arXiv subjects

Tosiron Adegbija

Publications and source records attributed to Tosiron Adegbija.

At least 19 recordsLinked to original sources

arXiv2026

Surrogates, Spikes, and Sparsity: Performance Analysis and Characterization of SNN Hyperparameters on Hardware

Spiking Neural Networks (SNNs) offer inherent advantages for low-power inference through sparse, event-driven computation. However, the theoretical energy benefits of SNNs are often decoupled from real hardware performance due to the opaque relationship between training-time choices and inference-time sparsity. While prior work has focused on weight pruning and compression, the role of training hyperparameters -- specifically surrogate gradient functions and neuron model configurations -- in shaping hardware-level activation sparsity remains underexplored. This paper presents a workload characterization study quantifying the sensitivity of hardware latency to SNN hyperparameters. We decouple the impact of surrogate gradient functions (e.g., Fast Sigmoid, Spike Rate Escape) and neuron models (LIF, Lapicque) on classification accuracy and inference efficiency across three event-based vision datasets: DVS128-Gesture, N-MNIST, and DVS-CIFAR10. Our analysis reveals that standard accuracy metrics are poor predictors of hardware efficiency. While Fast Sigmoid achieves the highest accuracy on DVS-CIFAR10, Spike Rate Escape reduces inference latency by up to 12.2% on DVS128-Gesture with minimal accuracy trade-offs. We also demonstrate that neuron model selection is as critical as parameter tuning; transitioning from LIF to Lapicque neurons yields up to 28% latency reduction. We validate on a custom cycle-accurate FPGA-based SNN instrumentation platform, showing that sparsity-aware hyperparameter selection can improve accuracy by 9.1% and latency by over 2x compared to baselines. These findings establish a methodology for predicting hardware behavior from training parameters. The RTL and reproducibility artifacts are at https://zenodo.org/records/18893738.

cs.AR↗

arXiv2026

Device-Circuit Co-Design of Variation-Resilient Read and Write Drivers for Antiferromagnetic Tunnel Junction (AFMTJ) Memories

Antiferromagnetic Tunnel Junctions (AFMTJs) offer picosecond switching and high integration density for in-memory computing, but their ultrafast dynamics and low tunnel magnetoresistance (TMR) make state-of-the-art MRAM interfaces unreliable. This work develops a device-circuit co-designed read/write interface optimized for AFMTJ behavior. Using a calibrated SPICE AFMTJ model as a baseline, we identify the limitations of conventional drivers and propose an asymmetric pulse driver (PD) for deterministic picosecond switching and a self-timed sense amplifier (STSA) with dynamic trip-point tuning for low-TMR sensing. Our experiments using SPICE and Monte Carlo evaluations demonstrate that the proposed circuits preserve AFMTJ latency and energy benefits while achieving robust read/write yield under realistic PVT and 3D integration parasitics, outperforming standard MRAM front-ends under the same conditions.

cs.AR↗

arXiv2026

Accelerating Post-Quantum Cryptography via LLM-Driven Hardware-Software Co-Design

Post-quantum cryptography (PQC) is crucial for securing data against emerging quantum threats. However, its algorithms are computationally complex and difficult to implement efficiently on hardware. In this paper, we explore the potential of Large Language Models (LLMs) to accelerate the hardware-software co-design process for PQC, with a focus on the FALCON digital signature scheme. We present a novel framework that leverages LLMs to analyze PQC algorithms, identify performance-critical components, and generate candidate hardware descriptions for FPGA implementation. We present the first quantitative comparison between LLM-driven synthesis and conventional HLS-based approaches for low-level compute-intensive kernels in FALCON, showing that human-in-the-loop LLM-generated accelerators can achieve up to 2.6x speedup in kernel execution time with shorter critical paths, while highlighting trade-offs in resource utilization and power consumption. Our results suggest that LLMs can minimize design effort and development time by automating FPGA accelerator design iterations for PQC algorithms, offering a promising new direction for rapid and adaptive PQC accelerator design on FPGAs.

cs.AR↗

arXiv2026

Antiferromagnetic Tunnel Junctions (AFMTJs) for In-Memory Computing: Modeling and Case Study

Antiferromagnetic Tunnel Junctions (AFMTJs) enable picosecond switching and femtojoule writes through ultrafast sublattice dynamics. We present the first end-to-end AFMTJ simulation framework integrating multi-sublattice Landau-Lifshitz-Gilbert (LLG) dynamics with circuit-level modeling. SPICE-based simulations show that AFMTJs achieve ~8x lower write latency and ~9x lower write energy than conventional MTJs. When integrated into an in-memory computing architecture, AFMTJs deliver 17.5x average speedup and nearly 20x energy savings versus a CPU baseline-significantly outperforming MTJ-based IMC. These results establish AFMTJs as a compelling primitive for scalable, low-power computing.

cs.AR↗

arXiv2026

Spiking Neural Network Architecture Search: A Survey

This survey paper presents a comprehensive examination of Spiking Neural Network (SNN) architecture search (SNNaS) from a unique hardware/software co-design perspective. SNNs, inspired by biological neurons, have emerged as a promising approach to neuromorphic computing. They offer significant advantages in terms of power efficiency and real-time resource-constrained processing, making them ideal for edge computing and IoT applications. However, designing optimal SNN architectures poses significant challenges, due to their inherent complexity (e.g., with respect to training) and the interplay between hardware constraints and SNN models. We begin by providing an overview of SNNs, emphasizing their operational principles and key distinctions from traditional artificial neural networks (ANNs). We then provide a brief overview of the state of the art in NAS for ANNs, highlighting the challenges of directly applying these approaches to SNNs. We then survey the state of the art in SNN-specific NAS approaches. Finally, we conclude with insights into future research directions for SNN research, emphasizing the potential of hardware/software co-design in unlocking the full capabilities of SNNs. This survey aims to serve as a valuable resource for researchers and practitioners in the field, offering a holistic view of SNNaS and underscoring the importance of a co-design approach to harness the true potential of neuromorphic computing.

cs.NE↗

arXiv2025

Deep Inverse Design for High-Level Synthesis

High-level synthesis (HLS) has significantly advanced the automation of digital circuits design, yet the need for expertise and time in pragma tuning remains challenging. Existing solutions for the design space exploration (DSE) adopt either heuristic methods, lacking essential information for further optimization potential, or predictive models, missing sufficient generalization due to the time-consuming nature of HLS and the exponential growth of the design space. To address these challenges, we propose Deep Inverse Design for HLS (DID4HLS), a novel approach that integrates graph neural networks and generative models. DID4HLS iteratively optimizes hardware designs aimed at compute-intensive algorithms by learning conditional distributions of design features from post-HLS data. Compared to four state-of-the-art DSE baselines, our method achieved an average improvement of 42.8% on average distance to reference set (ADRS) compared to the best-performing baselines across six benchmarks, while demonstrating high robustness and efficiency. The code is available at https://github.com/PingChang818/DID4HLS.

cs.AR↗

arXiv2024

Exploring the Sparsity-Quantization Interplay on a Novel Hybrid SNN Event-Driven Architecture

Spiking Neural Networks (SNNs) offer potential advantages in energy efficiency but currently trail Artificial Neural Networks (ANNs) in versatility, largely due to challenges in efficient input encoding. Recent work shows that direct coding achieves superior accuracy with fewer timesteps than traditional rate coding. However, there is a lack of specialized hardware to fully exploit the potential of direct-coded SNNs, especially their mix of dense and sparse layers. This work proposes the first hybrid inference architecture for direct-coded SNNs. The proposed hardware architecture comprises a dense core to efficiently process the input layer and sparse cores optimized for event-driven spiking convolutions. Furthermore, for the first time, we investigate and quantify the quantization effect on sparsity. Our experiments on two variations of the VGG9 network and implemented on a Xilinx Virtex UltraScale+ FPGA (Field-Programmable Gate Array) reveal two novel findings. Firstly, quantization increases the network sparsity by up to 15.2% with minimal loss of accuracy. Combined with the inherent low power benefits, this leads to a 3.4x improvement in energy compared to the full-precision version. Secondly, direct coding outperforms rate coding, achieving a 10% improvement in accuracy and consuming 26.4x less energy per image. Overall, our accelerator achieves 51x higher throughput and consumes half the power compared to previous work. Our accelerator code is available at: https://github.com/githubofaliyev/SNN-DSE/tree/DATE25

cs.AR↗

arXiv2024

System-Level Design Space Exploration for High-Level Synthesis under End-to-End Latency Constraints

Many modern embedded systems have end-to-end (EtoE) latency constraints that necessitate precise timing to ensure high reliability and functional correctness. The combination of High-Level Synthesis (HLS) and Design Space Exploration (DSE) enables the rapid generation of embedded systems using various constraints/directives to find Pareto-optimal configurations. Current HLS DSE approaches often address latency by focusing on individual components, without considering the EtoE latency during the system-level optimization process. However, to truly optimize the system under EtoE latency, we need a holistic approach that analyzes individual system components' timing constraints in the context of how the different components interact and impact the overall design. This paper presents a novel system-level HLS DSE approach, called EtoE-DSE, that accommodates EtoE latency and variable timing constraints for complex multi-component application-specific embedded systems. EtoE-DSE employs a latency estimation model and a pathfinding algorithm to identify and estimate the EtoE latency for paths between any endpoints. It also uses a frequency-based segmentation process to segment and prune the design space, alongside a latency-constrained optimization algorithm for efficiently and accurately exploring the system-level design space. We evaluate our approach using a real-world use case of an autonomous driving subsystem compared to the state-of-the-art in HLS DSE. We show that our approach yields substantially better optimization results than prior DSE approaches, improving the quality of results by up to 89.26%, while efficiently identifying Pareto-optimal configurations in terms of energy and area.

cs.AR↗

arXiv2024

Sparsity-Aware Hardware-Software Co-Design of Spiking Neural Networks: An Overview

Spiking Neural Networks (SNNs) are inspired by the sparse and event-driven nature of biological neural processing, and offer the potential for ultra-low-power artificial intelligence. However, realizing their efficiency benefits requires specialized hardware and a co-design approach that effectively leverages sparsity. We explore the hardware-software co-design of sparse SNNs, examining how sparsity representation, hardware architectures, and training techniques influence hardware efficiency. We analyze the impact of static and dynamic sparsity, discuss the implications of different neuron models and encoding schemes, and investigate the need for adaptability in hardware designs. Our work aims to illuminate the path towards embedded neuromorphic systems that fully exploit the computational advantages of sparse SNNs.

cs.AR↗

arXiv2024

Are LLMs Any Good for High-Level Synthesis?

The increasing complexity and demand for faster, energy-efficient hardware designs necessitate innovative High-Level Synthesis (HLS) methodologies. This paper explores the potential of Large Language Models (LLMs) to streamline or replace the HLS process, leveraging their ability to understand natural language specifications and refactor code. We survey the current research and conduct experiments comparing Verilog designs generated by a standard HLS tool (Vitis HLS) with those produced by LLMs translating C code or natural language specifications. Our evaluation focuses on quantifying the impact on performance, power, and resource utilization, providing an assessment of the efficiency of LLM-based approaches. This study aims to illuminate the role of LLMs in HLS, identifying promising directions for optimized hardware design in applications such as AI acceleration, embedded systems, and high-performance computing.

cs.AR↗

arXiv2024

CHIME: Energy-Efficient STT-RAM-based Concurrent Hierarchical In-Memory Processing

Processing-in-cache (PiC) and Processing-in-memory (PiM) architectures, especially those utilizing bit-line computing, offer promising solutions to mitigate data movement bottlenecks within the memory hierarchy. While previous studies have explored the integration of compute units within individual memory levels, the complexity and potential overheads associated with these designs have often limited their capabilities. This paper introduces a novel PiC/PiM architecture, Concurrent Hierarchical In-Memory Processing (CHIME), which strategically incorporates heterogeneous compute units across multiple levels of the memory hierarchy. This design targets the efficient execution of diverse, domain-specific workloads by placing computations closest to the data where it optimizes performance, energy consumption, data movement costs, and area. CHIME employs STT-RAM due to its various advantages in PiC/PiM computing, such as high density, low leakage, and better resiliency to data corruption from activating multiple word lines. We demonstrate that CHIME enhances concurrency and improves compute unit utilization at each level of the memory hierarchy. We present strategies for exploring the design space, grouping, and placing the compute units across the memory hierarchy. Experiments reveal that, compared to the state-of-the-art bit-line computing approaches, CHIME achieves significant speedup and energy savings of 57.95% and 78.23% for various domain-specific workloads, while reducing the overheads associated with single-level compute designs.

cs.CY↗

arXiv2024

STT-RAM-based Hierarchical In-Memory Computing

In-memory computing promises to overcome the von Neumann bottleneck in computer systems by performing computations directly within the memory. Previous research has suggested using Spin-Transfer Torque RAM (STT-RAM) for in-memory computing due to its non-volatility, low leakage power, high density, endurance, and commercial viability. This paper explores hierarchical in-memory computing, where different levels of the memory hierarchy are augmented with processing elements to optimize workload execution. The paper investigates processing in memory (PiM) using non-volatile STT-RAM and processing in cache (PiC) using volatile STT-RAM with relaxed retention, which helps mitigate STT-RAM's write latency and energy overheads. We analyze tradeoffs and overheads associated with data movement for PiC versus write overheads for PiM using STT-RAMs for various workloads. We examine workload characteristics, such as computational intensity and CPU-dependent workloads with limited instruction-level parallelism, and their impact on PiC/PiM tradeoffs. Using these workloads, we evaluate computing in STT-RAM versus SRAM at different cache hierarchy levels and explore the potential of heterogeneous STT-RAM cache architectures with various retention times for PiC and CPU-based computing. Our experiments reveal significant advantages of STT-RAM-based PiC over PiM for specific workloads. Finally, we describe open research problems in hierarchical in-memory computing architectures to further enhance this paradigm.

cs.CY↗

arXiv2024

SCART: Predicting STT-RAM Cache Retention Times Using Machine Learning

Prior studies have shown that the retention time of the non-volatile spin-transfer torque RAM (STT-RAM) can be relaxed in order to reduce STT-RAM's write energy and latency. However, since different applications may require different retention times, STT-RAM retention times must be critically explored to satisfy various applications' needs. This process can be challenging due to exploration overhead, and exacerbated by the fact that STT-RAM caches are emerging and are not readily available for design time exploration. This paper explores using known and easily obtainable statistics (e.g., SRAM statistics) to predict the appropriate STT-RAM retention times, in order to minimize exploration overhead. We propose an STT-RAM Cache Retention Time (SCART) model, which utilizes machine learning to enable design time or runtime prediction of right-provisioned STT-RAM retention times for latency or energy optimization. Experimental results show that, on average, SCART can reduce the latency and energy by 20.34% and 29.12%, respectively, compared to a homogeneous retention time while reducing the exploration overheads by 52.58% compared to prior work.

cs.CY↗

arXiv2024

ARC: DVFS-Aware Asymmetric-Retention STT-RAM Caches for Energy-Efficient Multicore Processors

Relaxed retention (or volatile) spin-transfer torque RAM (STT-RAM) has been widely studied as a way to reduce STT-RAM's write energy and latency overheads. Given a relaxed retention time STT-RAM level one (L1) cache, we analyze the impacts of dynamic voltage and frequency scaling (DVFS) -- a common optimization in modern processors -- on STT-RAM L1 cache design. Our analysis reveals that, apart from the fact that different applications may require different retention times, the clock frequency, which is typically ignored in most STT-RAM studies, may also significantly impact applications' retention time needs. Based on our findings, we propose an asymmetric-retention core (ARC) design for multicore architectures. ARC features retention time heterogeneity to specialize STT-RAM retention times to applications' needs. We also propose a runtime prediction model to determine the best core on which to run an application, based on the applications' characteristics, their retention time requirements, and available DVFS settings. Results reveal that the proposed approach can reduce the average cache energy by 20.19% and overall processor energy by 7.66%, compared to a homogeneous STT-RAM cache design.

cs.CY↗

arXiv2024

A high-level synthesis approach for precisely-timed, energy-efficient embedded systems

Embedded systems continue to rapidly proliferate in diverse fields, including medical devices, autonomous vehicles, and more generally, the Internet of Things (IoT). Many embedded systems require application-specific hardware components to meet precise timing requirements within limited resource (area and energy) constraints. High-level synthesis (HLS) is an increasingly popular approach for improving the productivity of designing hardware and reducing the time/cost by using high-level languages to specify computational functionality and automatically generate hardware implementations. However, current HLS methods provide limited or no support to incorporate or utilize precise timing specifications within the synthesis and optimization process. In this paper, we present a hybrid high-level synthesis (H-HLS) framework that integrates state-based high-level synthesis (SB-HLS) with performance-driven high-level synthesis (PD-HLS) methods to enable the design and optimization of application-specific embedded systems in which timing information is explicitly and precisely defined in state-based system models. We demonstrate the results achieved by this H-HLS approach using case studies including a wearable pregnancy monitoring device, an ECG-based biometric authentication system, and a synthetic system, and compare the design space exploration results using two PD-HLS tools to show how H-HLS can provide low energy and area under timing constraints.

cs.AR↗

arXiv2024

Skip the Benchmark: Generating System-Level High-Level Synthesis Data using Generative Machine Learning

High-Level Synthesis (HLS) Design Space Exploration (DSE) is a widely accepted approach for efficiently exploring Pareto-optimal and optimal hardware solutions during the HLS process. Several HLS benchmarks and datasets are available for the research community to evaluate their methodologies. Unfortunately, these resources are limited and may not be sufficient for complex, multi-component system-level explorations. Generating new data using existing HLS benchmarks can be cumbersome, given the expertise and time required to effectively generate data for different HLS designs and directives. As a result, synthetic data has been used in prior work to evaluate system-level HLS DSE. However, the fidelity of the synthetic data to real data is often unclear, leading to uncertainty about the quality of system-level HLS DSE. This paper proposes a novel approach, called Vaegan, that employs generative machine learning to generate synthetic data that is robust enough to support complex system-level HLS DSE experiments that would be unattainable with only the currently available data. We explore and adapt a Variational Autoencoder (VAE) and Generative Adversarial Network (GAN) for this task and evaluate our approach using state-of-the-art datasets and metrics. We compare our approach to prior works and show that Vaegan effectively generates synthetic HLS data that closely mirrors the ground truth's distribution.

cs.LG↗

arXiv2024

PULSE: Parametric Hardware Units for Low-power Sparsity-Aware Convolution Engine

Spiking Neural Networks (SNNs) have become popular for their more bio-realistic behavior than Artificial Neural Networks (ANNs). However, effectively leveraging the intrinsic, unstructured sparsity of SNNs in hardware is challenging, especially due to the variability in sparsity across network layers. This variability depends on several factors, including the input dataset, encoding scheme, and neuron model. Most existing SNN accelerators fail to account for the layer-specific workloads of an application (model + dataset), leading to high energy consumption. To address this, we propose a design-time parametric hardware generator that takes layer-wise sparsity and the number of processing elements as inputs and synthesizes the corresponding hardware. The proposed design compresses sparse spike trains using a priority encoder and efficiently shifts the activations across the network's layers. We demonstrate the robustness of our proposed approach by first profiling a given application's characteristics followed by performing efficient resource allocation. Results on a Xilinx Kintex FPGA (Field Programmable Gate Arrays) using MNIST, FashionMNIST, and SVHN datasets show a 3.14x improvement in accelerator efficiency (FPS/W) compared to a sparsity-oblivious systolic array-based accelerator. Compared to the most recent sparsity-aware work, our solution improves efficiency by 1.72x.

cs.AR↗

arXiv2024

Fine-Tuning Surrogate Gradient Learning for Optimal Hardware Performance in Spiking Neural Networks

The highly sparse activations in Spiking Neural Networks (SNNs) can provide tremendous energy efficiency benefits when carefully exploited in hardware. The behavior of sparsity in SNNs is uniquely shaped by the dataset and training hyperparameters. This work reveals novel insights into the impacts of training on hardware performance. Specifically, we explore the trade-offs between model accuracy and hardware efficiency. We focus on three key hyperparameters: surrogate gradient functions, beta, and membrane threshold. Results on an FPGA-based hardware platform show that the fast sigmoid surrogate function yields a lower firing rate with similar accuracy compared to the arctangent surrogate on the SVHN dataset. Furthermore, by cross-sweeping the beta and membrane threshold hyperparameters, we can achieve a 48% reduction in hardware-based inference latency with only 2.88% trade-off in inference accuracy compared to the default setting. Overall, this study highlights the importance of fine-tuning model hyperparameters as crucial for designing efficient SNN hardware accelerators, evidenced by the fine-tuned model achieving a 1.72x improvement in accelerator efficiency (FPS/W) compared to the most recent work.

cs.NE↗