macOS 15 Sequoia on Unsupported Macs Thread

[Sven G]

Same issue when trying to reproduce with a new user on the same machine?

Yes, exactly the same crash in a newly created admin test account: so, it looks like it’s something system-wide…

 

[Sven G]

… Beginning of the Motion crash log (main user):

-------------------------------------
Translated Report (Full Report Below)
-------------------------------------

Process: Motion [3796]
Path: /Applications/Motion.app/Contents/MacOS/Motion
Identifier: com.apple.motionapp
Version: 5.11 (442089)
Build Info: Motion-44000002090000000~8 (3B196c)
App Item ID: 434290957
App External ID: 878206477
Code Type: X86-64 (Native)
Parent Process: launchd [1]
User ID: 501

Date/Time: 2025-10-27 14:56:36.0367 +0100
OS Version: macOS 15.7.1 (24G231)
Report Version: 12
Anonymous UUID: […]


Time Awake Since Boot: 1100 seconds

System Integrity Protection: enabled

Crashed Thread: 6 Dispatch queue: com.apple.helium.rq.gpu1.ru0.gcd

Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: UNKNOWN_0xD at 0x0000000000000000
Exception Codes: 0x000000000000000d, 0x0000000000000000

Termination Reason: Namespace SIGNAL, Code 11 Segmentation fault: 11
Terminating Process: exc handler [3796]

VM Region Info: 0 is not in any region. Bytes before following region: 4473491456
REGION TYPE START - END [ VSIZE] PRT/MAX SHRMOD REGION DETAIL
UNUSED SPACE AT START
--->
__TEXT 10aa41000-10aa7a000 [ 228K] r-x/r-x SM=COW /Applications/Motion.app/Contents/MacOS/Motion

 

[amaze1499]

… Beginning of the Motion crash log (main user):

Long story short:

We are running on a 2012/13 Intel with OCLP:

• MetalOld.dylib, CoreImageOld.dylib, CoreWLANOld.dylib, plus AppleIntelHD4000GraphicsMTLDriver.
• We are using OCLP’s compatibility layers to expose modern Metal on HD4000 (officially unsupported).
  Motion 5.11 + Helium 3.1.0 use compute pipelines; the shim/driver combo borks the encoder lifecycle and the debug layer calls abort().
• Earlier trace also showed: CoreFoundation: doesNotRecognizeSelector:
  Helium: HGMetalFunctionCache::_createFunctionWithSource(...)

Newer Helium code path hitting unsupported/buggy Metal on HD4000. On Sequoia 15.7.2 + OCLP + HD4000, Motion 5.11’s Metal compute path is not stable. No amount of cache cleaning will fix a driver/shim API mismatch or a resource-lifetime bug in the compatibility layer.

Suggestion: Upgrade your hardware or wait for OCLP to fix it.

 

[Sven G]

^^^ And, as I mostly use my old MBP11,3 in clamshell mode with an external display, this evidently happens also with its Nvidia Kepler dGPU, which was in use in this case. Anyway, not an urgent issue (which, BTW, I discovered only because I reinstalled the software): if it is easily fixable by the OCLP devs, good; if not, it can definitely wait…