iMac late 2014 kernel panics

[hifivoice]

It happened 2 times last month, but last since last Friday I suffer from a structural problem with my iMac (late 2014, i5 3.5GHz). After a while it stalls with a screen that looks as below, and the computer reboots.

When I start the computer after o long time off, it can be used for quite some hours. When it gets into the status described above, the situation repeats itself quickly again. This makes a temperature/HW issue suspicious, and my suspicion goes to heavy usage of the graphic card (it happens mostly with GfX intensive situations; watching videos or looking at websites with graphics).

The computer sometimes leaves a kernel panic report (examples attached at the tail of this message). Some lines:

• CPU Machine Check Architecture Error Dump (CPU: Intel(R) Core(TM) i5-4690 CPU @ 3.50GHz, CPUID: 0x306C3)
• panic(cpu 2 caller 0xffffff80104db92d): Kernel trap at 0xffffff7f93565764, type 14=page fault
• Kernel Extensions in backtrace: com.apple.kext.AMDRadeonX4000(2.0.9)

In some cases the reboot starts with jittery vertical red lines. You can see the progress bar evolving, it re-starts again but then immediately without the red lines.

Stalling with vertical lines:

First boot attempt with red lines:

Immediate second boot attempt without red lines (which to me excludes problems with the LCD drivers):

For sure I did some obvious things:

• Reset PRAM, VRAM
• Run the Apple HW test (no issues)
• Did run Etrecheck (no issues)
• Re-install OS with full installer
  • I don’t think it OS related, as the stripes also appear when rebooting, when the OS has not been loaded yet. re-connect the RAM modules (Hynix)

Anyone recognising this situation or having tips?

 

[SecuritySteve]

Definitely a graphics card issue. Can you try increasing the fans to rule out a thermal problem? It may be that your first few boots are failing as the graphics card cools to a temperature that is workable.

 

[hifivoice]

Thx. Temperature is pretty fine overall (ambient is 23 now, GfX not above 40 degrees).

 

[SecuritySteve]

Yeah the temps seem fine.

Is there anything you can tell us that reproduces the problem? I.e. running a specific app under certain conditions, or happens X minutes after boot (estimated).

 

[chscag]

Agree with @SecuritySteve that it's a graphic card problem. Looks like it might be intermittent which are the worse kind of problems to resolve. The graphics card in that model is replaceable but requires a major take apart of the machine to get at it. Also, a new GPU can be quite expensive. Go to www.ifixit.com and look over the procedure for replacing the card.

 

[mikehalloran]

Don't ever diagnose one without a known good battery. 5 years is long enough for these symptoms to show although they are more common on the 2009–2013 right now.

Open it up, insert a CR2032. Use some blue painters' tape to hold it together.

See if you still have a problem.

If fixed and there's an HDD in there, order a BR2032 from Amazon — electrically identical but it's more heat resistant and is the OE battery. If SSD only, the CR2032 is fine.

 

[SecuritySteve]

Don't ever diagnose one without a known good battery. 5 years is long enough for these symptoms to show although they are more common on the 2009–2013 right now.

Open it up, insert a CR2032. Use some blue painters' tape to hold it together.

See if you still have a problem.

If fixed and there's an HDD in there, order a BR2032 from Amazon — electrically identical but it's more heat resistant and is the OE battery. If SSD only, the CR2032 is fine.

Given that his kernel panics are panicking at the Radeon kext, we know it is a GPU issue, and is not related to the PRAM battery. VRAM should be powered by the GPU socket itself.

 

[mikehalloran]

Given that his kernel panics are panicking at the Radeon kext, we know it is a GPU issue, and is not related to the PRAM battery. VRAM should be powered by the GPU socket itself.

And you may be right.

Given that I have seen this a lot, I'll stick with what I posted if you don't mind.

 

[hifivoice]

Yeah the temps seem fine.

Is there anything you can tell us that reproduces the problem? I.e. running a specific app under certain conditions, or happens X minutes after boot (estimated).

Reproducing is very arbitrary. Sometimes when watching video, other times when loading a web page, or when moving a finder window (in the latter case I saw the screen flicker once with the pattern, before the pattern appeared permanently). After a few seconds the computer reboots. When it happened once, the next time is a matter of minutes. I have to cool down the computer, and after that I can work with it again.

I forgot to attach a log file. Here is one (the first part of it):
Anonymous UUID: 5C31AD3E-520B-42D4-CD49-BE32328910CF


Sat Jul 6 15:12:00 2019


*** Panic Report ***

panic(cpu 2 caller 0xffffff80104db92d): Kernel trap at 0xffffff7f93565764, type 14=page fault, registers:

CR0: 0x000000008001003b, CR2: 0xffffffa3c503d000, CR3: 0x00000000153ef000, CR4: 0x00000000001626e0

RAX: 0xffffffa3c503d000, RBX: 0x0000000000005434, RCX: 0x0000000002000000, RDX: 0x0000000003000000

RSP: 0xffffffa3c503bc00, RBP: 0xffffffa3c503bc20, RSI: 0xffffff8052578d70, RDI: 0x0000000000000000

R8: 0x0000000000000002, R9: 0x0000000000000001, R10: 0x00000000e0000000, R11: 0xffffffa3c51ced90

R12: 0xffffff804f5e0a80, R13: 0xffffff834a82f000, R14: 0x0000000000000000, R15: 0xffffff805245ca00

RFL: 0x0000000000010246, RIP: 0xffffff7f93565764, CS: 0x0000000000000008, SS: 0x0000000000000000

Fault CR2: 0xffffffa3c503d000, Error code: 0x0000000000000002, Fault CPU: 0x2, PL: 0, VF: 1


Backtrace (CPU 2), Frame : Return Address

0xffffffa3c503b6d0 : 0xffffff80103aea2d

0xffffffa3c503b720 : 0xffffff80104e9e95

0xffffffa3c503b760 : 0xffffff80104db70a

0xffffffa3c503b7d0 : 0xffffff801035bb40

0xffffffa3c503b7f0 : 0xffffff80103ae447

0xffffffa3c503b910 : 0xffffff80103ae293

0xffffffa3c503b980 : 0xffffff80104db92d

0xffffffa3c503baf0 : 0xffffff801035bb40

0xffffffa3c503bb10 : 0xffffff7f93565764

0xffffffa3c503bc20 : 0xffffff7f93575107

0xffffffa3c503bc40 : 0xffffff7f9356564e

0xffffffa3c503bc80 : 0xffffff7f935861ab

0xffffffa3c503bcd0 : 0xffffff7f93574ed8

0xffffffa3c503bcf0 : 0xffffff7f9350c7f0

0xffffffa3c503bd40 : 0xffffff7f9350a7da

0xffffffa3c503bd80 : 0xffffff7f914e18b0

0xffffffa3c503bdf0 : 0xffffff7f93508e7c

0xffffffa3c503be40 : 0xffffff8010a2f5ab

0xffffffa3c503be80 : 0xffffff8010a2f265

0xffffffa3c503bf00 : 0xffffff8010a2e827

0xffffffa3c503bf50 : 0xffffff8010a30316

0xffffffa3c503bfa0 : 0xffffff801035b0ce

Kernel Extensions in backtrace:

com.apple.iokit.IOAcceleratorFamily2(404.9)[CC230675-4A72-3796-ADD2-53723A8A1E18]@0xffffff7f914b4000->0xffffff7f91558fff

dependency: com.apple.driver.AppleMobileFileIntegrity(1.0.5)[A367CF6B-B3EA-36FE-86E6-03B771CA8A69]@0xffffff7f91188000

dependency: com.apple.iokit.IOSurface(255.5.4)[029BB0E8-5E39-3E49-A97C-225D2B1A6467]@0xffffff7f91430000

dependency: com.apple.iokit.IOPCIFamily(2.9)[CC6A465F-5A24-304D-B9DF-8C27819CC214]@0xffffff7f90c95000

dependency: com.apple.iokit.IOGraphicsFamily(530.66)[2B24FC25-7174-3BEA-9701-E9E520137222]@0xffffff7f91454000

dependency: com.apple.iokit.IOReportFamily(47)[5772C11C-ED28-322A-910F-874583B23101]@0xffffff7f914ad000

com.apple.kext.AMDRadeonX4000(2.0.9)[6AD64509-77E3-32EB-8800-FCB397DC7DA0]@0xffffff7f93508000->0xffffff7f93945fff

dependency: com.apple.iokit.IOSurface(255.5.4)[029BB0E8-5E39-3E49-A97C-225D2B1A6467]@0xffffff7f91430000

dependency: com.apple.iokit.IOPCIFamily(2.9)[CC6A465F-5A24-304D-B9DF-8C27819CC214]@0xffffff7f90c95000

dependency: com.apple.iokit.IOGraphicsFamily(530.66)[2B24FC25-7174-3BEA-9701-E9E520137222]@0xffffff7f91454000

dependency: com.apple.iokit.IOAcceleratorFamily2(404.9)[CC230675-4A72-3796-ADD2-53723A8A1E18]@0xffffff7f914b4000


BSD process name corresponding to current thread: kernel_task


Mac OS version:

18F132


Kernel version:

Darwin Kernel Version 18.6.0: Thu Apr 25 23:16:27 PDT 2019; root:xnu-4903.261.4~2/RELEASE_X86_64

Kernel UUID: 7C8BB636-E593-3CE4-8528-9BD24A688851

Kernel slide: 0x0000000010000000

Kernel text base: 0xffffff8010200000

__HIB text base: 0xffffff8010100000

System model name: iMac15,1 (Mac-42FD25EABCABB274)


System uptime in nanoseconds: 22293270213

last loaded kext at 22288849203: com.apple.kext.AMDRadeonX4030HWLibs 1.0 (addr 0xffffff7f941f9000, size 7659520)

loaded kexts:
.... (Ed: list of more kexts, no suspicious stuff seen there)

 

[hifivoice]

From the log you can see that the last extension being loaded is related to the graphic card, and 4,4 microseconds later the iMacs panics.

Either that extension file is corrupt (unlikely; and I did a combo installation to rule out), or this is a sign the graphic card is used in a specific way that enables some faulty hardware (logic board, GPU, or power supply)

I have no tools to open up the Mac. From iFixit it is also clear the repair is quite labour intensive. So I’m afraid it’s going to be a repair(at least 3 weeks no computer), or a new iMac. :-(

 

[SecuritySteve]

Yeah this seems like an Apple store issue. I'd bring it in and show them everything you've posted here. Pics and panics included.

 

[Khaleal]

If SMC/PRAM reset and/or re-installation of the OS doesn't help then it's a GPU issue for sure.
It's soldered to the logic board, and to fix it you'll need to replace the whole logic board.
Now you have two options:
1. visit Apple store and see how much they'd charge to replace the board
2. buy the board secondhand off eBay (this for example: https://www.ebay.com/itm/iMac-A1419...878656?hash=item4698b79480:g:1sUAAOSwRkJdG2rD) and replace it yourself.

 

[hifivoice]

To inform whomever might have similar problems; logic board has been replaced, the iMac works fine again.

Before repair; I could easily make it crash by running the Heaven benchmark. When the GfX processor exceeded 80 degrees (typically within a minute), it would crash in a matter of seconds.