Early 2011 15 inch Macbook Pro crash

[coldwaves]

I had a early 2011 Macbook Pro 15 inch with 2.0GHz CPU, 128G SSD and 8GB RAM. Everything apple stock. Two days ago, the computer suddenly crashed with the display pixelated and splited. I was running a statistic computation during the crash, which uses 12.5% CPU. After that, I cannot boot into the OS. The apple logo will appear and then stays at the grey screen forever. I tried recovery mode or safe mode or boot from a bootable USB, none of these are successful and I always get the apple logo and then a blue screen with blue and black vertical strips. Resetting SMC or PRAM does not help. I can boot into the single user mode and I can use target mode to copy all my data out. I also tried taking out one RAM and putting it in different slots. That does not help either. I ran apple hardware test several times, both standard and extended. No trouble found. I ran fsck -fy on command line to check the hard drive, no trouble either.

Yesterday, I reboot it from single user mode and successfully boot into the OS. Since then, I can always boot into the OS and do normal light duty work, like email browsing. The computer behaves normally and nothing is sluggish. I can hook up to a cinema display and do the same work. But if I put a slight load to CPU (like running a R code using about 1/8 of CPU) or GPU (like playing a youtube video in 720P), it will crash. Then I can boot it again when it is cooled down. I cannot boot it right after crash when it is hot. I looked at CPU and GPU temperature before crash, around 80 is the highest I see and the fan speed goes up correctly.

I saw a long thread in Apple forum talking about the crash in 2011 MBP but my case seems to be a little different. I am confused as why it works fine with all the programs but cannot handle any stress. Would it be a cooling system issue or would it be a failure in CPU and GPU? It is odd that both CPU and GPU fail at the same time. Or could it be a software issue? All the software and firmware is up to date.

 

[yjchua95]

I had a early 2011 Macbook Pro 15 inch with 2.0GHz CPU, 128G SSD and 8GB RAM. Everything apple stock. Two days ago, the computer suddenly crashed with the display pixelated and splited. I was running a statistic computation during the crash, which uses 12.5% CPU. After that, I cannot boot into the OS. The apple logo will appear and then stays at the grey screen forever. I tried recovery mode or safe mode or boot from a bootable USB, none of these are successful and I always get the apple logo and then a blue screen with blue and black vertical strips. Resetting SMC or PRAM does not help. I can boot into the single user mode and I can use target mode to copy all my data out. I also tried taking out one RAM and putting it in different slots. That does not help either. I ran apple hardware test several times, both standard and extended. No trouble found. I ran fsck -fy on command line to check the hard drive, no trouble either.

Yesterday, I reboot it from single user mode and successfully boot into the OS. Since then, I can always boot into the OS and do normal light duty work, like email browsing. The computer behaves normally and nothing is sluggish. I can hook up to a cinema display and do the same work. But if I put a slight load to CPU (like running a R code using about 1/8 of CPU) or GPU (like playing a youtube video in 720P), it will crash. Then I can boot it again when it is cooled down. I cannot boot it right after crash when it is hot. I looked at CPU and GPU temperature before crash, around 80 is the highest I see and the fan speed goes up correctly.

I saw a long thread in Apple forum talking about the crash in 2011 MBP but my case seems to be a little different. I am confused as why it works fine with all the programs but cannot handle any stress. Would it be a cooling system issue or would it be a failure in CPU and GPU? It is odd that both CPU and GPU fail at the same time. Or could it be a software issue? All the software and firmware is up to date.

It's the infamous Radeon GPU failure that's been striking thousands of early and late-2011 MBPs with Radeon GPU, due to a manufacturing flaw with the GPU in all the 15"/17" logic boards with the Radeon GPU. Mine was also hit by Radeongate.

 

[coldwaves]

It's the infamous Radeon GPU failure that's been striking thousands of early and late-2011 MBPs with Radeon GPU, due to a manufacturing flaw with the GPU in all the 15"/17" logic boards with the Radeon GPU. Mine was also hit by Radeongate.

I read the long thread on Apple forum and it does seem to be a very common issue for that year's model. What confuses me is that the computer will crash when only CPU is under slight load and GPU is totally free. If it is a GPU issue, I would not expect it to crash when GPU is not under any load.

But anyway, if it is not hard drive and not RAM, it is something on the logic board. In terms of repair cost, it will all be the same.

 

[TheIguana]

I read the long thread on Apple forum and it does seem to be a very common issue for that year's model. What confuses me is that the computer will crash when only CPU is under slight load and GPU is totally free. If it is a GPU issue, I would not expect it to crash when GPU is not under any load.

If you look at iFixit's teardown of the 2011 MBP, the GPU is upstream of the CPU on the heat pipe, when the CPU heats up some of the heat ultimately gets transferred to the GPU because of the thermal design. That could have something to do with why even when the GPU isn't under load that it could become unstable.

 

[coldwaves]

If you look at iFixit's teardown of the 2011 MBP, the GPU is upstream of the CPU on the heat pipe, when the CPU heats up some of the heat ultimately gets transferred to the GPU because of the thermal design. That could have something to do with why even when the GPU isn't under load that it could become unstable.

Image
Image

Thank you for your insight. I am thinking along the same lines. It seems to be related to temperature. The computer works totally fine and it just suddenly dies as long as there is a little bit load on GPU or CPU. Would you think it could be the thermal paste? I am not sure whether thermal paste can have such significant influence.

 

[TheIguana]

Thank you for your insight. I am thinking along the same lines. It seems to be related to temperature. The computer works totally fine and it just suddenly dies as long as there is a little bit load on GPU or CPU. Would you think it could be the thermal paste? I am not sure whether thermal paste can have such significant influence.

Doubtful it is the thermal paste, I had an early 2011 MBP that did the same thing. My best guess is it was caused by a materials failure on the chip or soldering, but I don't think anyone has dug into the science of why these chips seem to be failing. It has been reported as a known issue as coldwaves alluded to: https://www.macrumors.com/2014/01/17/2011-macbook-pro-gpu-glitches/

Ultimately, I had my Mac replaced by Apple with a Retina MBP under AppleCare because the issue kept reoccurring even after a few logic board replacements. Dunno if this is an option for you, if it is an out of warranty repair expect it to be >$500 for a new logic board.

 

[coldwaves]

Doubtful it is the thermal paste, I had an early 2011 MBP that did the same thing. My best guess is it was caused by a materials failure on the chip or soldering, but I don't think anyone has dug into the science of why these chips seem to be failing. It has been reported as a known issue as coldwaves alluded to: https://www.macrumors.com/2014/01/17/2011-macbook-pro-gpu-glitches/

Ultimately, I had my Mac replaced by Apple with a Retina MBP under AppleCare because the issue kept reoccurring even after a few logic board replacements. Dunno if this is an option for you, if it is an out of warranty repair expect it to be >$500 for a new logic board.

I agree with you. I doubt thermal paste can have such significant influence since it will crash only under very little load and the temperature before crash is not that high. I am out of warranty so I will probably just get a new computer.

Glad it works out for you. It seems to be a wide spread issue. Hope apple do something right to correct it.

 

[MagicBoy]

Mine went in a similar way a couple of weeks ago, just got it back.

Depending on where you live and the local consumer protection laws, it might be worth asking Apple nicely. I got the logic board replaced at no cost to myself.

The rest of the Mac is likely fine, it just fails to boot when it can't initialise the AMD GPU. Target disk mode ran fine on mine, so I could get the files I needed off the machine in the meantime.

 

[coldwaves]

Mine went in a similar way a couple of weeks ago, just got it back.

Depending on where you live and the local consumer protection laws, it might be worth asking Apple nicely. I got the logic board replaced at no cost to myself.

The rest of the Mac is likely fine, it just fails to boot when it can't initialise the AMD GPU. Target disk mode ran fine on mine, so I could get the files I needed off the machine in the meantime.

I can use target mode to get all data out as well. I went ot Apple store and the genius did hardware test and found nothing wrong. Then I demonstrate how to make it crash by playing a HD video. Then the computer cannot pass the cooling system test. The genius said it is something with the logic board and the way to repair is to replace it. He suggest me to mail it to apple for a flat rate repair, which will cost about 350 including shipping and tax. He also checked that there is no official replacement programs for this model yet.

I did more experiment on the laptop today. I put stress on CPU by running a statistical computation but did not use anything that requires the discrete graphic. The computer did computation for hours without any problem. I did the same thing with the laptop hook up to the cinema display. It will crash within 5 minutes. It seems like the heat from CPU makes GPU hot and GPU cannot cool down properly. If I just browse the web and check email with a display hooked up, the computer is fine for a long time. These three things lead to think that the problem is that GPU cannot cool down properly.

I don't know why GPU cannot cool down. The only thing I can think of to try is to reapply the thermal paste. Since I am out of warranty, I will properly just go ahead and try the thermal paste to see if that solves the issue. I doubt it will solve the issue but it is probably worth a try. I will report back what I find out.

 

[saturnotaku]

Since I am out of warranty, I will properly just go ahead and try the thermal paste to see if that solves the issue.

You have little to lose at this point, so that would be the most sensible thing to do at the moment. Please do report back when you've finished.

 

[triplelucky]

It would be interesting to see how the testing would go if you install gfxcardstatus and select only the integrated graphics card.

 

[coldwaves]

You have little to lose at this point, so that would be the most sensible thing to do at the moment. Please do report back when you've finished.

Got thermal paste reapplied this evening and it is as what I expect. No luck. Same problem still exist.

Apple does apply an excess amount of thermal paste. There is much more on the CPU than GPU on my laptop. Now, the computer can do anything fine if it does not use the discrete graphic card. Another obervation is that how long I can use it on discrete graphic card seems to depend on the temperature. If the computer has been running for a while and it is warm or hot, I can use discrete graphic for very short time or no time before it crashes. If I use it when the computer is cool, I can use it for a few minutes before it gives me the blue screen.

So in short, it seems to be a discrete card failure, not a thermal paste issue.

 

[coldwaves]

It would be interesting to see how the testing would go if you install gfxcardstatus and select only the integrated graphics card.

Just post an update here. I have installed gfxcardstatus. As long as I can boot into OS, I have no problem using it by using integrated graphic card only. It will crash once the discrete card is involved. The problem now is that it sometimes does not boot and just stay in the grey screen after the Apple logo. So I just keep the laptop on or in sleep without shutting it down.

I decide to get a new laptop and keep this one as it is. If this is such a wide spread issue in early 2011 models, hope Apple can address it.

 

[yjchua95]

If this is such a wide spread issue in early 2011 models, hope Apple can address it.

It's a very widespread issue, and Apple won't address it, because:

1. AMD doesn't manufacture the 6490M, 6750M and 6770M anymore, so Apple can't re-solder new GPUs onto the logic board.

2. That means Apple will have to give all of us a substantial discount off our next Mac, or maybe even give us brand-new rMBPs with 750M. Either way, both options are too costly for Apple. So Apple would rather screw over us 2011 MBP users.

 

[coldwaves]

It's a very widespread issue, and Apple won't address it, because:

1. AMD doesn't manufacture the 6490M, 6750M and 6770M anymore, so Apple can't re-solder new GPUs onto the logic board.

2. That means Apple will have to give all of us a substantial discount off our next Mac, or maybe even give us brand-new rMBPs with 750M. Either way, both options are too costly for Apple. So Apple would rather screw over us 2011 MBP users.

I agree with you and sadly this is what I think will most likely to happen.

 

[MagicBoy]

Any reason the 2012 revision board with the nVidia 650 won't slot straight in?

They should be able to still get the parts for those in quantity.

 

[yjchua95]

Any reason the 2012 revision board with the nVidia 650 won't slot straight in?

They should be able to still get the parts for those in quantity.

It does. My friend actually did that, so now he has a mid-2012 cMBP logic board (2.7GHz i7, 1GB GT650M and 16GB 1600MHz RAM) in his 2011 MBP's chassis.

 

[thundersteele]

I can use target mode to get all data out as well. I went ot Apple store and the genius did hardware test and found nothing wrong. Then I demonstrate how to make it crash by playing a HD video. Then the computer cannot pass the cooling system test. The genius said it is something with the logic board and the way to repair is to replace it. He suggest me to mail it to apple for a flat rate repair, which will cost about 350 including shipping and tax. He also checked that there is no official replacement programs for this model yet.

Just get the flat rate repair. It is very affordable compared to the full cost of a logic board replacement or compared to the price of the machine. Keep the documentation, if a replacement program is put in place later you can claim the money back.

 

[MagicBoy]

It does. My friend actually did that, so now he has a mid-2012 cMBP logic board (2.7GHz i7, 1GB GT650M and 16GB 1600MHz RAM) in his 2011 MBP's chassis.

Any idea what the cost is?

 

[coldwaves]

Just get the flat rate repair. It is very affordable compared to the full cost of a logic board replacement or compared to the price of the machine. Keep the documentation, if a replacement program is put in place later you can claim the money back.

Yes, that is the most cost effective way to repair. Apple quote me around 350 to get it fixed. But, if the hardware itself is indeed problematic, the replaced new board will most likely suffer the same issue. They have a 90 days warranty for the repair though.

----------

You have little to lose at this point, so that would be the most sensible thing to do at the moment. Please do report back when you've finished.

Another thing I notice is that the computer runs much cooler now. The fan keep at the basic speed around 2000rpm for most of the time. Before, the fan will spin very fast even just watching a youtube video. Now doing the same thing does not increase the fan speed at all.

I am not sure whether this is the effect of thermal paste or it is because I totally disable the discrete card so no heat is generated from the GPU. One thing I did is that when I run statistic computation, it is much cooler than before. No GPU is used in running such computation so my sense is that reapplying the thermal paste does help a lot in cooling it down.