CNTKs roots are from Speech type data which inherently have a notion of time. The architecture of the toolkit support efficient recurrence from ground up. Also

CNTKs roots are from Speech type data which inherently have a notion of time. The architecture of the toolkit support efficient recurrence from ground up. Also the toolkit focusses on handling large production scale data workloads which implies additional engineering efficiencies built into the toolkit. I am a Microsoft employee.