Fast and Flexible Spiking Network Simulations on Massively Distributed On-Chip Memory Accelerators
The paper demonstrates a method to accelerate brain research simulations on commercial systems with large distributed on-chip memory.
The paper presents a novel method that achieves a significant speedup on commercial systems, unlike previous work which required custom-built FPGA systems.
Before reading this…
Applications
- →Brain research
- →Neural networks simulation
To understand this paper, make sure you know these concepts first:
- Knowledge of neural networks and brain researchfind papers →
- Understanding of simulation acceleration techniquesfind papers →
Abstract
More Like ThisStudying brain function and dysfunction in spiking networks with realistic synaptic densities pushes simulation technology beyond its current limits. It requires simulations substantially faster than biological real-time to cover long model times and enable efficient iterative exploration, particularly when connectivity is plastic. The modelling community aspires to a $100\times$ speedup. Yet on conventional many-core systems this remains elusive, bounded by inter-node communication and memory access. So far, significant speedup has only been shown on a custom-built FPGA system. We demonstrate the potential of commercial systems with large distributed on-chip memory to accelerate brain research. To exploit the local SRAM, we introduce a group-based mapping and communication strategy tackling spike-routing challenges on massively distributed systems. On Graphcore IPUs, the cortical microcircuit runs at a record $22.7\times$ faster than biological real-time with negligible build times and a reusable compiled graph. By design, the algorithm makes no model- or hardware-specific assumptions, providing a general strategy for advancing neuronal simulation technology.