Neuromorphic computing systems are embracing memristors to implement high density and low power synaptic storage as crossbar arrays in hardware. These systems are energy efficient in executing Spiking Neural Networks (SNNs). We observe that long bitlines and wordlines in a memristive crossbar are a major source of parasitic voltage drops, which create current asymmetry. Through circuit simulations, we show the significant endurance variation that results from this asymmetry. Therefore, if the critical memristors (ones with lower endurance) are overutilized, they may lead to a reduction of the crossbar's lifetime. We propose eSpine, a novel technique to improve lifetime by incorporating the endurance variation within each crossbar in mapping machine learning workloads, ensuring that synapses with higher activation are always implemented on memristors with higher endurance, and vice versa. eSpine works in two steps. First, it uses the Kernighan-Lin Graph Partitioning algorithm to partition a workload into clusters of neurons and synapses, where each cluster can fit in a crossbar. Second, it uses an instance of Particle Swarm Optimization (PSO) to map clusters to tiles, where the placement of synapses of a cluster to memristors of a crossbar is performed by analyzing their activation within the workload. We evaluate eSpine for a state-of-the-art neuromorphic hardware model with phase-change memory (PCM)-based memristors. Using 10 SNN workloads, we demonstrate a significant improvement in the effective lifetime.
6 November 2019 in IEEE Computer Architecture Letters
Neuromorphic hardware with non-volatile memory (NVM) can implement machine learning tasks in an energy-efficient manner. Unfortunately, NVMs such as phase change memory (PCM) require high voltages for correct operation, which are supplied from on-chip charge pumps. If a charge pump is activated too frequently, its internal CMOS devices do not recover from stress, accelerating their aging and leading to negative bias temperature instability (NBTI) generated defects in the charge pump. Forcefully discharging a stressed charge pump can lower the aging rate of its CMOS devices, but makes the neuromorphic hardware unavailable to perform computations while its charge pumps are being discharged. This negatively impacts performance such as latency and accuracy. In this paper, we propose a novel framework to exploit workload-specific performance and lifetime trade-offs in neuromorphic computing. Our framework first extracts the precise times at which a charge pump in the hardware is activated to support neural computations within a workload. This timing information is then used with a characterized NBTI reliability model to estimate the charge pump's aging during the workload execution. We use our framework to evaluate workload-specific performance and reliability impacts of using 1) different SNN mapping strategies and 2) different charge pump discharge strategies.
22 September 2020 in IEEE Embedded Systems Letters
With growing model complexity, mapping Spiking Neural Network (SNN)-based applications to tile-based neuromorphic hardware is becoming increasingly challenging. This is because the synaptic storage resources on a tile, viz. a crossbar, can accommodate only a fixed number of pre-synaptic connections per post-synaptic neuron. For complex SNN models that have many pre-synaptic connections per neuron, some connections may need to be pruned after training to fit onto the tile resources, leading to a loss in model quality, e.g., accuracy. In this work, we propose a novel unrolling technique that decomposes a neuron function with many pre-synaptic connections into a sequence of homogeneous neural units, where each neural unit is a function computation node, with two pre-synaptic connections. This spatial decomposition technique significantly improves crossbar utilization and retains all pre-synaptic connections, resulting in no loss of the model quality derived from connection pruning. We integrate the proposed technique within an existing SNN mapping framework and evaluate it using machine learning applications on the DYNAP-SE state-of-the-art neuromorphic hardware. Our results demonstrate an average 60% lower crossbar requirement, 9x higher synapse utilization, 62% lower wasted energy on the hardware, and between 0.8% and 4.6% increase in model quality.
The paper develops a methodology for the online built-in self-testing of deep neural network (DNN) accelerators to validate the correct operation with respect to their functional specifications. The DNN of interest is realized in the hardware to perform in-memory computing using non-volatile memory cells as computational units. Assuming a functional fault model, we develop methods to generate pseudorandom and structured test patterns to detect hardware faults. We also develop a test-sequencing strategy that combines these different classes of tests to achieve high fault coverage. The testing methodology is applied to a broad class of DNNs trained to classify images from the MNIST, Fashion-MNIST, and CIFAR-10 datasets. The goal is to expose hardware faults which may lead to the incorrect classification of images. We achieve an average fault coverage of 94% for these different architectures, some of which are large and complex.
Large-scale neuromorphic architectures consist of computing tiles that communicate spikes using a shared interconnect. We propose ADIONA, a dynamic segmented bus interconnect to address design scalability while reducing energy and latency of spike traffic. A DIONA c onsists o f p arallel bus lanes arranged in a ladder-shaped structure that allows any tile to connect to another, offers multiple routing options for communication links, and provides a high level of customization for different mapping scenarios and use cases. Each lane in the ladder bus is partitioned into segments using lightweight buffer-less switches. Based on compile-time communication information, these switches can be dynamically reconfigured at runtime to execute the target application. Our dynamic segmented bus interconnect substantially enhances hardware utilization, improves fault tolerance, and offers adaptability to execute different applications on a single hardware platform. We evaluate ADIONA using three synthetic and three realistic machine learning workloads on a cycle-accurate neuromorphic simulator. Our results show that ADIONA reduces energy consumption by 2.1×, latency by 40×, and interconnect area by 2×, compared to a state-of-the-art interconnect for neuromorphic systems.