Mac will not wake up from sleep, reboot or cold boot - BLACK SCREEN OR STUCK AT BOOTING APPLE LOGO

I cannot update the message, so here is a new one:

Also, "/Library/Logs/DiagnosticReports/Sleep Wake Failure_2018-11-01-104733_Ma.diag" shows:

Thread 0x465 Thread name "IOGraphicsWorkLoop" 9 samples (1-9) priority 93 (base 81) cpu time 0.089

<IO tier 0>

*9 call_continuation + 23 (kernel + 620887) [0xffffff8000297957] 1-9

*9 IOWorkLoop::threadMain() + 22 (kernel + 7054358) [0xffffff80008ba416] 1-9

*9 IOWorkLoop::runEventSources() + 433 (kernel + 7056945) [0xffffff80008bae31] 1-9

*9 IOInterruptEventSource::checkForWork() + 287 (kernel + 7062991) [0xffffff80008bc5cf] 1-9

*9 IOFramebuffer::systemWork(OSObject*, IOInterruptEventSource*, int) + 211 (IOGraphicsFamily + 48111) [0xffffff7f81a4abef] 1-9

*9 IOFramebuffer::checkPowerWork(IOFBController*, unsigned int) + 140 (IOGraphicsFamily + 65156) [0xffffff7f81a4ee84] 1-9

*9 IOFramebuffer::checkPowerWork(unsigned int) + 503 (IOGraphicsFamily + 68841) [0xffffff7f81a4fce9] 1-9

*9 AMDFramebuffer::setAttribute(unsigned int, unsigned long) + 3747 (AMDFramebuffer + 42595) [0xffffff7f83f64663] 1-9

*9 AMDFramebuffer::setAttribute(unsigned int, unsigned long) + 833 (AMDFramebuffer + 39681) [0xffffff7f83f63b01] 1-9

*9 AMDFramebuffer::doSetPowerState(unsigned int) + 147 (AMDFramebuffer + 69219) [0xffffff7f83f6ae63] 1-9

*9 AMDFramebuffer::setSystemPowerState(unsigned int) + 98 (AMDFramebuffer + 70146) [0xffffff7f83f6b202] 1-9

*9 AMDFramebuffer::doWake() + 2066 (AMDFramebuffer + 72274) [0xffffff7f83f6ba52] 1-9

*9 AMDFramebuffer::handleEvent(int, void*) + 162 (AMDFramebuffer + 38642) [0xffffff7f83f636f2] 1-9

*9 IOFramebuffer::deliverFramebufferNotification(int, void*) + 181 (IOGraphicsFamily + 59317) [0xffffff7f81a4d7b5] 1-9

*9 AMDRadeonX4200_AMDGraphicsAccelerator::iofbNotificationHandler(OSObject*, void*, IOFramebuffer*, int, void*) + 43 (AMDRadeonX4200 + 13325) [0xffffff7f8368440d] 1-9

*9 AMDRadeonX4200_AMDGraphicsAccelerator::internalDidWake() + 273 (AMDRadeonX4200 + 10851) [0xffffff7f83683a63] 1-9

*9 AMDRadeonX4200_AMDEllesmereHardware::powerUp() + 23 (AMDRadeonX4200 + 516533) [0xffffff7f836ff1b5] 1-9

*9 AMDRadeonX4200_AMDHardware::powerUp() + 208 (AMDRadeonX4200 + 386714) [0xffffff7f836df69a] 1-9

*9 AMDRadeonX4200_AMDHardware::powerUpHWEngines() + 34 (AMDRadeonX4200 + 385978) [0xffffff7f836df3ba] 1-9

*9 AMDRadeonX4200_AMDVIPM4Engine::powerUp() + 111 (AMDRadeonX4200 + 449425) [0xffffff7f836eeb91] 1-9

*9 AMDRadeonX4200_AMDHWChannel::waitForIdle() + 99 (AMDRadeonX4200 + 291029) [0xffffff7f836c80d5] 1-9

*9 AMDRadeonX4200_AMDVIPM4Channel::isIdle() + 34 (AMDRadeonX4200 + 446218) [0xffffff7f836edf0a] 1-9

*9 AMDRadeonX4200_AMDHWRegisters::read(unsigned int) + 40 (AMDRadeonX4200 + 340450) [0xffffff7f836d41e2] (running) 1-9

Eight days later without sleeping Mac nor rebooting, so far so good. I guess that is a GPU or motherboard (logic board) hardware issue.

Thanks. No problems were found by EtreCheckPro, except "Minor Issues: Abnormal shutdown - Your machine shut down abnormally". Anyway, I guess the application only checks for software conflicts, not hardware ones. And as said above, the issue also arises booting from external disks with newly installed macOS. So, no software conflicts should be there.

Additionally, I forgot to mention before (and these forums do not allow to edit messages after a few minutes of posting, which is not good) that the Mac is protected by an Uninterruptible Power Supply (UPS).

Thanks. I tried what is indicated in the Apple article HT200553 of your link to no avail. The kernel panic happened only once after many, many, many (I guess, 200 or more) of such sleep/wakeup and failed reboots/cold boots failures over a long period of time. I can post the EtreCheckPro test, but I guess it does not make much sense doing it for a boot from an external disk with nothing inside but a freshly installed macOS from scratch, and as said before the issue also happens in such environment and does not happen when booting any of the remaining four identical iMacs.

In case it is useful, I found several instances of "GPU Reset" like this at "/Library/Logs/DiagnosticReports/Kernel_2018-10-19-181953_Mac.gpuRestart":

Fri Oct 19 18:19:53 2018

Event: GPU Reset

Date/Time: Fri Oct 19 18:19:53 2018

Application: WindowServer

Path:

Tailspin: /Library/Logs/DiagnosticReports/gpuRestart2018-10-19-181953.tailspin

GPUSubmission Trace ID: 0

OS Version: Mac OS X Version 10.12.6 (Build 16G1510)

Graphics Hardware: Radeon Pro 580

Signature: e

Report Data:

GPURestartReportStart

------------------------

Hung Channels: GFX

------------------------

[00] AccelChannel: GFX

Pending Command from : GLCtx

PendingCommandTimestamp: 0x00822dda, TotalDWords: 0x00000000, GART Offset=0x0000000000000000, stamp_idx=0, estamp=0x00822dda

PendingCommandStart:

PendingCommandEnd

------------------------

[00] GFX HWChannel : Enabled, NotIdle

LastReadTimestamp : 0x00822dd9

NextSubmitTimestamp : 0x00822dee

[GFX, ts:0x00822dda]: No semaphore wait

[GFX, ts:0x00822dda]: semaphore signal: 0xffffff8070450000

HWSemaphore 0xffffff8070450000 Signal Event: [GFX channel[0] TS:0x00822ded lastReadTS:0x00822dd9]

------------------------

SRBM_CNTL : 0x00000000 | CP_ME_CNTL : 0x00000000

SRBM_STATUS: 0x20020a40 | SRBM_STATUS2: 0x00000000

GRBM_STATUS: 0xa0003028 | GRBM_STATUS2: 0x70000008

------------------------

GPU HangFlags 0x00000007: AsicHangState 0x00000007, AsicResetRequirement 0x00000002

IndividualEngineHang: 1

NonEngineBlockHang : 1

FenceNotRetired : 1

PerEngineReset : 0

FullAsicReset : 1

HangEngineBitmap : 0x0000000e

Hung Engines Detected by Cail: GRAPHICS COMPUTE0 COMPUTE1

------------------------

AMDRadeonX4200_AMDEllesmereGraphicsAccelerator PCIe Device: [1:0:0] State: ENABLED

Configuration: deviceBits: 0x940167df, capabilityBits: 0x30840300

TotalVideoRAMBytes: 0x0000000200000000 (8589934592)

------------------------

[00] PM4 Engine : Enabled, NotIdle

[00] GFX Channel: Enabled, NotIdle, lastReadTimestamp 0x00822dd9, nextSubmitTimestamp 0x00822dee

[01] C0 Channel: Enabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

[02] C1 Channel: Enabled, Idle, lastReadTimestamp 0x0000087e, nextSubmitTimestamp 0x00000881

[03] KIQ Channel: Enabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

Hung Blocks Detected by CAIL: 0x00008006

HWRing[1]-[Me,Queue,Pipe]=[1,0,0]:

PQ_RPTR 0x00000000, PQ_WPTR 0x00000000, HQD_ACTIVE 0x00000000, HQ_STATUS [0x00000080, 0x00000000]

HWRing[2]-[Me,Queue,Pipe]=[1,0,1]:

PQ_RPTR 0x00000000, PQ_WPTR 0x00000000, HQD_ACTIVE 0x00000000, HQ_STATUS [0x00008040, 0x00000000]

HWRing[3]-[Me,Queue,Pipe]=[1,0,3]:

PQ_RPTR 0x00000000, PQ_WPTR 0x00000058, HQD_ACTIVE 0x00000001, HQ_STATUS [0x00008040, 0x00000000]

------------------------

[01] SDMA Engine : Enabled, Idle

[04] sDMA Channel: Enabled, Idle, lastReadTimestamp 0x0006ab8a, nextSubmitTimestamp 0x0006ab8f

------------------------

[02] SDMA Engine : Enabled, Idle

[05] sDMA Channel: Enabled, Idle, lastReadTimestamp 0x00126f2c, nextSubmitTimestamp 0x00126f3f

------------------------

[03] UVD Engine : Disabled, Idle

[06] UVD Channel: Disabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

[07] UVDE Channel: Disabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

[08] UVDE Channel: Disabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

------------------------

[04] SAMU Engine : Disabled, Idle

[09] SAMU Channel: Disabled, Idle, lastReadTimestamp 0x0000001d, nextSubmitTimestamp 0x0000001e

[10] SAMU Channel: Disabled, Idle, lastReadTimestamp 0x0000001b, nextSubmitTimestamp 0x0000001c

------------------------

[05] VCE Engine : Disabled, Idle

[11] VCE Channel: Disabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

[12] VCEL Channel: Disabled, Idle, lastReadTimestamp 0x00000000, nextSubmitTimestamp 0x00000001

------------------------

: ** GPU Debug Info Start **

…/… matrix of numbers here deleted for simplicity …/…

Now, when trying to update the message above, I got:

Your content could not be saved. Please try again in a few moments. Click here to refresh this page.

Yet, my report was lost. Here it is again:

Additionally, "/Library/Logs/DiagnosticReports/Sleep Wake Failure_2018-11-01-104733_Mac.diag" shows:

ÉYES$,$A$|Á DumpGPURestart /com.apple.crashreporter.writereport.gpuRestart UNBUNDLED ||| kernel ((null)) %…hexadecimal code here…‰$Ç(]u [ ∂ YÂP ˆˇˇˇˇP $l$ÄeÜdaemon$}¿$Äz$~)ÜKernel$~^$ÄØ$~®áUNKNOWN$~ÎÑnoop$

More: searching the full disk for files containing (all words)

machine check

(as Apple document HT200553 indicates) finds no hits, albeit the file above shows it.

I forgot to say that the issue happens all the time if the Mac has been sleeping for a long time (one or two hours or so), but at least on one instance it arose after a few minutes of sleep (five or so). Additionally, the failed rebooting usually arises after the first, second or third trial, but on one instance it took 11 reboots in a row to show when I was testing. Really amazing.

Two days with no problem using the Mac, but without rebooting or sutting down the Mac, besides not putting it to sleep. Now I leave it on overnight with "Apple - System Preferences - Energy Saver - Power" configured as:

- Prevent computer from sleeping automatically when the display is off (as has always been)

- Put hard disks to sleep when possible

And "Apple - System Preferences - Displays - Display" with Brightness as low as allowed.

I will keep it like that until sent to repair.

Good news. The problem has gone after replacing the main mother logic board (keeping CPU and SSD which were not soldered). To anyone having this problem, I recommend the following to troubleshoot:

1. Try to reboot or cold boot from an external disk (disconnect before all peripherals except Apple keyboard and mouse in case they are wired), with nothing inside but macOS installed. Configured to ask for username and password is rather quick: if the Mac stops at the booting black screen with white Mac icon logo and progress bar, there is the problem. If it shows the login window to enter username and password, do not do it; just restart again. And so on for 15 trials or so in a row (in one instance, it took me 12 trials in a row to show the problem, although usually showed after one to three trials).

2. If the problem goes away, it may be a software issue. Check, update or uninstall third-party applications installed when issue showed first. I wish there was now something like Conflict Catcher (for Mac OS 9) from Casady & Greene

https://en.wikipedia.org/wiki/Conflict_Catcher

to troubleshoot also the latest versions of macOS. If the problem persists, try to change PRAM.

3. If the problem goes away, the culprit was badly sit or defectiva RAM. If the problem persists, probably it is a GPU issue, like in my case above. Replace it. If soldered, it will require to replace the full mainboard-motherboard-logicboard (I write all names for indexing and search purposes), as in the example above.

I must say that this Mac was received by end August 2017, and has been used about 15 hours/day, seven days a week. The issue arose after about one year. Luckily, it was still under the 2-year warranty period. Other identical four iMacs have not shown any issue so far, albeit they have been used much less.

Happy New Year!

Oops! When I said "change PRAM", I meant "change RAM". On the other hand, try first to reset PRAM and if the problem persists, reset SMC. If the problem persists, do as indicated in step 1 and following, above. Maybe rebuilding directory with DiskWarrior and after that repairing with DiskUtility could help.