The BrainScaleS-2 multi-chip system: Interconnecting continuous-time neuromorphic compute substrates
By: Joscha Ilmberger, Johannes Schemmel
Potential Business Impact:
Makes brain-like computers learn much faster.
The BrainScaleS-2 SoC integrates analog neuron and synapse circuits with digital periphery, including two CPUs with SIMD extensions. Each ASIC is connected to a Node-FPGA, providing experiment control and Ethernet connectivity. This work details the scaling of the compute substrate through FPGA-based interconnection via an additional Aggregator unit. The Aggregator provides up to 12 transceiver links to a backplane of Node-FPGAs, as well as 4 transceiver lanes for further extension. Two such interconnected backplanes are integrated into a standard 19in rack case with 4U height together with an Ethernet switch, system controller and power supplies. For all spike rates, chip-to-chip latencies -- consisting of four hops across three FPGAs -- below 1.3$μ$s are achieved within each backplane.
Similar Papers
Neuromorphic Processor Employing FPGA Technology with Universal Interconnections
Hardware Architecture
Builds cheap computer brains that learn fast.
SynapticCore-X: A Modular Neural Processing Architecture for Low-Cost FPGA Acceleration
Hardware Architecture
Builds smart computer chips cheaply for AI.
Scenario-Aware Control of Segmented Ladder Bus: Design and FPGA Implementation
Neural and Evolutionary Computing
Makes computer brains work faster and use less power.