16" MBP GPU Panics

[klagermkii]

My worry with trusting the Geekbench 5 issue is that those kinds of logs are very similar to what you'd see for any long-running GPU compute kernel that exceeds an internal timeout for completing. And the reason you'd be seeing it with Geekbench 5 and not 4 is that they increased the complexity of the test to put more strain on newer hardware. That could easily then take too long on something weak like a UHD 630.

(I get the same messages on my MBP 16" and I don't have any panics, and I've also seen very similar messages when messing around with GPGPU programming elsewhere)

 

[StolenPhone]

UPDATE: Catalina beta seems to do the trick.

Ok... I just got off the phone with Apple. They said they did some board testing on my LB and they said the technicians did find an anomaly.

However, because we all know Apple swaps LBs, even when the wind blows, and the fact that they said my new machine will be here on Wed, I will run the Catalina beta and come back here with the results.

Thank you for being one to test this with the beta.... it at least allows to see if some of these are hardware or not.

 

[gazwas]

Really looking forward to your results with the Catalina beta.

I get the impression the response you’re had from Apple sound like they are clueless.

 

[StolenPhone]

Ok...I got my hands on Catalina 10.15.4. Beta 3 and tried it out. I still panics, but even worse now. I can't even boot into normal mode, even after going into SafeMode.

Luckily Apple said my new machine will get here early this week, so I will report back if the new machine panics at all.

 

[Squuiid]

Ok...I got my hands on Catalina 10.15.4. Beta 3 and tried it out. I still panics, but even worse now. I can't even boot into normal mode, even after going into SafeMode.

Luckily Apple said my new machine will get here early this week, so I will report back if the new machine panics at all.

Sorry to hear that. I think your issue is indeed rather different to others. I hope Apple are able to make this right for you.

 

[StolenPhone]

They are... new machine arrives today.

Yes, my issue is definitely different from others, but my point as OP of this discussion was there are a group of folks experiencing what I have as shown in the original post.

I think it may be good for people to try out catalina 10.15.4 beta 3 (or even wait for the release) and run geek bench. If it still panics, its likely the same problem I have.

Anyways thank you to those who have responded in this thread regarding Geekbench and Cat 10.15.4B3 as it really helps narrow and identify if this is a hardware problem.

I'm hoping all of the information here will help others determine whether they need a new LB or not.

 

[Squuiid]

Now seeing this on my 2019 16".
This 2019 MacBook Pro really is a disaster of a machine.

panic(cpu 1 caller 0xfffffff00bbcf7b4): x86 CPU CATERR detected
Debugger message: panic
Memory ID: 0x6
OS version: 17P54244e
macOS version: 19E242d
Kernel version: Darwin Kernel Version 19.4.0: Wed Feb 19 00:13:44 PST 2020; root:xnu-6153.100.206.0.1~33/RELEASE_ARM64_T8010
Kernel UUID: 81B2E9D0-E02F-34BB-89CC-BCF071907D18
iBoot version: iBoot-5540.100.194
secure boot?: YES
x86 EFI Boot State: 0x16
x86 System State: 0x0
x86 Power State: 0x0
x86 Shutdown Cause: 0x1
x86 Previous Power Transitions: 0x10001000100
PCIeUp link state: 0x89473614
Paniclog version: 13
Kernel slide: 0x0000000003b98000
Kernel text base: 0xfffffff00ab9c000
mach_absolute_time: 0x1b78ba14fbd
Epoch Time: sec usec
Boot : 0x5e57d9b7 0x000da2a5
Sleep : 0x5e5e8328 0x0005cf48
Wake : 0x5e5e839e 0x000f17f3
Calendar: 0x5e5e85d3 0x00078ab4

Panicked task 0xffffffe000484e00: 3298 pages, 221 threads: pid 0: kernel_task
Panicked thread: 0xffffffe000716a40, backtrace: 0xffffffe01570b4b0, tid: 347
lr: 0xfffffff00b460678 fp: 0xffffffe01570b4f0
lr: 0xfffffff00b4604d4 fp: 0xffffffe01570b560
lr: 0xfffffff00b57ab38 fp: 0xffffffe01570b610
lr: 0xfffffff00ba2962c fp: 0xffffffe01570b620
lr: 0xfffffff00b45fe3c fp: 0xffffffe01570b990
lr: 0xfffffff00b460194 fp: 0xffffffe01570b9e0
lr: 0xfffffff00c296934 fp: 0xffffffe01570ba00
lr: 0xfffffff00bbcf7b4 fp: 0xffffffe01570ba30
lr: 0xfffffff00bbc0be4 fp: 0xffffffe01570baa0
lr: 0xfffffff00bbc2b0c fp: 0xffffffe01570bb50
lr: 0xfffffff00bbc031c fp: 0xffffffe01570bbe0
lr: 0xfffffff00bb87780 fp: 0xffffffe01570bc10
lr: 0xfffffff00b996660 fp: 0xffffffe01570bc50
lr: 0xfffffff00b995ee4 fp: 0xffffffe01570bc90
lr: 0xfffffff00ba34514 fp: 0x0000000000000000

 

[gazwas]

Now seeing this on my 2019 16".
This 2019 MacBook Pro really is a disaster of a machine.

panic(cpu 1 caller 0xfffffff00bbcf7b4): x86 CPU CATERR detected
Debugger message: panic
Memory ID: 0x6
OS version: 17P54244e
macOS version: 19E242d
Kernel version: Darwin Kernel Version 19.4.0: Wed Feb 19 00:13:44 PST 2020; root:xnu-6153.100.206.0.1~33/RELEASE_ARM64_T8010
Kernel UUID: 81B2E9D0-E02F-34BB-89CC-BCF071907D18
iBoot version: iBoot-5540.100.194
secure boot?: YES
x86 EFI Boot State: 0x16
x86 System State: 0x0
x86 Power State: 0x0
x86 Shutdown Cause: 0x1
x86 Previous Power Transitions: 0x10001000100
PCIeUp link state: 0x89473614
Paniclog version: 13
Kernel slide: 0x0000000003b98000
Kernel text base: 0xfffffff00ab9c000
mach_absolute_time: 0x1b78ba14fbd
Epoch Time: sec usec
Boot : 0x5e57d9b7 0x000da2a5
Sleep : 0x5e5e8328 0x0005cf48
Wake : 0x5e5e839e 0x000f17f3
Calendar: 0x5e5e85d3 0x00078ab4

Panicked task 0xffffffe000484e00: 3298 pages, 221 threads: pid 0: kernel_task
Panicked thread: 0xffffffe000716a40, backtrace: 0xffffffe01570b4b0, tid: 347
lr: 0xfffffff00b460678 fp: 0xffffffe01570b4f0
lr: 0xfffffff00b4604d4 fp: 0xffffffe01570b560
lr: 0xfffffff00b57ab38 fp: 0xffffffe01570b610
lr: 0xfffffff00ba2962c fp: 0xffffffe01570b620
lr: 0xfffffff00b45fe3c fp: 0xffffffe01570b990
lr: 0xfffffff00b460194 fp: 0xffffffe01570b9e0
lr: 0xfffffff00c296934 fp: 0xffffffe01570ba00
lr: 0xfffffff00bbcf7b4 fp: 0xffffffe01570ba30
lr: 0xfffffff00bbc0be4 fp: 0xffffffe01570baa0
lr: 0xfffffff00bbc2b0c fp: 0xffffffe01570bb50
lr: 0xfffffff00bbc031c fp: 0xffffffe01570bbe0
lr: 0xfffffff00bb87780 fp: 0xffffffe01570bc10
lr: 0xfffffff00b996660 fp: 0xffffffe01570bc50
lr: 0xfffffff00b995ee4 fp: 0xffffffe01570bc90
lr: 0xfffffff00ba34514 fp: 0x0000000000000000

Under what circumstances and is yours from the November/December batch?

UPS just delivered mine and not sure what to make of all these recent reports.

 

[Mark_EL]

Under what circumstances and is yours from the November/December batch?

UPS just delivered mine and not sure what to make of all these recent reports.

You could run the Benchmark 5 Metal GPU test and see if it performs as it should. With that I mean, just to check if it doesn‘t give any kernel panics. If it doesn’t, I’d not worry about it anymore and just start using en enjoying the machine. If it gives you any trouble, you‘ll notice it soon enough. But I think that IRLC you will not notice any problems with the Intel GPU. Because if the Intel can’t handle it, it will switch over to the dGPU anyway. Unless you were handed a lemon, like the OP of course.

As you know, I have done the Benchmark 5 test and came to the conclusion that the Intel GPU performs a lot worse than it should. But I do think it is Catalina related, just as a lot of other issues. As long as my MBP switches over to the dGPU when needed (and it doesn’t give me kernel panics) I am not bothered by it too much. I also decided not to actively search for any other issues. It will only stress me out....

 

[matram]

Now seeing this on my 2019 16".
This 2019 MacBook Pro really is a disaster of a machine.

panic(cpu 1 caller 0xfffffff00bbcf7b4): x86 CPU CATERR detected

I believe CATERR is a detected but unrecoverable fatal error in the Intel CPU.

 

[Norbert Mikołajczyk]

Now seeing this on my 2019 16".
This 2019 MacBook Pro really is a disaster of a machine.

panic(cpu 1 caller 0xfffffff00bbcf7b4): x86 CPU CATERR detected
Debugger message: panic
Memory ID: 0x6
OS version: 17P54244e
macOS version: 19E242d
Kernel version: Darwin Kernel Version 19.4.0: Wed Feb 19 00:13:44 PST 2020; root:xnu-6153.100.206.0.1~33/RELEASE_ARM64_T8010
Kernel UUID: 81B2E9D0-E02F-34BB-89CC-BCF071907D18
iBoot version: iBoot-5540.100.194
secure boot?: YES
x86 EFI Boot State: 0x16
x86 System State: 0x0
x86 Power State: 0x0
x86 Shutdown Cause: 0x1
x86 Previous Power Transitions: 0x10001000100
PCIeUp link state: 0x89473614
Paniclog version: 13
Kernel slide: 0x0000000003b98000
Kernel text base: 0xfffffff00ab9c000
mach_absolute_time: 0x1b78ba14fbd
Epoch Time: sec usec
Boot : 0x5e57d9b7 0x000da2a5
Sleep : 0x5e5e8328 0x0005cf48
Wake : 0x5e5e839e 0x000f17f3
Calendar: 0x5e5e85d3 0x00078ab4

Panicked task 0xffffffe000484e00: 3298 pages, 221 threads: pid 0: kernel_task
Panicked thread: 0xffffffe000716a40, backtrace: 0xffffffe01570b4b0, tid: 347
lr: 0xfffffff00b460678 fp: 0xffffffe01570b4f0
lr: 0xfffffff00b4604d4 fp: 0xffffffe01570b560
lr: 0xfffffff00b57ab38 fp: 0xffffffe01570b610
lr: 0xfffffff00ba2962c fp: 0xffffffe01570b620
lr: 0xfffffff00b45fe3c fp: 0xffffffe01570b990
lr: 0xfffffff00b460194 fp: 0xffffffe01570b9e0
lr: 0xfffffff00c296934 fp: 0xffffffe01570ba00
lr: 0xfffffff00bbcf7b4 fp: 0xffffffe01570ba30
lr: 0xfffffff00bbc0be4 fp: 0xffffffe01570baa0
lr: 0xfffffff00bbc2b0c fp: 0xffffffe01570bb50
lr: 0xfffffff00bbc031c fp: 0xffffffe01570bbe0
lr: 0xfffffff00bb87780 fp: 0xffffffe01570bc10
lr: 0xfffffff00b996660 fp: 0xffffffe01570bc50
lr: 0xfffffff00b995ee4 fp: 0xffffffe01570bc90
lr: 0xfffffff00ba34514 fp: 0x0000000000000000

How this happened?
I'm stressing hard my mbp16 with video and animation inside After Effects and some encoding, had not a single crash.
No crash after wake from sleep issues neither.

 

[gazwas]

How this happened?
I'm stressing hard my mbp16 with video and animation inside After Effects and some encoding, had not a single crash.
No crash after wake from sleep issues neither.

Driving it hard is not the problem.

The issue is with the Intel GPU build into the CPU that seems to have either driver issues in Catalina or user reported hardware failures from an unknown batch and quantity of CPU’s.

 

[StolenPhone]

New machine arrived from Apple... no GPU panics.

Production info:

Production year : 2020
Production week : 9 (March)
Model introduced: 2019
Factory: C0 (Quanta Computer (Subsidiary = Tech Com) China)

Nice new LB... Geekbench was run. Got the errors that others got under 10.5.3 which is normal and to be expected for that version. NO panic!!! So its good. I have a good running machine at this stage.

If you get the panic, please run your serial number at the website as shown earlier in the thread to check your MFR batch number (year/week). I can attest mine for March has the issue resolved.

 

[insited]

This is 'philux' from apple forums. My 3rd machine was a board replacement, bringing me to my 4th. The 4th is exhibiting the same issues with absolutely nothing plugged in and bare minimum software.

I have the same early issues that have devolved in to freezing & restarting within the week already:
- Display flickering waking from sleep
- True tone flickering on & off even though it is switched off
- Display 'jumping' to the right, then back to centre again (20 - 30ms, hard to spot)

On machine 1, 2 & 3 the above issues were there straight away, then a week later whenever opening anything that made GPU switch to AMD, the macbook freezes, either restarting itself or me having to manually restart to escape freeze. This makes it unsuable in a 'real world' application.

These issues are nothing like what thousands of others are reporting in regards to Catalina bugs. There is a small percentage with hardware issues and the pile-on of fanbois trying to tell me it's my fault and to gimp my use of the machine to rectify, coupled with Apple doing the same, is mind numbingly frustrating.

I have a 2011 13" mbp, a 2011 15" mbp, 2015 27" iMac, 2016 mbp no touch bar and they all run fine, with High Sierra or Catalina as well.

In reality the biggest issue is the way Apple treats the situation. If you get a faulty iPhone or iPad out the box, they are replaced straight away. For some reason a faulty macbook out the box entails you having to drive around town for it to be assessed.

The authorized Apple Repair centre I took mine to said in a lot of cases, the Apple Hardware Diagnostic is entirely useless since issues need to trigger while scanning hardware, which a lot of the time they won't. They also said basically all the issues I'm having are because of a hardware fault on the logicboard. In my city, the replacements are all from the original 2019 batch still. Great.

My biggest issue is any household appliance I have in my home, the manufacturer issues a replacement for a fault out the box. Apple forcing me to get my new machine 'repaired' turns it in to a refurb, which leaves me asking why I don't then get a discount, or why I don't refund and buy a refurb.

Considering how many bugs macOS has now especially in relation to the T2 chip (and not including the bad batch of MBP 2019 16"), It just works should now be It *just* works 😂

What's truly pathetic is since Apple are currently denying mbp 2019 16" faults, my quickest course of action if I want to use the exact machine I paid for would be to refund the one I have now, then fly over to USA (same keyboard layout) and purchase the same machine but from a region of the world where the rectified 2020 batch is now available. I could have all that done before my return call and Apple's solution to my 4th going on 5th replacement now.

 

[gazwas]

New machine arrived from Apple... no GPU panics.

Production info:

Production year : 2020
Production week : 9 (March)
Model introduced: 2019
Factory: C0 (Quanta Computer (Subsidiary = Tech Com) China)

Nice new LB... Geekbench was run. Got the errors that others got under 10.5.3 which is normal and to be expected for that version. NO panic!!! So its good. I have a good running machine at this stage.

If you get the panic, please run your serial number at the website as shown earlier in the thread to check your MFR batch number (year/week). I can attest mine for March has the issue resolved.

My machine also arrived yesterday and has the exact same date and factory info you posted above. The fact you seem to have zero issues is comforting. It came with 10.15.2 installed which I found strange (?) so I've decided to just use it and stop worrying over all these issues. I've also decided I'm not going to bother updating my machine's OS until 10.15.4 is released as 10.15.3 looks to be a pig with the 16" MBP even though in the release notes part of the update was to address some issue with this machine.

 

[Norbert Mikołajczyk]

I did G5 Metal test using Intel chipset, lots of errors in console but no kernel panic.

 

[StolenPhone]

This is 'philux' from apple forums. My 3rd machine was a board replacement, bringing me to my 4th. The 4th is exhibiting the same issues with absolutely nothing plugged in and bare minimum software.

Hi philux... welcome to where we can answer you without fear and retribution from Apple moderating comments.

Are you getting your hardware from Apple or a fixit location? You probably need to go straight to apple. Check your serial number in the website listed earlier and find out the manufacturing date. What you need to do (which has been moderated at Apple) is escalate the issue to the executive team. You do this by emailing Tara Bunch - head of Apple care (tbunch@apple.com) or Tim Cook (tcook@apple.com). Explain to them your issue and how disappointed you are. Those emails get caught by the executive team and then they begin getting things done for you. Be sure you list your case numbers at it helps them with your history. This is how I ended up with a brand new machine.
[automerge]1583332287[/automerge]

I did G5 Metal test using Intel chipset, lots of errors in console but no kernel panic.

You are good to go... no hardware issues and just ned the new Catalina version.
[automerge]1583332370[/automerge]

My machine also arrived yesterday and has the exact same date and factory info you posted above. The fact you seem to have zero issues is comforting. It came with 10.15.2 installed which I found strange (?) so I've decided to just use it and stop worrying over all these issues. I've also decided I'm not going to bother updating my machine's OS until 10.15.4 is released as 10.15.3 looks to be a pig with the 16" MBP even though in the release notes part of the update was to address some issue with this machine.

Yep... good to go. 10.15.3 is fine... I updated mine with it. I ran Cat 10.15.4 Beta 3 too and the errors go away.

I'm glad you got a new one and no defects.

 

[insited]

StolenPhone I'll give this a try after this latest attempt. I said on the phone to Apple that by them admitting the product had a major fault out the box by repairing, they've denied an Australian consumer their rights (repair/replace/refund). So now they've gotten nasty and their legal team is calling me tomorrow apparently to discuss if I am within my right for a replacement or refund. I want a replacement it can't be that hard. Logic board number 4 and now they are threatening with their legal team.

I spoke with the ACCC after that and they advised me that if Apple refuse to do anything I'll have to take them to small claims court ($67 fee) and most likely win. The fact Apple have placed me here for a product that's been faulty out the box and now on my 4th faulty one is just totally insane, their legal dept. and culture of blaming their users is totally out of control.

Crazy company.

 

[Brian Y]

Just to note, I've done some hardware level work with AMD GPUs in the past - this error (the panic around PowerPlay) won't be caused by the intel chip - that's a red herring.

PowerPlay is AMD's power management technology. This, almost certainly, looks to be a driver issue, which would explain why multiple MLB swaps does nothing.

 

[insited]

I have the model advertised on their website that their sales pitch is based around, mine doesn't work as advertised. It freezes launching final cut pro x, how many are experiencing that? My experience should be consistent with everyone else's if what you say is true.

https://www.apple.com/au/macbook-pro-16/

 

[StolenPhone]

Just to note, I've done some hardware level work with AMD GPUs in the past - this error (the panic around PowerPlay) won't be caused by the intel chip - that's a red herring.

PowerPlay is AMD's power management technology. This, almost certainly, looks to be a driver issue, which would explain why multiple MLB swaps does nothing.

Not true and not the panic. Read above.

I could run the machine on the AMD processor only and no panic whatsoever if I didn't let it sleep. Only on Intel would it happen. My chats with Apple confirmed it as well.

Although yes, the panic is a part of AMD's driver (Powerplay). But I believe it has to do with the switch over from Intel, i.e. the driver attempting to believe whether it should go to AMD or not.

Also, Apple confirmed it was a bad batch of LBs. So far its Nov/Dec showing the issue. Its not a driver or everyone would have the problem... not a select few.
[automerge]1583339570[/automerge]

StolenPhone I'll give this a try after this latest attempt. I said on the phone to Apple that by them admitting the product had a major fault out the box by repairing, they've denied an Australian consumer their rights (repair/replace/refund). So now they've gotten nasty and their legal team is calling me tomorrow apparently to discuss if I am within my right for a replacement or refund. I want a replacement it can't be that hard. Logic board number 4 and now they are threatening with their legal team.

I spoke with the ACCC after that and they advised me that if Apple refuse to do anything I'll have to take them to small claims court ($67 fee) and most likely win. The fact Apple have placed me here for a product that's been faulty out the box and now on my 4th faulty one is just totally insane, their legal dept. and culture of blaming their users is totally out of control.

Crazy company.

I really don't think you need to go that far. Please try what I said above. The executive team appears to bend over backwards to get you fixed.

You may also want to try the Catalina 10.15.4 beta 3 and see if it helps you. As I said, if you get the panic in Geekbench, its defective. If its just slow with errors in Geekbench, then its fine and you should try the beta to see if your issues are resolved.

 

[gazwas]

You may also want to try the Catalina 10.15.4 beta 3 and see if it helps you. As I said, if you get the panic in Geekbench, its defective. If its just slow with errors in Geekbench, then its fine and you should try the beta to see if your issues are resolved.

Beta 4 was released yesterday. Have you tried the latest version yet?

 

[multidima]

Hi! I put Catalina 10.15.4 beta version 4
Nothing changed
I did not have a panic, but I experience constant freezes

Geekbench5

Also constant errors in the console
virtual void IOAccelEventMachineFast2::checkGPUProgress() - Signaling hardware error on channel 8..

void IOAccelEventMachine2::signalHardwareError(eRestartRequest, int32_t): GPURestartSignaled stampIdx=2 type=2 prevType=0 numStamps=64

void IOAccelFenceMachine::fence_timeout(IOTimerEventSource *): IntelAccelerator prodding blockFenceInterrupt

IntelAccelerator driver returned kIOReturnNotReady for transaction [ID=8900, IOSurfaceID=5]

IntelAccelerator previous NotReady transaction [ID=8900, IOSurfaceID=5] sent. kr=0x0

 

[visualinventor]

That is not true, i have a base build i9, 16GB 1TB, 5500M 4GB with freezes, kernel panics. And with the geekbench test i got the same errors with Intel GPU. Also low geekbench score. I have 3 days left to return....

Yep you're right turns out my Macbook was a base model as well. We have a similar model - mine has 32gb ram.
[automerge]1583416725[/automerge]

@visualinventor Is there any chance this is a Geekbench 5 bug?

Have you ran Geekbench 4 and got the same errors?

I ran GB5 on my 2017 Macbook pro and there wasn't the same issues and the GeekBench score was 4000+. I don't believe it's Geekbench but I will test more in depth tonight when I get my machine back. You can refer back to that post on the apple forums. It could be Catalina since my 2017 was on Mojave (but I kind of doubt it).

https://discussions.apple.com/thread/250905859?answerId=252207204022#252207204022

[automerge]1583417036[/automerge]

Hi! I put Catalina 10.15.4 beta version 4
Nothing changed
I did not have a panic, but I experience constant freezes

Geekbench5
View attachment 897404

Also constant errors in the console
virtual void IOAccelEventMachineFast2::checkGPUProgress() - Signaling hardware error on channel 8..

void IOAccelEventMachine2::signalHardwareError(eRestartRequest, int32_t): GPURestartSignaled stampIdx=2 type=2 prevType=0 numStamps=64

This is the error/reason I took mine into Apple. I have Applecare - so it was covered. It appears to me in the logs, the GPU is stuck in a reboot loop of some kind. Thats what I was seeing anyway. Go through your activity logs and see how many restarts you see when you run the test.

 

[visualinventor]

There seems to be a lot of ‘disproving’ going on in this thread. I will find out tonight what Apple did to fix my machine but I am guessing it will be the exact same fix as @StolenPhone - a new Logicboard.

I have a MBP 2012 retina and I ran the GB5 test on it. It was ~1700.
Bottom line is, if you have a New MacBook Pro 16” and it gets a similar intel gpu metal score as a 2012 MBP Retina then there is something wrong. There is no real debate to be had there. This is what I told the Apple store and they conceded that point, no questions asked.

Your mileage may vary with the solution to the fix, but for a brand new machine we should be looking at a much higher benchmark than that of a MBP 2017 or of a MBP Retina 2012.

FWIW I would expect the ~4800 intel metal gpu score that someone posted earlier in the thread, to be what it is “supposed to be”. That would be an 800pt increase from my 2017 MBP.