Q: 2011 MacBook Pro and Discrete Graphics Card

@ tjohnsoniv1200

tjohnsoniv1200 wrote:

 

Not sure.  Found a bunch of 'Kernal 0x00000000', but nothing specifically with the 'GPU Hang State = 0x00000000'

On the left side you can see 'system.log', you have to check there. And also go to the last section 'private/var/log', there are some more system.log.1.gz…

Enjoy your 

sorry no luck finding it

Orcinus wrote:

 

ante@macbook:~$ kextstat | grep AMD

   85    2 0xffffff7f81071000 0x6d000    0x6d000    com.apple.kext.AMDSupport (1.1.4) <84 78 11 10 7 5 4 3 1>

   86    0 0xffffff7f810de000 0x1d3000   0x1d3000   com.apple.kext.AMD6000Controller (1.1.4) <85 78 11 10 5 4 3 1>

  103    0 0xffffff7f8142b000 0x3a1000   0x3a1000   com.apple.AMDRadeonX3000 (1.1.4) <102 78 11 7 5 4 3 1>

  114    0 0xffffff7f819b1000 0x14000    0x14000    com.apple.kext.AMDFramebuffer (1.1.4) <85 78 11 10 7 5 4 3 1>

 

 

Note the AMDFramebuffer.kext loaded.

Yes, integrated graphics are by Intel. But the video outputs (both LVDS and DisplayPort) go through AMD GPU's framebuffer.

 

So no, i am not absolutely wrong, no matter how many exclamation marks you use.

Stop parroting the same info over and over again. Do some research instead.

 

apple_power wrote:

 

@ Orcinus

Orcinus wrote:

 

That's the fallback driver kicking in.

By removing *all* drivers containing the string "AMD" you've also removed the one supporting the integrated (Intel) GPU. So Mavericks booted using a fallback "safe" driver.

 

Sorry, you are absolutely wrong!! The internal GPU does not relay on any AMD*.kext…

 

Think first… The internal GPU is made by Intel not by AMD who makes the AMD GPU which used to be ATI before it was acquired by AMD!!

 

Enjoy your 

Very smart, for sure, when you still run the AMD, all AMD.kext are showing up. Go to your 'System Information' and check what Graphics Card is active…

And sure the externel display is not working when you disable the AMD and ATI drivers.

Enjoy your 

tjohnsoniv1200 wrote:

 

sorry no luck finding it

Then you got an other problem…

Can you tell me how exactly your computer behaves? Can you post some pictures?

Enjoy your 

tjohnsoniv1200 wrote:

 

I called apple support this morning and spoke to a Sr tech.  He mentioned the AMD 6750M version was a known issue and they were offering repairs Unfortunately, my computer is running the AMD 6490M.  They are forwarding my case to the Engineering department.  With 1000's of people having the exact same issue and with 85 pages of complaints on this forum alone..not to mention all of the other forums/news articles/lawsuits about this exact issue you would think Apple would want to address this.

This is great news!

An update on my early 2011 MBP (AMD 6750M)... the genius bar kept it overnight and ran more hardware tests and concluded the 16GB OWC RAM failed the test. They tested in twice and reseated it just to make sure. That was Tuesday. Today I was watching a video and I got yellow vertical lines (what color will it be next?! yay!) then the screen went black, the fans ramped up to full blast and the speakers emmitted repeating bouts of static. When I hard restarted and I arrived on my desktop, I had some random dime-sized blip of static appear and disappear on the screen. Yikes. Off to the Apple store I go again! I will press them about this a lot more seriously when I go tomorrow.

When mine was acting up, I never had anything about the GPU in my system logs. I scoured them after every crash (so very frequently) but there was nothing there.

For everyone here with this problem, what was the general purpose of their MacBook Pro? Keeping up with this thread, I noticed for many their MBP was primiarily used for video editing, which fully utilizes the GPU.  Majoriry of those here perform GPU extensive-tasks on their MBP daily?

I used it for software development.  Some java libraries (swing) are hip to the gpu and turn it on when rendering certain controls.  It killed the mac in 23 months.

Has anyone here replaced the motherboard, and then sold the computer (with a clean conscience).

Web development, light Photoshop, iOS development and occasional games in Windows 7 bootcamp. Very rarely with an external monitor. But more recently web browsing only because I can't use my personal computer in my current job. It died after 19 months.

When my MBP first started acting up the most GPU intensive things I would do on it were watching 720p YouTube videos and then after my first logic board replacement I started playing games and it broke again like two weeks later. So now I'm on my second logic board, third display and second hard drive. Now I use gfxCardStatus and just use integrated graphics and no longer watch HD videos nor play games on it until Apple does something to resolve this issue

I'm not convinced this is true. Why wouldn't similar but not exactly the same hardware failures throw a variety of errors? In my opinion it takes a much more definitive answer than just common error logs between the people reporting the problem. Additionally it's obvisouly affecting more than one chip.

Same symptoms across more than one chip will throw multiple errors for the same root issue IMHO.

From what I've gathered via this thread it seems that the hard evidence points to non iron solder in combination with an excess of thermal paste on the GPU being the root cause. Which has been patched by private engineers re-soldering the GPU and removing the excess paste.

Correct me if I am wrong, but that is what makes sense to me. The exact symptoms unite the thread, not the specific arbitrary error messages console is spitting out.

@apple_power are you saying that anyone that doesn't have 'GPU Hang State = 0x00000000' in their system.log has a problem different that yours?

Just because it's not reported in the same manner as you see doesn't mean the root cause is not the same. The reporting may depend on OS and drivers. And the problem may manifest in various ways and generate different errors.

I don't think we should be excluding people just because of this error message that you are seeing. Is ANYONE else seeing this error message in system.log?

I'm not seeing it but I'm able to reproduce the problem by simply re-enabling my ATI drivers (I'm running Lion) and restarting. Then it'll crash before the login screen (just a gray screen) and sometimes boot up to the Finder but then crash moments later. There is no sign of 'GPU Hang State' in the log.

I've seen graphical glitches just as others have shown - split screen, interlacing (offset lines of pixels): https://dl.dropboxusercontent.com/u/1383484/discrete-glitch.jpg

Though I've never seen bands of colors as many have described.

Maybe we are encountering different problems but 'GPU Hang State = 0x00000000' is not evidence enough to call it at this stage IMHO.

I'd pored over the logs on my failed logic boards and that isn't a log message that I remember seeing.  There were various other failure messages that coincided with crashes or unsuccessful boots, but nothing that gave me any hints as to what was happening.

It wouldn't surprise me if there are multiple  problems represented within the thousands of messages on the threads about 2011 MacBook Pro graphics issues, but I'm also not convinced that we have any tried-and-true means of differentiating between them.  A failing/broken GPU could exhibit any number of different symptoms and errors.

At least in my case (Lion with ATI*.* drivers removed from Extensions dir) I find that my MBP runs with the Intel integrated graphics perfectly. This is very different from when I start up in Safe Mode. In Safe Mode the graphics are slow and stuttery. Screen refreshes appear clunky.

For me disabling the ATI* files doesn't seem to effect the integrated graphic in a negative way.

Running on the integrated with the following files disabled I have a system that is stable and fast enough for casual use. Photoshop and VMWare fusion both warn me that graphics are not up to the job but still continue to run reasonably well.

krayzi:~ mike$ ls -l /System/Library/Extensions\ \(disabled\)/

total 0

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI1300Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI1600Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI1900Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI2400Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI2600Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI3800Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI4600Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI4800Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI5000Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATI6000Controller.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIFramebuffer.kext

drwxr-xr-x  3 root  wheel  102 Apr 10  2012 ATIRadeonX1000.kext

drwxr-xr-x  3 root  wheel  102 Apr 10  2012 ATIRadeonX1000GA.plugin

drwxr-xr-x  3 root  wheel  102 Apr 10  2012 ATIRadeonX1000GLDriver.bundle

drwxr-xr-x  3 root  wheel  102 Apr 10  2012 ATIRadeonX1000VADriver.bundle

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX2000.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX2000GA.plugin

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX2000GLDriver.bundle

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX2000VADriver.bundle

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX3000.kext

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX3000GA.plugin

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX3000GLDriver.bundle

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATIRadeonX3000VADriver.bundle

drwxr-xr-x  3 root  wheel  102 Apr 18  2012 ATISupport.kext

SpaceBacon wrote:

 

...From what I've gathered via this thread it seems that the hard evidence points to non iron solder in combination with an excess of thermal paste on the GPU being the root cause. Which has been patched by private engineers re-soldering the GPU and removing the excess paste.

 

Correct me if I am wrong, but that is what makes sense to me. The exact symptoms unite the thread, not the specific arbitrary error messages console is spitting out.

 

We don't really have hard evidence of much.  I guess we know for a fact that solder joints are failing, but that doesn't necessarily mean that it was a bad batch of solder.  It could be as simple as a mistake in Apple's thermal recipe when they cooked all of this.  Or it could be a combination of things.  It could be brittle solder combined with the CTE mismatch between the package substrate and PCB (Why does 2011 MBP have corner bonds, but not 2012 MPB?).  It would be nice to see some ersascope images of the failures that have occurred. 

Those choosing to reball with lead are doing so to target more variables at once.  I would choose the same given what we currently know.

The ludicrous amount of thermal compound is probably not very helpful, but my guess is that it's mostly playing the role of bringing these failures on a little sooner than might have otherwise happened.