Apple support 1st and 2nd line, repair shop and possibly development in software ticket screening). In other words, if someone in Apple management reads this and  -- and had I provided the case #s -- the right reaction would NOT be to reprimand the involved employees; they did everything right given the information and probably guidelines given to them, responsively and professionally. Had I rated them on one of the ubiquitous post-contact customer satisfaction surveys, I would have given them a perfect score. So that is not where the problem is.

Still, the 2 issues were grossly misshandled, knowing now that discrete GPU issues in this model type are wide-spread.

There is an organization within Apple which with little lag probably clearly sees the data on problem reports and repair incidents around the world. With near certainty, those are being tracked meticulously as substantial costs are involved. They probably have all the lab resources to reproduce and find root-causes, it not alone then with e.g. the GPU manufacturers. If they have more information and see high fault incidents, and are able to extrapolate the numbers of affected machines as it is simple sample to population statistics, what do they do? Apparently, this part of Apple or Apple as such does not proactively act on clearly significant field hardware problems that are affecting very many machines. This is a conjecture based on my experiences above, and please don't sue me on it. 

- Cheers,

an early 2011 rMBP owner with 2 motherboard swaps  

Does anyone have any suggestions/thoughts on what to do if switching to integrated mode using gfxcardstatus (after rebooting from a freeze) causes the system to immidiately freeze again? It appears that others have experienced this behavior as well (see comments copied below).

I'm afraid to do/work on anything without being able to switch to integrated mode because the system will crash w/o any warning.

One suggestion earlier in the discussion (page 62) was to reinstall (remove?) the AMD drivers:

David Nazar wrote:

 

I found something that can maybe help :

Solution given by Davide Pasca
To follow on my previous suggestion. It's actually better to boot in shell mode, instead of waiting for the Mac to overheat.

- Press Cmd + s at start, and wait for the shell come up
- At the prompt, type:

mount -uw /
mv /System/Library/Extensions/AMD* ~/Desktop/
shutdown -h now


..restart and it will work. I just tested it with Mavericks, that reinstalled the AMD drivers and wouldn't boot again.

 

Source : http://www.mactrast.com/2013/10/owners-early-2011-macbook-pro/

 

I have not tried yet.but I will tonight.

Is this the right approach if using gfxcardstatus to switch to integrated starts crashing the system??

Thanks.

========

odarellmc wrote:

 

gfxCardStatus now freezes my computer anytime I try to switch to discrete graphics. Anyone else experienced this?

I am experiencing this as well now. Anyone else??

odarellmc wrote:

 

sorry i meant my computer freezes when i try to use gfxCardStatus to switch to intergrated only....anyone experieincing this...sorry for the confusion earlier....

(Just incase there was any confusion)

odarellmc wrote:

 

Am i the the only one experiencing the loss of the start-up chime due to this issue?

I also experienced no startup chime just before I became unable to switch gfx to integrated mode after rebooting from a freeze.

The solution provided by Davide Pasca will remove the drivers for the problematic AMD descrete graphics card. This will mean the system will not attempt to use the GPU and will instead default to using the integrated graphics.

          mv /System/Library/Extensions/AMD* ~/Desktop/

That will move all extensions (drivers) with a file name starting with "AMD" to the desktop.

IMHO This should only be done as a last resort if gfxCardStatus is not working for you ie. MBP crashes before gfxCardStatus can do its thing. Unfortunately that is the case for many people - like me.

It seems that most versions of OSX without these drivers will perform very poorly - sluggish rendering seems to be the main complaint. According to reports I've read this sluggish rendering make the computer unusable for real work.

And once you've removed the drivers they can be tricky to reinstall if you want to because permissions on the drivers need to be corrected once you move them back into the Extensions directory. Not a problem if you're familiar with *nix permissions and the CLI.

On my install of OSX Lion these drivers all start with "ATI" rather than "AMD" so the command to move them is:

          mv /System/Library/Extensions/ATI* ~/Desktop/

On my install of Lion with the ATI driver removed I have a very stable and usable system now. It doesn't crash but the dGPU is not available so the graphics are not as good as they could be in certain applications (many games, Photoshop etc). But the rendering does NOT feel sluggish as some people have reported. If I start up in Safe mode then I encounter the sluggish rendering issues. I suspect that in Safe mode even the integrated graphics drivers are disabled.

I'm not planning to upgrade the OS beyond Lion because I suspect that removing the AMD drivers in future versions of the OS also causes the integrated graphics drivers to be disabled and this causes the sluggish graphics.

I'd be interested to hear from anyone that is running a more recent OS and has successfully removed the AMD drivers and still has good rendering performance.

I'm still waiting for Apple to address this issue - either a recall or a replacement logic board which will not fail again.

Add another to the list I guess. Just started getting the distortion and freezes/reboots today. 15" MBP

AMD Radeon HD 6750M and Intel HD Graphics 3000 with the Hi-Res anti-glare display.

Depends on what you're working on.

Close EVERY Finder window. Finder will lock your system more than any program, even with Integrated Graphics Only selected.

If it's graphical in nature, good luck.

I do an HDR rendering and it just tanks.

I can't believe Apple has let this get this out of control. It's sad because 3 people brought their units into the Apple Store in Palm Beach Gardens, last night.

Most were turned away with the plausible denial.

Finally I showed one lady this thread and the Cult of Mac article.

The employees know, but they can't admit they know.

Or they become ex-employees.

Added.

Don't worry, the fanbois will find a way to blame you for it.

Did you immunize your machine, I mean, have AppleCare?

The situation is only getting and only going to get worse.

Before yesterday, I had been tracking the issue in Mail. We would see at most 20 replies a day. Sometimes not even 3-5.

Yesterday there were 122, and already as of 7:50 EST tonight 81 new replies.

The reason it is getting worse is simple.

This is not a software issue. It's not even technically the GPU.

The solder and lack of adequate thermal paste have created a monumental failure that cannot be resolved with logic board replacements.

What is Apple supposed to do here?  I really have no idea.

The issue is more than widespread. It will get probably every unit, regardless of preventative measures.

todzilla78 is correct!

I first replied to this thread a couple months ago when I experienced the issues described here. I kept my email notifications on to track the progress but the last few days it's been getting really really annoying. I'm going to have to turn it off as my inbox is blowing up, but I'm happy to see the issue is really getting this much attention. Several posts here, Facebook groups, news articles, blog posts and petitions. It's only a matter of time before Apple has not choice but to do something about it. They can't keep lying to us for much longer - claiming that this issue is not very widespread is complete BS. Keep making noise people.

ciu5781 wrote:

 

D3us wrote:

 

ciu5781 wrote:

 

Is this caused by heat? If so, it's users fault. Not apple's.

Really ciu5781?

So, it's the user's and not a desing or production process fault that the solder balls between gpu and motherboard loose contact?

How did you come to that conclusion?

Because when the user uses it, it gets to warm?

So, one should not use it to avoid it getting warm and breaks?

 

Explain your conclusion please?

I don't have confirmation it's caused by heat. But if it is, yes, you should avoid it getting hot. It shorten the lifespan of your machine. Why do you think there are plenty of laptop cooler in the market? Because people want to cool it to expand the life of the machine. Unlike inside the desktop machine like Mac Pro the condition inside the mobile machine is worst.

They even fail on watching youtube videos.

If Apple puts in a 2nd graphics chip for better performance it should make sure it's cooled enough.

Not the users responsability.

For me, the more macbooks fail the better, as I repair them.

But I am not like that.

I really feel sorry for all the users experiencing this problem.

Which is entirely to blame Apple/foxconn/production process/designers for.

Not the user.

What you say is like a car manufacturer selling a car with a turbo for extra power telling the customer it's his responsability to make sure it gets extra cooling if he wants to use that Turbo.

This thread has been blowing up tremendously lately! I'm so glad it is finally getting the attention we have all been waiting for! I agree, it is only a matter of time at this point.

I'm still waiting on a reply from Tim Cook. I sent a very lengthy email on Wednesday and have not recieved a reply yet. I'm trying to be patient, but I am running out of it just as everyone else here. I'll keep everyone notified if and when I recieve a response!

I have my macbook paperweight sit down for a month now. This situation is really annoying.

Do you think theres something we can do (as users) to speed up the process of apple acknownledging this issue?

Santi RT