Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Mac Pro (Late 2013) GPU (Driver) Issues

What do you think is the source of the graphics issues on Mac Pro (Late 2013) ?

  • Hardware

    Votes: 69 52.3%

  • Software

    Votes: 30 22.7%

  • Hardware & Software

    Votes: 32 24.2%

  • Other

    Votes: 1 0.8%
  • Total voters
    132
Well, I can answer my question. First and complete restart of computer after crash. While watching a youtube video. Log shows it is driver crash related to window server. HD4000 active in MBP15 mid 2012(non-retina). macOS Sierra PB2.

Any thoughts, ideas?
 
Hello,

Here is my story, maybe the information helps for someone...

A few months ago I experienced the 'GPU driver appears to be hung' problem almost every day. So, I brought it back to the store. Two weeks later I went back to pick the mac up, they said the D300 GPUs were replaced. That evening, I think it took about 5 minutes until the spinning beachball was there again. So I brought the mac again to the store and said that the issue was still there. Three weeks later (I think June 25), I received the mac again but this time the employee said he didn't receive information about what the repair center had done :/ Back at home I had to configure OS X again so I assume they just reinstalled OS X this time. Surprisingly, the issue didn't came back until yesterday. And it could be coincidence, but yesterday I installed El Capitan 10.11.6.

At this moment I think that the graphic issues are related to the OS version (in my case 10.11.5 was a good one then). Any thoughts on this?
 
jorenjm said:
Hello,

Here is my story, maybe the information helps for someone...

A few months ago I experienced the 'GPU driver appears to be hung' problem almost every day. So, I brought it back to the store. Two weeks later I went back to pick the mac up, they said the D300 GPUs were replaced. That evening, I think it took about 5 minutes until the spinning beachball was there again. So I brought the mac again to the store and said that the issue was still there. Three weeks later (I think June 25), I received the mac again but this time the employee said he didn't receive information about what the repair center had done :/ Back at home I had to configure OS X again so I assume they just reinstalled OS X this time. Surprisingly, the issue didn't came back until yesterday. And it could be coincidence, but yesterday I installed El Capitan 10.11.6.

At this moment I think that the graphic issues are related to the OS version (in my case 10.11.5 was a good one then). Any thoughts on this?
Click to expand...

I recently bought a Mac Pro 2013 and sent the first one back. I purchased refurbish to save on costs. The second one seemed to work better, but noticed a few graphic issues when using FCPX. Called tech support and they said they would call me back...never did. But while I was waiting 10.11.6 came out and to my surprise it fixed my graphics issues.

Here are some suggestions:

1. Find out the date of when your system was assembled (mine fortunately was April 2016-or at least repaired). Look on Internet for small program to collect info on your system.

2.make a USB OS X 10.11.6 install drive to use to install. Download the latest version from App Store.

3. Reinstall OS X 10.11.6 from scratch, boot into repair mode before, wipe disk clean first.

4. Install OS X from USB

If problem still exists, then download the 10.11.6 combo update and apply. Don't use the auto App Store update.

5. If problem continues, apply the combo update again (sounds funny, but try).

This fixed my issues and my Mac Pro is happy. But runs hot (expected and is normal unfortunately)

If your Mac Pro was assembled in early 2015, then get the graphics card replaced. Look up the recall on the Mac Pro 2013 for more info.

I think the apple tech did not call me back because he knew that the update would fix my issues that was coming out in a few days. He said that he would do some research on the issues I told and call back. Of course he could not tell me that the update was coming in two days and probably waited for me to call back if issues after apply update continued.

These steps helped my issues and hope it helps.
 
Since updating to 10.11.6 I have had a couple of issues.

1. Some odd looking rectangle of texture in an iTunes window. Appeared to be part of the desktop texture being displayed within iTunes. This oddity disappeared when the mouse pointer was moved to certain places on the screen.

2. I just had a lockup. The first thing I noticed going was my wireless connection. Then the beachball. I haven't had much in the way of lockups on this replacement machine. And this seems a little bit different to the ones I've had before in that I lost network connectivity. Which is kind of awkward, because I couldn't ssh in to reboot the machine.

I absolutely love this machine when it's working, which is most of the time. But it's difficult to have complete confidence when at any moment the thing may freeze up and need to be powered off manually. I really hope this is addressed, because were it not for these serious issues, it would be the best Mac I've ever used.
 
Recently I had exactly the same issues with 2010 cMP, random beachball freezes and the reason was faulty PCIe SATA 3 controller. I replaced it and the problem disappeared.
 
  • Like
Reactions: Infrared
Bluetooth not working after booting up today. Had to reboot via ssh to get it back.

That's not happened before. 10.11.6.

[doublepost=1470330504][/doublepost]The 10.11.6 combo has an EFI files directory containing these files:

Screen Shot 2016-08-04 at 17.57.13.png


Mac Pro identifier is bottom right.
 
Last edited:
Just a note that the 10.11.6 update seems to have caused a return of this issue in our recently built 8 core dual D700 32GB RAM MacPro. I have had FCPX crashes during stabilization of all 4 projects run since the update. Seems to crash toward the end of the last clip in the project. Two seem to be a dispatchqueue thread, one seemed to be a dominantmotiontrackerworker thread, and the other seems to be a GPU reset by the kernel. It had been running fine under 10.11.5, I think.

EDIT:
As an update, someone suggested moving the FCPX library to another drive, so I moved it to my second SSD (Same model Samsung pro) and have not seen a crash in three similar projects. I probably just jinxed me :)

EDIT 2:

Yeah a jinx, the last four projects failed along the way, after 6 good ones.
 
Last edited:
Infrared said:
What's the most straightforward way of reverting to 10.11.5?

Thanks.
Click to expand...

My experience is that El Capitan is just rubbish. cMP problems, Retina 13" Mid 2014 problems...
For me, most modern but stable and usable OS X is Yosemite 10.10.5. I will try Sierra but I'll skip El Cap.

Most straightforward way is to download fresh copy, create bootable USB and erase boot disk during setup.

Creating Bootable USB command:
sudo /Applications/Install\ OS\ X\ Yosemite.app/Contents/Resources/createinstallmedia --volume /Volumes/USB_Disk_Name --applicationpath /Applications/Install\ OS\ X\ Yosemite.app --nointeraction
 
  • Like
Reactions: filmak
loby said:
I recently bought a Mac Pro 2013 and sent the first one back. I purchased refurbish to save on costs. The second one seemed to work better, but noticed a few graphic issues when using FCPX. Called tech support and they said they would call me back...never did. But while I was waiting 10.11.6 came out and to my surprise it fixed my graphics issues.

Here are some suggestions:

1. Find out the date of when your system was assembled (mine fortunately was April 2016-or at least repaired). Look on Internet for small program to collect info on your system.

2.make a USB OS X 10.11.6 install drive to use to install. Download the latest version from App Store.

3. Reinstall OS X 10.11.6 from scratch, boot into repair mode before, wipe disk clean first.

4. Install OS X from USB

If problem still exists, then download the 10.11.6 combo update and apply. Don't use the auto App Store update.

5. If problem continues, apply the combo update again (sounds funny, but try).

This fixed my issues and my Mac Pro is happy. But runs hot (expected and is normal unfortunately)

If your Mac Pro was assembled in early 2015, then get the graphics card replaced. Look up the recall on the Mac Pro 2013 for more info.

I think the apple tech did not call me back because he knew that the update would fix my issues that was coming out in a few days. He said that he would do some research on the issues I told and call back. Of course he could not tell me that the update was coming in two days and probably waited for me to call back if issues after apply update continued.

These steps helped my issues and hope it helps.
Click to expand...

Started to notice problems again. Screen went black and had to push the start button to turn off the system and turn it on again. 10.11.6 seems to run better, but not all of the issues are fix.. :(
[doublepost=1471093784][/doublepost]
loby said:
Started to notice problems again. Screen went black and had to push the start button to turn off the system and turn it on again. 10.11.6 seems to run better, but not all of the issues are fix.. :(
Click to expand...
What runs better on the Mac Pro 2013, Mavericks, Yosemite or El Capitan?
 
loby said:
Started to notice problems again. Screen went black and had to push the start button to turn off the system and turn it on again. 10.11.6 seems to run better, but not all of the issues are fix.. :(
[doublepost=1471093784][/doublepost]
What runs better on the Mac Pro 2013, Mavericks, Yosemite or El Capitan?
Click to expand...
I haven't tried Yosemite, but El Capitan made my mac pro 2013 freeze so bad. With Mavericks I had no freeze at all.
[doublepost=1471171565][/doublepost]Don't you guys think that one should keep his mac with its original OSX when possible? I have a macbook 2009 with Snow Leopard and it's still fast, 0 issues. I remember when I tried to upgrade it to Lion: wifi started to drop and fans were noisy and out of control. Reverted back to SL and still fine today.
 
I know of Mac Pros that run all three OSXs that were released since the nMP launched without any big issues. However they returned their nMP up to 6-7 times until they worked fine.
 
merlinmage said:
However they returned their nMP up to 6-7 times until they worked fine.
Click to expand...

That confirms my strategy. As its main purpose is not "being away for repair" I found my way to live with the little odds that remain for now. Using 10.11.6 I have a quite stable system, that sometimes locks up, needs the TB-monitor to be un- and replugged - save work - reboot - carry on... In almost all the cases it locks up once and then runs a whole day without any problem. And I am doing "heavy" work in Cinema 4D, Logic and Final Cut Pro.
 
Here's my update for my Mac Pro with D300

1st servicing: Replaced graphics card
2nd servicing: Replace CPU board and graphics card
3rd servicing: Apple had someone to come down to take my Mac Pro. It's the same issue: screen freezes, music still plays, you can unplug and replug the monitor get control and restart. I've tried OS 10.10.5 and also 10.11.6 and both still exhibit the freezes randomly. It's so difficult to replicate the problem. Some days there will be several freezes in short intervals, and then it can go for a few days without freezes. There's something wrong with Mac OS and the the Mac Pro for sure. So many people are having this mysterious problem! And all the while my Mac Mini is running perfectly with the same OS.
-_-
 
  • Like
Reactions: Infrared
It's surprising to me that this is still going on after a couple of years. I really hope Apple can resolve this soon for once and all.
 
I'm experimenting with Firefox instead of Safari. There are a lot of reports about of other macs freezing and that has been linked to Safari. I guess it can't hurt to try Firefox for a while, although I have to say I'm not especially keen on it.

This is just blind guessing, really, and I'm categorically not attributing these faults to Safari or any other potential culprit until we know more.
 
  • Like
Reactions: Michaelgtrusa and grahamperrin
Pardo83 said:
I haven't tried Yosemite, but El Capitan made my mac pro 2013 freeze so bad. With Mavericks I had no freeze at all.
[doublepost=1471171565][/doublepost]Don't you guys think that one should keep his mac with its original OSX when possible? I have a macbook 2009 with Snow Leopard and it's still fast, 0 issues. I remember when I tried to upgrade it to Lion: wifi started to drop and fans were noisy and out of control. Reverted back to SL and still fine today.
Click to expand...

Problem is apple makes you update if you want to use the latest version of logic and FCPX which the Mac Pro 2013 was made for.

Interesting that I am not having too many issues with 10.11.6 except now my screen going black at times with my thunderbolt monitor all of a sudden and will not go back. Have to unplug and plug it out of the thunderbolt port to again to resolve it and get the screen to turn on. Has never happened to me except recently.
[doublepost=1471781781][/doublepost]
axantas said:
That confirms my strategy. As its main purpose is not "being away for repair" I found my way to live with the little odds that remain for now. Using 10.11.6 I have a quite stable system, that sometimes locks up, needs the TB-monitor to be un- and replugged - save work - reboot - carry on... In almost all the cases it locks up once and then runs a whole day without any problem. And I am doing "heavy" work in Cinema 4D, Logic and Final Cut Pro.
Click to expand...
Having the same issues. I hope apple reads these posts.
 
Incidentally, on the couple of occasions when the system froze and I logged in via ssh, I noticed that spindump was running. There are logs visible in the console app that appear to relate to this (e.g., stack snapshots).

It's hard to know what is important or might be related to this particular issue. I did notice this though:

Code: 
Heaviest stack for the main thread of the target process:
  10  start + 1 (libdyld.dylib + 13741) [0x7fff865cd5ad]
  10  ??? (WindowServer + 3966) [0x104ef7f7e]
  10  CGXServer + 7273 (CoreGraphics + 2680753) [0x7fff8725c7b1]
  10  CGXRunOneServicesPass + 1075 (CoreGraphics + 2672380) [0x7fff8725a6fc]
  10  mach_msg_trap + 10 (libsystem_kernel.dylib + 69490) [0x7fff88d44f72]
*10  ipc_mqueue_receive_continue + 0 (kernel + 812688) [0xffffff80002c6690]


Process:  WindowServer [170]
Path:  /System/Library/Frameworks/ApplicationServices.framework/Frameworks/CoreGraphics.framework/Resources/WindowServer
Architecture:  x86_64
Parent:  launchd [1]
UID:  88
Task size:  18932 pages
CPU Time:  0.027s

  Thread 0x46e  DispatchQueue 1  10 samples (1-10)  priority 79 (base 79)  cpu time 0.027s
  <IO policy important>
  10  start + 1 (libdyld.dylib + 13741) [0x7fff865cd5ad] 1-10
  10  ??? (WindowServer + 3966) [0x104ef7f7e] 1-10
  10  CGXServer + 7273 (CoreGraphics + 2680753) [0x7fff8725c7b1] 1-10
  10  CGXRunOneServicesPass + 1075 (CoreGraphics + 2672380) [0x7fff8725a6fc] 1-10
  10  mach_msg_trap + 10 (libsystem_kernel.dylib + 69490) [0x7fff88d44f72] 1-10
  *1  ipc_mqueue_receive_continue + 0 (kernel + 812688) [0xffffff80002c6690] 1
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>
  <os_trace: Activity 0x37 CGSLogMessage + 171 (CoreGraphics + 1021902)>

And there are loads of those CGSLogMessage instances.

Also found this:

Code: 
19/08/2016 19:03:11.175 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.180 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.184 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.189 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.193 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.198 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.202 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.208 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.212 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.217 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.222 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.227 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.231 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.236 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.240 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.245 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.249 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.254 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.259 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.263 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.268 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
19/08/2016 19:03:11.272 WindowServer[170]: GPU Driver appears to be hung (over 5 continuous seconds of unreadiness)
:
 
Last edited:
Since the 7,1 nMP is around the corner I wouldn't assume Apple intends to fix it. The Pro Apple train left the building a long time ago.
 
My opinion:
The problem, what I see here is, that all the comments here cannot help anyone nor Apple. Maybe there are 100
people out there having problems, thats all, I cannot asses this by reading this thread.
What I miss is a list of everyone, something like below, of his setup better in a new thread.
And that everyone posts only once to check the quantity of Mac Pro problems:

..... My setup .....
-Problem and how often:
-Mac Pro date of purchase:
-Mac Pro was in service y/n, once, twice...:
-GPU was changed y/n:
-Other components changed:
-System version installed (list if serveral):
-System clean fresh install, SSD initialized y/n:
-System software clean without 3rd party drivers y/n:
-Monitor type and connector type:
-Any USB, HDMI or Thunderbolt hardware attached aside of monitor:
-Test with powered USB hub y/n:
-Test result with no hardware attached:
-RAM original or changed, type and capacity:
-Wifi is on, is off:
-What software crashing or beachball:
-User action at time of crash:
-Fan speed at time of crash:
-Other points of interest:

I personally think, that Apple will not do much developement to resolve problems
other than changing gpu and system reinstall.
Like the Cinema black Display problem or Thunderbolt display problem.

First of all, try to encircle the problem. All this comments here are not really helpful.

When nothing helps, in the US you have a big advantage over much other countries:
Do a class action !
Do it before the Mac Pro will be replaced or production stops.

The Mac Pro is the most expensive Apple product, not a 200 bucks fone for your kids,
and often your daily work depends on it.
 
Last edited:
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back