I cannot use threadgroup_barrier into a compute kernel on El Capitan with my MacPro (D500) :
It always crashes the compiler with this message after a call to newComputePipelineStateWithFunction :
>> ERROR: Failed creating a compute kernel: Error Domain=CompilerError Code=1 "Compiler encountered an internal error" UserInfo=XXXXXXXXX {NSLocalizedDescription=Compiler encountered an internal error}
Even a minimal kernel that contains only the barrier instruction produces a bug.
Is someone able to use a barrier into a compute kernel ?