Q: 2011 MacBook Pro and Discrete Graphics Card

Tried to reboot in normal mode but ended up at the blue screen. Back in safe mode now, which seems to work without gfxCardStatus and smcFanControl. I'm out of ideas.

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.

After having had my laptop repaired three times (with Apple claiming having replaced the logic board with a new one each time) and the screen as well the last time, I am almost certain that the whole thing has to do with the latest SMC firmware update (http://support.apple.com/kb/dl1633). I have experienced the computer going wrong everytime I had that firmware installed. I am heading down to the Apple Store for the fifth time to try to either get my money back or check in the computer for another time.

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.

Tried it... Unfortunately it didn't work for me

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

That the whole point of this thread.  The discrete GPU kills the machine after it's physical connection to the motherboard becomes unstable. 

You use gfxCardStatus to prevent it from switching on.  Your options:  Get used to using the integrated GPU, get a new motherboard, or get a new computer.

Sorry for spamming, but a problem that I think has been mentioned before, I just can't find the post.

While in safe mode (the only mode that works for me) I switched gfxCardStatus to "Discrete only", which worked fine. I then tried to switch to "Integrated only", which resulted in the prompt "You can't switch to Integrated Only while the following item is running: External Display".

When I open "Displays" in system preferences I can't see any external displays, however, SwitchResX shows an "Unknown External Display". The last time I used an external display was a couple of weeks ago, so quite a while before the problems started.

Is there any way to force a universal disable of external displays?

Flamenco13 wrote:

 

That the whole point of this thread.  The discrete GPU kills the machine after it's physical connection to the motherboard becomes unstable. 

 

You use gfxCardStatus to prevent it from switching on.  Your options:  Get used to using the integrated GPU, get a new motherboard, or get a new computer.

I'm using my MBP with only integrated graphics and I have to say, it is not flawless. Yesterday I was watching a video using the Air Video Server streaming through my ATV and at some point the MBP rebooted itself.

I also had the 'external monitor issue' with no monitor attached.  It prevented me from booting, so gfxCardStatus was of no use.

Boot into single user mode, then move the discrete GPU drivers out so the boot sequence will not try to enable it.  You will have BAD graphics, but bad graphics is better than no graphics.

see https://discussions.apple.com/message/22710793#22710793.

Flamenco13: I've already moved the drivers thanks to your tip, that was what let me boot into safe mode in the first place. Before I moved the drivers I couldn't boot into safe mode. Updating the system resulted in having to move the drivers again. I'm currently in safe mode and able to switch to discrete graphics without any problems, however trying to switch to intergrated resulted in the error message above.

I am following all the threads about this problem as I have the early 2011 model and have strong indications that this SMC update has, at least a bit, something with this problem. Coming from the windows/linux world and assembling my computers for more than 17 years, I already broke my fair share of motherboards due to bad bios upgrades. With time you learn that you apply a bios upgrade only to solve a problem that is happening in your motherboard. If your computer is running fine you don't need to do anything.

Even attesting this, I don't believe that Apple had pushed a bad SMC update. Maybe it trigged the problem, but it is not the main cause.

When the 6750m was launched I read about some people experimenting to over clock this chip:

By default 6750M runs 600/800 (core/mem) at 1.00V, however these clocks run fine at 0.855V. This means it ispossible to overclock using lower than stock voltage which will enable faster performance and lower operating temperatures (e.g. 750/950 at 0.955V). A win/win situation. You can always choose to overvolt (>1V) for a higher overclock (above 800 core). I'm not responsible for any damage to your hardware as a result of this guide, exercise caution and keep an eye on your GPU's temperature.

FINAL UPDATE : ~800/900 @ 0.955V is the best all-round overclock. Performance gains through overvolting and going above 800 core and 900 mem are minimal and not recommended.

If this is really true, this chip is suffering the effects of over-voltage. In fact most of the images about this problem shows a typical Vram error. The banding and blocks of color are the result of the bad sectors that couldn't be mapped out. Maybe it is the reason where some people could boot and use the discrete gpu until an affected area of the memory is requested.

Well, this is just conjectures from my part, based in everything I am reading about this problem.

My SMC is this one:

SMC Version (system):    1.69f3

Maybe my conclusions are completely wrong, but I am curious to know how many of you didn't update the firmware. I believe that this update could have trigged I voltage rise on the gpu and Vram. As the overclockers concluded the chip is already overclocked by default, a minimal voltage change could fry the components.

(just to add to the situation, I checked the 'software update' right now and the SMC 1.7 isn't showing anymore. Weird…)

*sorry for my bad english :-/

I never had attached a display to my laptop, but gfxCardStatus still showed it as having one attached.  If you removed your drivers, you are probably using the integrated even though the gfxCardStatus menu says discrete. 

I could never find a way to force the system to ignore the phantom external monitor.

You might be right about the integrated drivers being used despite the menu saying discrete. Anyways, I took some screenshots in case anyone is interested: http://imgur.com/a/MiFIf

The system profiler shows the following info for the discrete gfx card:

AMD Radeon HD 6490M:

  Chipset Model:          AMD Radeon HD 6490M

  Type:          GPU

  Bus:          PCIe

  PCIe Lane Width:          x8

  VRAM (Total):          256 MB

  Vendor:          ATI (0x1002)

  Device ID:          0x6760

  Revision ID:          0x0000

  ROM Revision:          113-C0170H-521

  gMux Version:          1.9.23

  EFI Driver Version:          01.00.521

  Displays:

Display:

  Resolution:          1680 x 1050

  Pixel Depth:          32-Bit Color (ARGB8888)

  Main Display:          Yes

  Mirror:          Off

  Online:          Yes

And the following for the integrated:

Intel HD Graphics 3000:

  Chipset Model:          Intel HD Graphics 3000

  Type:          GPU

  Bus:          Built-In

  VRAM (Total):          384 MB

  Vendor:          Intel (0x8086)

  Device ID:          0x0116

  Revision ID:          0x0009

  gMux Version:          1.9.23

Might be just a coincidence, but my mbp started to present the issues right after the SMC update.

So, there comes the question : Is there a way to rollback a SMC update ?

@GregD.MacUser I'm parisian too, pls send me an email to t@_-d_-e-_s-_-a-_DOT_-r-_-s-_ (sans les tirets et underscores)