V7 vs. v8 anisotropy rotation

I notice some regression in v8 regarding anisotropy rotation:

7.34.23267.11001

8.1.23325.13001

here is the file:

cycles-aniso-v7-vs-v8.3dm (4.2 MB)

thanks for reporting, logged as RH-78939 anisotropic in Rhino 8 behaves different than in Rhino 7