Convolutional neural networks (CNNs) are progressively deployed on embedded systems, which is challenging because their computational and energy requirements need to be satisfied by devices with limited resources and power supplies. For instance, they are implemented in the Internet of Things or edge computing, i.e. in applications using low-power and low-performance microcontroller units (MCUs). Monocore MCUs are not tailored to respond to the computational and energy requirements of CNNs because of their limited resources and power supplies, but a multicore MCU could overtake these limitations. This paper experimentally compares three algorithms scheduling CNNs on embedded systems at two different levels (neuron and layer ones) and evaluates their performances in terms of the makespan and energy consumption. The results show that the algorithm called SNN outperforms other two algorithms (STD and STS) and that the scheduling at layer level significantly reduces the energy consumption.