Performance Analysis of Edge and In-Sensor AI Processors: A Comparative Review

• URL: http://arxiv.org/abs/2603.08725v1
• Date: Wed, 18 Feb 2026 07:23:55 GMT
• Title: Performance Analysis of Edge and In-Sensor AI Processors: A Comparative Review
• Abstract summary: This review examines the rapidly evolving landscape of ultra-low-power edge processors.<n>We categorize commercially available and research-grade platforms according to their compute paradigms, power envelopes, and memory hierarchies.<n>We benchmark a 336 million Multiply-Accumulate (MAC) segmentation model (PicoSAM2) on three representative processors.
• Score: 9.350342235384081
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: This review examines the rapidly evolving landscape of ultra-low-power edge processors, covering heterogeneous Systems-on-Chips (SoCs), neural accelerators, near-sensor and in-sensor architectures, and emerging dataflow and memory-centric designs. We categorize commercially available and research-grade platforms according to their compute paradigms, power envelopes, and memory hierarchies, and analyze their suitability for always-on and latency-critical Artificial Intelligence (AI) workloads. To complement the architectural overview with empirical evidence, we benchmark a 336 million Multiply-Accumulate (MAC) segmentation model (PicoSAM2) on three representative processors: GAP9, leveraging a multi-core RISC-V architecture augmented with hardware accelerators; the STM32N6, which pairs an advanced ARM Cortex-M55 core with a dedicated neural architecture accelerator; and the Sony IMX500, representing in-sensor stacked-Complementary Metal-Oxide-Semiconductor (CMOS) compute. Collectively, these platforms span MCU-class, embedded neural accelerator, and in-sensor paradigms. The evaluation reports latency, inference efficiency, energy efficiency, and energy-delay product. The results show a clear divergence in hardware behavior, with the IMX500 achieving the highest utilization (86.2 MAC/cycle) and the lowest energy-delay product, highlighting the growing significance and technological maturity of in-sensor processing. GAP9 offers the best energy efficiency within microcontroller-class power budgets, and the STM32N6 provides the lowest raw latency at a significantly higher energy cost. Together, the review and benchmarks provide a unified view of the current design directions and practical trade-offs that are shaping the next generation of ultra-low-power and in-sensor AI processors.

Related papers

• KScaNN: Scalable Approximate Nearest Neighbor Search on Kunpeng [46.35664429179457]
  A naive port of existing x86 ANNS algorithms to ARM platforms results in a substantial performance deficit.<n>We introduce KScaNN, a novel ANNS algorithm co-designed for the Kunpeng 920 ARM architecture.
  arXiv  Detail & Related papers  (2025-11-05T09:01:32Z)
• Energy efficiency analysis of Spiking Neural Networks for space applications [43.91307921405309]
  Spiking Neural Networks (SNN) are highly attractive due to their theoretically superior energy efficiency.<n>This work presents a numerical analysis and comparison of different SNN techniques applied to scene classification for the EuroSAT dataset.
  arXiv  Detail & Related papers  (2025-05-16T16:29:50Z)
• Ultra-Low-Power Spiking Neurons in 7 nm FinFET Technology: A Comparative Analysis of Leaky Integrate-and-Fire, Morris-Lecar, and Axon-Hillock Architectures [0.0]
  Neuromorphic computing aims to replicate the brain's remarkable energy efficiency and parallel processing capabilities for large-scale artificial intelligence applications.<n>We present a comprehensive study of three spiking neuron circuit architectures-Leaky--and-Fire (LIF), Morris-Lecar (ML), and Axon-Hillock (AH)-implemented in a 7 nm FinFET technology.<n>Our results show that the AH design achieves the highest throughput, demonstrating multi-gigahertz firing rates (up to 3 GHz) with attojoule energy costs.
  arXiv  Detail & Related papers  (2025-04-21T03:06:39Z)
• Exploring energy consumption of AI frameworks on a 64-core RV64 Server CPU [1.1134856914044027]
  In today's era of rapid technological advancement, artificial intelligence (AI) applications require large-scale, high-performance, and data-intensive computations, leading to significant energy demands.<n> Hardware manufacturers are developing new, efficient, and specialized solutions, with the RISC-V architecture emerging as a prominent player.<n>Software developers are creating new algorithms and frameworks, yet their energy efficiency often remains unclear.
  arXiv  Detail & Related papers  (2025-04-03T08:27:10Z)
• Co-design of a novel CMOS highly parallel, low-power, multi-chip neural network accelerator [0.0]
  We present the NV-1, a new low-power ASIC AI processor that greatly accelerates parallel processing (> 10X) with dramatic reduction in energy consumption. The resulting device is currently being used in a fielded edge sensor application.
  arXiv  Detail & Related papers  (2024-09-28T15:47:16Z)
• Enhancing Dropout-based Bayesian Neural Networks with Multi-Exit on FPGA [20.629635991749808]
  This paper proposes an algorithm and hardware co-design framework that can generate field-programmable gate array (FPGA)-based accelerators for efficient BayesNNs. At the algorithm level, we propose novel multi-exit dropout-based BayesNNs with reduced computational and memory overheads. At the hardware level, this paper introduces a transformation framework that can generate FPGA-based accelerators for the proposed efficient BayesNNs.
  arXiv  Detail & Related papers  (2024-06-20T17:08:42Z)
• TeMPO: Efficient Time-Multiplexed Dynamic Photonic Tensor Core for Edge AI with Compact Slow-Light Electro-Optic Modulator [44.74560543672329]
  We present a time-multiplexed dynamic photonic tensor accelerator, dubbed TeMPO, with cross-layer device/circuit/architecture customization. We achieve a 368.6 TOPS peak performance, 22.3 TOPS/W energy efficiency, and 1.2 TOPS/mm$2$ compute density. This work signifies the power of cross-layer co-design and domain-specific customization, paving the way for future electronic-photonic accelerators.
  arXiv  Detail & Related papers  (2024-02-12T03:40:32Z)
• EPIM: Efficient Processing-In-Memory Accelerators based on Epitome [78.79382890789607]
  We introduce the Epitome, a lightweight neural operator offering convolution-like functionality. On the software side, we evaluate epitomes' latency and energy on PIM accelerators. We introduce a PIM-aware layer-wise design method to enhance their hardware efficiency.
  arXiv  Detail & Related papers  (2023-11-12T17:56:39Z)
• Evaluating Emerging AI/ML Accelerators: IPU, RDU, and NVIDIA/AMD GPUs [14.397623940689487]
  Graphcore Intelligence Processing Unit (IPU), Sambanova Reconfigurable Dataflow Unit (RDU), and enhanced GPU platforms are reviewed. This research provides a preliminary evaluation and comparison of these commercial AI/ML accelerators.
  arXiv  Detail & Related papers  (2023-11-08T01:06:25Z)
• FPGA-optimized Hardware acceleration for Spiking Neural Networks [69.49429223251178]
  This work presents the development of a hardware accelerator for an SNN, with off-line training, applied to an image recognition task. The design targets a Xilinx Artix-7 FPGA, using in total around the 40% of the available hardware resources. It reduces the classification time by three orders of magnitude, with a small 4.5% impact on the accuracy, if compared to its software, full precision counterpart.
  arXiv  Detail & Related papers  (2022-01-18T13:59:22Z)
• MAPLE: Microprocessor A Priori for Latency Estimation [81.91509153539566]
  Modern deep neural networks must demonstrate state-of-the-art accuracy while exhibiting low latency and energy consumption. Measuring the latency of every evaluated architecture adds a significant amount of time to the NAS process. We propose Microprocessor A Priori for Estimation Estimation MAPLE that does not rely on transfer learning or domain adaptation.
  arXiv  Detail & Related papers  (2021-11-30T03:52:15Z)
• An In-Memory Analog Computing Co-Processor for Energy-Efficient CNN Inference on Mobile Devices [4.117012092777604]
  We develop an in-memory analog computing (IMAC) architecture realizing both synaptic behavior and activation functions within non-volatile memory arrays. Spin-orbit torque magnetoresistive random-access memory (SOT-MRAM) devices are leveraged to realize sigmoidal neurons as well as binarized synapses. A heterogeneous mixed-signal and mixed-precision CPU-IMAC architecture is proposed for convolutional neural networks (CNNs) inference on mobile processors.
  arXiv  Detail & Related papers  (2021-05-24T23:01:36Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.