@JeffTutorials We're using more GPTs, as they now run natively on Tesla TRIP chip vs needing to round trip to iGPU