Benchmarking Learning Efficiency in Deep Reservoir Computing