Search papers, labs, and topics across Lattice.
This paper details the design and development of a neuromorphic silicon suite comprising four digital IP blocks tailored for edge computing, including a PVT sensor, a stochastic LIF neuron, an on-chip STDP controller, and a memristive-crossbar controller. These components are implemented in standard-cell CMOS on the SkyWater 130 nm process and are designed to facilitate compact, configurable hardware that integrates probabilistic inference and local learning. The suite achieves efficient performance metrics, including a 50 MHz operational target and low power consumption, while passing extensive verification tests, highlighting its potential for advancing neuromorphic computing applications.
A unified suite of neuromorphic silicon blocks achieves efficient edge processing with low power consumption, paving the way for advanced probabilistic inference in compact hardware.
Edge neuromorphic systems need compact, configurable hardware that combines probabilistic inference, local learning, and an interface to emerging analogue memory. We present four interface-compatible digital IP blocks implemented as standard-cell CMOS on the SkyWater 130 nm process: a process, voltage and temperature (PVT) sensor built from five selectable ring oscillators that also provides a jitter-based true-random-number generator and a frequency-bounds health monitor; a stochastic leaky integrate-and-fire (LIF) neuron with a configurable LFSR, a programmable activation table, and a refractory period; an on-chip spike-timing-dependent plasticity (STDP) controller with a programmable curve and reward-modulated, eligibility-trace, and anti-Hebbian modes; and a memristive-crossbar controller supporting forming, set, reset, read, and automated current-voltage sweep with current-compliance limiting and half-select biasing. All four blocks share a common serial peripheral interface (SPI) register file; the sensor also exposes a parallel readout. Each occupies a single tile at a 50 MHz target. The suite was verified with 99 cocotb tests at register-transfer and gate level (all passing) and taken through an open standard-cell flow, then submitted for tapeout via the Tiny Tapeout shared-silicon programme. Mapped to the open cell library, each block occupies a post-synthesis cell area of 9.3 to 10.6 thousand square micrometres, places at 61 to 70 per cent tile utilisation, meets the 50 MHz constraint with positive setup and hold margin after clock-tree synthesis, and draws an estimated 0.64 to 0.70 mW under a default switching-activity assumption. The contribution is a coherent, openly released set of building blocks unified by one register interface and one verification flow. All results are from simulation and the implementation flow; no fabricated silicon is reported.
