Mac Pro 5,1 Kernel Panics very frequently (Nvidia related?)

[leon771]

Good morning,

I have a problem with my Mac Pro 5,1 Kernel Panicking very frequently (sometimes every 2 minutes), then stable for a week with no issue.

This morning I was filling in a form on the internet and the Mac Pro shutdown then restarted.

This is the info that it provided.
From what I can tell, it would appear to be related to my graphics card.

Spoiler: Panic

Anonymous UUID: E66ABFD5-8BD0-2D08-CFE7-C8502D1CA3F9

Mon Nov 19 06:35:00 2018

*** Panic Report ***

panic(cpu 0 caller 0xffffff7fa089bf1a): NVRM[0/5:0:0]: Read Error 0x006100c0: CFG 0xffffffff 0xffffffff 0xffffffff, BAR0 0x8a000000 0xffffffa75bfe8000 0x0e4000a2, D0, P3/4

Backtrace (CPU 0), Frame : Return Address

0xffffffa75ab739b0 : 0xffffff801fc6b1c6

0xffffffa75ab73a00 : 0xffffff801fd93a74

0xffffffa75ab73a40 : 0xffffff801fd85d44

0xffffffa75ab73ab0 : 0xffffff801fc1d1e0

0xffffffa75ab73ad0 : 0xffffff801fc6ac3c

0xffffffa75ab73c00 : 0xffffff801fc6a9fc

0xffffffa75ab73c60 : 0xffffff7fa089bf1a

0xffffffa75ab73ce0 : 0xffffff7fa0969e19

0xffffffa75ab73d40 : 0xffffff7fa230d2fb

0xffffffa75ab73d50 : 0xffffff7fa090917b

0xffffffa75ab73d70 : 0xffffff7fa09faf2e

0xffffffa75ab73dd0 : 0xffffff7fa0870d30

0xffffffa75ab73e90 : 0xffffff7fa0836aaa

0xffffffa75ab73ed0 : 0xffffff801fca4634

0xffffffa75ab73f40 : 0xffffff801fca4195

0xffffffa75ab73fa0 : 0xffffff801fc1c557

Kernel Extensions in backtrace:

com.apple.nvidia.driver.NVDAResman(10.3.2)[6DE4BC26-EEC1-379B-AD41-532FB0A529CD]@0xffffff7fa082f000->0xffffff7fa0b08fff

dependency: com.apple.iokit.IOPCIFamily(2.9)[D02EA6D7-7A0E-3F9D-B179-75AC2228E449]@0xffffff7fa0494000

dependency: com.apple.iokit.IONDRVSupport(519.21)[8E0EE518-0D53-38C0-B308-2C6CD9EC5B7B]@0xffffff7fa0813000

dependency: com.apple.iokit.IOGraphicsFamily(519.21)[E5B53C5D-BF6E-3D72-9091-A80DFE5426C9]@0xffffff7fa07bd000

dependency: com.apple.AppleGraphicsDeviceControl(3.20.14)[D3CF5F89-2DA6-3F73-995D-F5DDE191E192]@0xffffff7fa0823000

com.apple.nvidia.driver.NVDAGK100Hal(10.3.2)[60E264DA-1FBB-37D2-A834-8A15DBD562D3]@0xffffff7fa228a000->0xffffff7fa2437fff

dependency: com.apple.nvidia.driver.NVDAResman(10.3.2)[6DE4BC26-EEC1-379B-AD41-532FB0A529CD]@0xffffff7fa082f000

dependency: com.apple.iokit.IOPCIFamily(2.9)[D02EA6D7-7A0E-3F9D-B179-75AC2228E449]@0xffffff7fa0494000



BSD process name corresponding to current thread: kernel_task



Mac OS version:

17G3025

Is there any way to determine if it is a graphics card fault, or if it is a fault related to the GPU perhaps failing?

Around 2 weeks ago I had the Nvidia web driver installed and it was causing all sorts of problems. I uninstalled the driver as well as CUDA (as I currently don't need it). It would appear to me that it is the Apple suppled driver that is causing me the problem(s) now.

I haven't noticed any graphical glitches other than the occasional menu bar at the top of the screen going black and then returning to normal every now and then.

System is a 2010 Mac Pro.
2 x 3.33GHz CPU
64GB RAM
Zotac GTX 680 2GB that was flashed with the EVGA ROM about 2-3 years ago.

Thank you

 

[georgechew]

Greetings,
Search up in the forum regarding northbridge heatsink issue. A very likely cause of it.

 

[leon771]

Greetings,
Search up in the forum regarding northbridge heatsink issue. A very likely cause of it.

I’m familiar with that issue, but why would it lead to an Nvidia kernel panic?

 

[georgechew]

I’m familiar with that issue, but why would it lead to an Nvidia kernel panic?

I'm no expert to explain how, but kernel panic on my 5,1 was solved when I rectified the northbridge heatsink issue.

 

[LightBulbFun]

Good morning,

I have a problem with my Mac Pro 5,1 Kernel Panicking very frequently (sometimes every 2 minutes), then stable for a week with no issue.

This morning I was filling in a form on the internet and the Mac Pro shutdown then restarted.

This is the info that it provided.
From what I can tell, it would appear to be related to my graphics card.

Spoiler: Panic

Anonymous UUID: E66ABFD5-8BD0-2D08-CFE7-C8502D1CA3F9

Mon Nov 19 06:35:00 2018

*** Panic Report ***

panic(cpu 0 caller 0xffffff7fa089bf1a): NVRM[0/5:0:0]: Read Error 0x006100c0: CFG 0xffffffff 0xffffffff 0xffffffff, BAR0 0x8a000000 0xffffffa75bfe8000 0x0e4000a2, D0, P3/4

Backtrace (CPU 0), Frame : Return Address

0xffffffa75ab739b0 : 0xffffff801fc6b1c6

0xffffffa75ab73a00 : 0xffffff801fd93a74

0xffffffa75ab73a40 : 0xffffff801fd85d44

0xffffffa75ab73ab0 : 0xffffff801fc1d1e0

0xffffffa75ab73ad0 : 0xffffff801fc6ac3c

0xffffffa75ab73c00 : 0xffffff801fc6a9fc

0xffffffa75ab73c60 : 0xffffff7fa089bf1a

0xffffffa75ab73ce0 : 0xffffff7fa0969e19

0xffffffa75ab73d40 : 0xffffff7fa230d2fb

0xffffffa75ab73d50 : 0xffffff7fa090917b

0xffffffa75ab73d70 : 0xffffff7fa09faf2e

0xffffffa75ab73dd0 : 0xffffff7fa0870d30

0xffffffa75ab73e90 : 0xffffff7fa0836aaa

0xffffffa75ab73ed0 : 0xffffff801fca4634

0xffffffa75ab73f40 : 0xffffff801fca4195

0xffffffa75ab73fa0 : 0xffffff801fc1c557

Kernel Extensions in backtrace:

com.apple.nvidia.driver.NVDAResman(10.3.2)[6DE4BC26-EEC1-379B-AD41-532FB0A529CD]@0xffffff7fa082f000->0xffffff7fa0b08fff

dependency: com.apple.iokit.IOPCIFamily(2.9)[D02EA6D7-7A0E-3F9D-B179-75AC2228E449]@0xffffff7fa0494000

dependency: com.apple.iokit.IONDRVSupport(519.21)[8E0EE518-0D53-38C0-B308-2C6CD9EC5B7B]@0xffffff7fa0813000

dependency: com.apple.iokit.IOGraphicsFamily(519.21)[E5B53C5D-BF6E-3D72-9091-A80DFE5426C9]@0xffffff7fa07bd000

dependency: com.apple.AppleGraphicsDeviceControl(3.20.14)[D3CF5F89-2DA6-3F73-995D-F5DDE191E192]@0xffffff7fa0823000

com.apple.nvidia.driver.NVDAGK100Hal(10.3.2)[60E264DA-1FBB-37D2-A834-8A15DBD562D3]@0xffffff7fa228a000->0xffffff7fa2437fff

dependency: com.apple.nvidia.driver.NVDAResman(10.3.2)[6DE4BC26-EEC1-379B-AD41-532FB0A529CD]@0xffffff7fa082f000

dependency: com.apple.iokit.IOPCIFamily(2.9)[D02EA6D7-7A0E-3F9D-B179-75AC2228E449]@0xffffff7fa0494000



BSD process name corresponding to current thread: kernel_task



Mac OS version:

17G3025

Is there any way to determine if it is a graphics card fault, or if it is a fault related to the GPU perhaps failing?

Around 2 weeks ago I had the Nvidia web driver installed and it was causing all sorts of problems. I uninstalled the driver as well as CUDA (as I currently don't need it). It would appear to me that it is the Apple suppled driver that is causing me the problem(s) now.

I haven't noticed any graphical glitches other than the occasional menu bar at the top of the screen going black and then returning to normal every now and then.

System is a 2010 Mac Pro.
2 x 3.33GHz CPU
64GB RAM
Zotac GTX 680 2GB that was flashed with the EVGA ROM about 2-3 years ago.

Thank you

looking at the panic Log I do indeed see the NVIDIA Kepler Graphics drivers in back trace

so I would try another video card if you have one on hand and see if the system stops Kernel panicking

 

[Asgorath]

panic(cpu 0 caller 0xffffff7fa089bf1a): NVRM[0/5:0:0]: Read Error 0x006100c0: CFG 0xffffffff 0xffffffff 0xffffffff, BAR0 0x8a000000 0xffffffa75bfe8000 0x0e4000a2, D0, P3/4

These panics are an indication of a hardware problem. That is, the driver attempted to read a register or some other value from the GPU, and got back 0xffffffff instead. This could easily be caused by a northbridge problem, I'm pretty sure the cMP is old enough that the north bridge is involved in CPU-GPU communication (or could have messed with the PCIe transfers to the point that the GPU was no longer accessible by the CPU).

 

[leon771]

Northbridge temps are ok.
Rivet is in one piece still.
Does this imply issue with GPU perhaps?
GPU RAM?

I’ve transferred an RX580 into it and see how it goes over the next few weeks. If it’s all good then I’m happy to think it’s the 680GTX.

 

[leon771]

No issues at all with system since installed RX 580.

I think my GTX 680 may be dying.

I will install it in my 5,1 flashes 4,1 at home and see if that system starts crashing.

 

[Norbert Mikołajczyk]

I’ve got similar problem with two of my gtx 680 under Mojave when rendering multiple instances of current after effects. It renders for some time and then bang, restart and the backtrace is similar and leads to the Nvidia driver...

 

[leon771]

No issues at all with system since installed RX 580.

I think my GTX 680 may be dying.

I will install it in my 5,1 flashes 4,1 at home and see if that system starts crashing.

I have had my GTX 680 installed in my 4,1 flashed 5,1 for 2 weeks now. I haven't had a single crash in all this time.
I am thinking that I have some sort of kext corruption on my real 5,1 installation that was causing the GTX 680 to fail in that system.
I only started to have issues when I installed an Nvidia web driver update (the pulled 108 version).
Uninstalled the web driver and associated CUDA seemed to "fix" the problem, but was still getting crashes with the Apple supplied kext.

In my 4,1 which hasn't had an Nvidia web driver on its installation, everything is rock solid with the GTX 680.
I'm using the same power cables to power the card in both systems, so there isn't a power supply issue.

My 5,1 which crashes with the GTX 680 has been running an RX 580 flawlessly all this time as well.

 

[bwinter88]

I have had my GTX 680 installed in my 4,1 flashed 5,1 for 2 weeks now. I haven't had a single crash in all this time.
I am thinking that I have some sort of kext corruption on my real 5,1 installation that was causing the GTX 680 to fail in that system.
I only started to have issues when I installed an Nvidia web driver update (the pulled 108 version).
Uninstalled the web driver and associated CUDA seemed to "fix" the problem, but was still getting crashes with the Apple supplied kext.

In my 4,1 which hasn't had an Nvidia web driver on its installation, everything is rock solid with the GTX 680.
I'm using the same power cables to power the card in both systems, so there isn't a power supply issue.

My 5,1 which crashes with the GTX 680 has been running an RX 580 flawlessly all this time as well.

I just wanted to add my experience to this thread. I've had this exact kernel panic for months now and just discovered the cause of the problem.

Are you ready?

It was a bad DIMM memory module!

The crash report backtraces to a lot of Nvidia-related stuff, but I think it's because I am using graphics-accelerated software alongside all my memory-intensive tasks.

Memory fills up and hits a bad sector on the bad module—> boom. kernel panic.

For those of you who find this thread who are having this issue, first go to your system profiler and make sure all your installed DIMM modules are displayed and recognized. Then proceed to download Rember and test all your memory modules.

I removed the malignant DIMM module, haven't had a crash since. Hooray!