MacBook Pro 2011 17" hard freeze

I have a very interesting development....when you run the test in integrated mode it's almost unreal as to how fast the "integrated" graphics performs compared to the discrete. I looked @ the numbers and both are coming in almost as quick if not quicker on the integrated side! I followed up with a test in dynamic mode and sure enough it crashed. I'm now thinking that we do in fact have a driver issue as no way should the integrated gfx be coming in anywhere near the discrete solutions. Recommendation....keep your machine in integrated mode while apple sorts this out.

Update... Even more evidence of a driver issue. I can no longer run the test with integrated enabled so I'm guessing there must be a glitch somewhere that enabled me 2 do it before.

Message was edited by: Chullo

@Chullo - cheers.

I have never sen the reported fan speed shown as higher than 5500 in any monitoring utility regardless of how much stress I put on the system.

In fact just run some more stress tests on both internal and descrete GPU and in nether case did it exceed 5500 or so.

Wierd ???

Chullo - BTW - I am finding the opposite for the applications I am using. Ableton Live running 64 tracks is much more efficient when using descrete rather than integrated which suprised me. I didnt expect any difference.

Message was edited by: Khazul

One way to test the thermal question is to get two computers: one that exhibits the problem and one that does not. Set them side by side idling (just the OS). If the heat wasn't transferring fast enough from the problem computer, you would expect it to have a consistently higher temperature (mean each +/- the noise over 10 minutes or so), even under a low process load. You'd want them side by side so you had the same ambient conditions.

If you can talk that guy with three macs into getting a fourth, then he could do it --- what's another three grand on the amex 🙂. Or perhaps someone could go to an apple store and find a working machine with the same configuration as their broken one and try it out.

Checking in - same problem with a CTO 15" 2.2GHz Hi Res Screen and 7200rpm 500GB drive. Ordered on opening day of sale and arrived 4th March-ish.

Consistently locks up in WoW or doing a Handbrake encode using discrete or dynamic GPU modes. Doesn't lockup in integrated. No truly high temps, tends to crash between 65-75 degrees on the AMD card.

Spoke to Apple Support in Sydney from about the 7th March (so they've known about it in Sydney since then and they're replacing it (waiting for a courier to actually turn up to pick up the original).

Not holding out a **** of a lot of hope that a new one won't show the same issues...looking less and less like a bad batch or an isolated issue.

John, I've tried this on ALL the new machines they had @ my local apple store and they all failed. Much to the surprise of the genius "helping" me. This is a systemic issue and I'm still pretty surprised that Adrian's machine doesn't get killed by it. So....there is hope.

Chullo - I have it forced into Integrated mode - i showing in menubar... But I'm unable to launch the test even after shutting down SmallLux and even restarting the machine in I mode.

Thoughts? I have to be doing something super-simple-stupid wrong.

This machine will lock up quite happily on its own without some kind of fan speed assistance - usually just takes a few seconds from hitting play in an audio editing application like logic or ableton live when I have alot of tracks and processing in the project.

Just tried something called fan control and finally Im seeing it go upto 5900 with core temps a couple of degrees lower - so some improvement.

As for powerful machines, this is my first mac, my normal audio workstation is a windows 7/64 that runs all its cores permenantly at 4Ghz (ie no turbo boost, or thottling just BIOS locked at 4GHz) - its an abused i7/920, but with some excellent cooling that mean under under that abuse, core temps emina quite low. It also 1600Mhz memory etc - so in theory shold be alot faster than the MBP for what I do.

The big suprise however - when I can get the MBP under control cooling wise, is seems to be able to match the performance of my PC running ableton live. Granted - it doesnt stress the machine as hard as your 3d rendering when Im editing, but only because I need to leave some headroom for it to manage real time low latency multi-channel audio with no drop outs etc. The only real difference is UI responsiveness remains noticeably better on the PC under extreme audio processing loads. This is also another difference I notice between the intel and radeon GPU in the MBP - UI of the audio application I use are far more responsive when the Radeon is active and the machine is under heavy load.

The PC at this time is not running optimally, when first installed I guess it was about 25% faster for audio work than it is now.

Anyway - just wish the MBP was stable - else I cant use it for working on people's music when they are there - dont look good having a machine stiff on you, never mind risk to precious recordings when you have to hard reboot it etc.

I guess we each have own own performance tests according to what we do 🙂

Well generally a good idea to be polite lol, usually the people who get in touch with can't do anything about it anyways....unless u were bumped up or something? It's a tough issue and I personally know 3 people having this issue, plus random machines @ the apple store having the same issue and a genius suggesting that given that statistic my replacement may have the same issue! Still hoping it's fixable or something like that bad batch of thermal paste 😉 because otherwise it's gonna b a serious issue. I can't even video skype with this for longer than 10mins. I use VMWare while watching my CPU temp so as not to lose any valuable information. Apple has usually been on their game with this stuff, but they certainly dropped the ball on this one.

Jkirker, naw it's not you. It won't allow me 2 do it anymore either and I'm not sure how to reproduce whatever glitch that allowed me and one other guy on the forum to do it. Very upset that I restarted my machine now though, cuz whatever it was did the trick 😟.

Just for giggles, I ran smallluxGPU again for another 1000 seconds on my new MBP and again I didn't encounter any lockups. Interesting. This is a stretch, but perhaps there's a piece of software installed on my system that's preventing the lockup?

Solution.....run the smallluxgpu in discrete only, press esc to quit that instance. Switch to integrated mode (w/o closing the app). You can then rerun the test in that mode w/o a hang and having a stable temp, in my case the temp is even dropping! I'm going to operate the machine like this for the next couple hours and see if it hangs. There is now no doubt that there is also a driver problem here.

Message was edited by: Chullo

OK, I got it to run in Integrated Only mode by...

1) Put it in Dynamic Switching mode
2) Launch SmallLuxGPU
3) Start a test 1024/768 8+GPU
4) Throw it into Integrated only mode while test is running
5) Immediately cancel test
6) Restart test

I'm not sure if it's truly in integrated only mode because the app was already launched - however so far, after about 4 minutes of watching and hammering out this message I haven't frozen up yet and my temps are hovering at 95/96/97C and below.

It's interesting... I'm going to hit post before this thing locks up and let it run...

Lol, I guess I beat u to the post 😉. It's definitely still using the discrete, but it appears that the integrated driver is a lot more stable than the discrete one as by now, on my machine, I would be guaranteed a hang! So what I plan to do is quit the test but leave the app open and see if I get any hangs under this "new GPU setting".

Update....no need to leave the app open as it won't allow you to rerun the test w/o doing the same thing over again.

I was bumped up with Applecare, and then I sent my computer in to get it "fixed". I realize that it's likely I'll get a new computer with the same problem, but until they hit a certain number, they wont realize there is something wrong.

After my computer got to the repair place, I got an email from some lady named Liz saying that my computer qualified for something called "Apple's Capture Program". This basically meant they wanted to keep my machine and send me a new one. That way they could study my machine and understand what's wrong with it. Since I had her email address, I decided I'd send her relevant information about the problem I experienced.

It seems the base assumption from their support staff is that customers are paranoid about systemic problems. As such they tend to be dismissive of any notions coming from their customers. My thoughts are: Only overwhelming evidence will dissuade them of this notion. And getting over this notion is the first step in fixing the problem. So until someone from Apple says to me: Yes we acknowledge there is a problem, it's widespread, and we're working to fix it. Then I'm going to keep on providing them with evidence.

So when my new computer comes in next week, the first thing I'm going to do is run that test with darwin ports. If it works, then great. If that crashes it, I'm going to call Liz and John (the Apple care guy helping me out) and say: How statistically likely do you think it is that I would get two computers with the exact same anomalous problem?

BTW My money is on a driver/firmware issue 🙂

JohnH - my rep kind of "acknowledged" an issue on the phone however she said that engineering didn't disclose said issue. My primary machine is going directly to the engineers as well as soon as the replacement arrives. I'm not sure if they want these two that I'll be taking back to the apple store... 😉