Power, Energy and Speed of Embedded and Server Multi-Cores applied to Distributed Simulation of Spiking Neural Networks: ARM in NVIDIA Tegra vs Intel Xeon quad-cores

This short note regards a comparison of instantaneous power, total energy consumption, execution time and energetic cost per synaptic event of a spiking neural network simulator (DPSNN-STDP) distributed on MPI processes when executed either on an embedded platform (based on a dual-socket quad-core ARM platform) or a server platform (INTELbased quad-core dual-socket platform). We also compare the measuer with those reported by leading custom and semi-custom designs: TrueNorth and SPiNNaker. In summary, we observed that: 1we spent 2.2 micro-Joule per simulated synaptic event on the “embedded platform”, approx. 4.4 times lower than what was spent by the “server platform”; 2the instantaneous power consumption of the “embedded platform” was 14.4 times better than the “server” one; 3the server platform is a factor 3.3 faster. The “embedded platform” is made of NVIDIA Jetson TK1 boards, interconnected by Ethernet, each mounting a Tegra K1 chip including a quad-core ARM [email protected]. The “server platform” is based on nodes which are dual-socket, quad-core Intel Xeon CPUs ([email protected]). The measures were obtained with the DPSNN-STDP simulator (Distributed Simulation of Polychronous Spiking Neural Network with synaptic Spike-Timing Dependent Plasticity) developed by INFN, that already proved its efficient scalability and execution speed-up on hundreds of similar “server” cores and MPI processes, applied to neural nets composed of several billions of synapses.