Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

2013 MacPro - GPU hang failure to restart GPU

I purchased a 2013 MacPro and I was having frequent GPU Hangs. The symptom is the screen would be frozen but I could login from remote via ssh (not screen sharing) and see the system.log was spewing about not being able to restart GPU, and that the GPU had hung.


I went back and forth with applecare over it and after some failed debugging attempts including re-installing I returned the mac pro and they sent me a new one.

The original macpro would crash every day or so. I've had the new one a few weeks and it just crashed today with the same message for the first time.


I never really got to the bottom of which it was happening with AppleCare. Has anyone ever seen this be a software problem? I am using the DualLink DVI adapter which is not what I would call a quality product, but I'm not sure I see how it could be causing this.


I'm going to call AppleCare again about it but I wanted to see if anyone else had faced similar issue. I haven't seen much about it with searches.



Here is snippit from log.


May 19 01:06:10 habanero kernel[0]: [6:0:0] GPU Hang State = 0x0000002e

May 19 01:06:10 habanero kernel[0]: ** GPU ASIC Log Start **

May 19 01:06:10 habanero kernel[0]: 0x00006810

May 19 01:06:10 habanero kernel[0]: 0x000000e3

.. a lot of spew

May 19 01:06:16 habanero kernel[0]: ** GPU ASIC Log End **

May 19 01:06:16 habanero kernel[0]: : ** GPU Debug Info End **

May 19 01:06:16 habanero kernel[0]: GPURestartReportEnd

May 19 01:06:16 habanero kernel[0]: Trying restart GPU ...

May 19 01:06:16 habanero kernel[0]: [6:0:0] Failed to reset hung GPU! result=0x00000097 ulResetBlockBitmap=0x00008041 ulResetResultBitmap=0x00008000

May 19 01:06:17 habanero kernel[0]: [6:0:0] GPU Hang State = 0x00000000

May 19 01:06:17 habanero kernel[0]: GPU hang:

May 19 01:06:17 habanero kernel[0]: Trying restart GPU ...

May 19 01:06:19 habanero com.apple.launchd[1] (com.apple.DumpGPURestart): Throttling respawn: Will start in 8 seconds

May 19 01:06:27 habanero com.apple.launchd[1] (com.apple.DumpGPURestart): Throttling respawn: Will start in 10 seconds

.. goes on until I forced powered it down

Mac Pro, OS X Mavericks (10.9.2), 2013 model

Posted on May 19, 2014 8:26 AM

Reply
34 replies

Jun 2, 2014 2:39 PM in response to Roger Jones

It hasn't happened again since I reported it. On my original MacPro 2013 it was happening about once a day.I haven't had time to call apple again on it. If it happens again I will.


When I got the replacement macpro I went ahead and did all the things they asked me to do before which weren't likely to work, but wanted to rule it out. Namely I wiped and reinsatlled from scratch as well as ran the diagnostics boot mode on it. Even the old defective one passed all those tests though.

Jun 5, 2014 6:33 PM in response to Bradford W. Miller

It happened to me again with 10.9.3. But this time it was a bit different It was able to restart the GPU, but it did it many many times. The screen was pretty much freaking out while this was going on. Freezing, being 'digital staticy' if that makes any sense,a nd then ultimately going completely black while leaving me with a mouse pointe r that moved around.


I escalated to Applecare and was able to convince them to let me send my logs. When this error happens the logs are pretty verbose. I am hopeful someone with the right tools can pinpoint what the problem is.


Love my new mac pro but this sure is annoying.

Jun 11, 2014 4:50 PM in response to realpdm

I'm having a similar... or identical issue.

My main screen goes a little twitchy occasionally. The computer appears to go to sleep, with both displays off, and then shuts itself down after a while.


2014-06-11 15:18:39.000 kernel[0]: : ** GPU Debug Info End **

2014-06-11 15:18:39.000 kernel[0]: ------------------------

2014-06-11 15:18:39.000 kernel[0]: GPURestartReportEnd

2014-06-11 15:18:39.000 kernel[0]: Trying restart GPU ...

2014-06-11 15:18:39.694 com.apple.launchd[1]: (com.apple.DumpGPURestart) Throttling respawn: Will start in 4 seconds

2014-06-11 15:18:39.000 kernel[0]: [6:0:0] Failed to reset hung GPU! result=151 ulResetBlockBitmap=0x00008000 ulResetResultBitmap=0x00008000

2014-06-11 15:18:39.000 kernel[0]: vtd[2] fault: device 6:0:0 reason 0x6 R:0xffffff1000

2014-06-11 15:18:39.000 kernel[0]: vtd[3] fault: device 6:0:0 reason 0x6 R:0xffffff1000

2014-06-11 15:18:42.000 kernel[0]: [6:0:0] GPU HangState 0x0000000e, HangFlags 0x00000007: IndividualEngineHang 1, NonEngineBlockHang 1, FenceNotRetired 1, PerEngineReset 0, FullAsicReset 1

2014-06-11 15:18:44.000 kernel[0]: ** GPU ASIC Log Start **

2014-06-11 15:18:44.000 kernel[0]: : 0x00006798 : 0x000000e3 : 0x00000001 : 0x00000018 : 0x0000a060 : 0x00000001 : 0x00000001 : 0x00006798

... repeat ...


Mac Pro:

Model Identifier: MacPro6,1

Processor Name: Quad-Core Intel Xeon E5

Processor Speed: 3.7 GHz

Number of Processors: 1

Total Number of Cores: 4

L2 Cache (per Core): 256 KB

L3 Cache: 10 MB

Memory: 12 GB

Boot ROM Version: MP61.0116.B05

SMC Version (system): 2.20f18

Illumination Version: 1.4a6

OS X 10.9.3 (13D65)

AMD FirePro D700

Stock RAM

Jun 11, 2014 5:00 PM in response to thisma

thisma, you may want to call AppleCare. My next step with my case is to bring it to an applestore and let them capture some debugging information on the HW. They apparently will send some special thing to the store for some specific isolation of the problem. I did get them to analyze my logo files with all the asis debug info in there.


I've had it happen on two different macpros. I hope they can resolve it with a firmware update. Or maybe it is as simple as something coming lose in shipping. It is an entirely new product and assembly logistics. It seems either are just as likely.

Jul 6, 2014 8:57 PM in response to realpdm

Hi there realpdm - have you resolved this? I'm having the same exact issue as well with a new Mac Pro (6.1 late 2013). I'm currently running OSX 10.9.3. It was having the same issue running on 10.9.2 though. I've taken the machine into the apple store for a genius appt. They kept the machine for 48 hours and were not able to reproduce the problem so they told me there was nothing wrong with my machine. However, it's still freezing daily with every reboot. It's arbitrary and I cannot predictably reproduce the problem. And that makes things very difficult.


I am getting the same "GPU hang state" message in Console consistently with every freeze though. Not sure if it would be beneficial to paste those logs here but it's getting extremely frustrating sitting on a machine that is essentially unusable with constant daily freezes.


At first I thought maybe bc the problem was GPU-related, it could be the Mini DisplayPort to DVI adapter I was using for one of my two connected monitors. I was advised that bc I was using a 3rd party non-apple adapter I should instead use the Apple branded adapter for this since it was supposedly tested and is supported. It didn't fix the issue.


I've got a follow up call into Apple Care. They haven't called me back.


Would love love to know if anyone successfully made it to the other end of this same issue??

Jul 6, 2014 9:15 PM in response to Roger Jones

Interesting, Roger Jones. Thanks for the quick reply. Curious, how long have you been running 10.9.4? It's only been out a few days, no? Were you having the GPU-related freeze-ups described here on a daily basis too? Just trying to get a sense of how much of a noticeable difference updating to 10.9.4 made for you with this problem.


I use my machine for audio recording and editing and the latest version of Avid Pro Tools which I use is only supported for use up through OSX 10.9.3 at the moment. But if 10.9.4 truly fixes the potential bug, I'd consider making the jump in order to have a better working machine!

Oct 7, 2014 11:39 AM in response to realpdm

Hi Roger, I wanted to check back in on this thread with you. Have your GPU hangs stopped? Are you still running 10.9.4. I took my Mac Pro (6.1 trash can) back into Apple a while back and after much stress testing they were finally able to see what I was talking about and as a result replaced both graphics cards in my machine. It seemed to help for a little bit.. but then.... again.... freeze ups started happening again. I am REALLY frustrated. Again, hard to predictably reproduce them... but I am now running 10.9.4 and they are still happening. Still getting similar "GPU Hang" messages in Console preceded pretty consistently by a "Sound assertion" message. I understand this has nothing to do with a sound error though. Not sure what to make of all this. Not sure what other steps I can try... perhaps other than a full wipe of my machine and reinstall everything from a Time Machine backup? Or maybe install everything from scratch? I suppose I can take it back into the Apple Store... but it will be my 3rd trip and not feeling very confident that they'll be able to resolve it. This is starting to get ridiculous.

Nov 20, 2014 10:55 AM in response to Fleadomexp

Fleadomexp,


I'm having the same issues... it's killing me! Any resolution?


I'm also running Avid Pro Tools 11.2.1 (HDX) on 10.9.4. Sometimes I'll go a few days with no crash, then I might have several within an hour (like this morning... with clients in the room). Strangely, we have an identical system (aside from a Sync HD, different controller, and different screens) that doesn't seem to have this problem. Just hoping someone can shed some light on this.


Thanks.

2013 MacPro - GPU hang failure to restart GPU

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple ID.