Mitigating Compiler Fusion-Induced Power Bursts in Mobile NPU Inference as the Battery Depletes

• URL: http://arxiv.org/abs/2607.16555v1
• Date: Fri, 17 Jul 2026 23:44:41 GMT
• Title: Mitigating Compiler Fusion-Induced Power Bursts in Mobile NPU Inference as the Battery Depletes
• Abstract summary: Mobile devices increasingly rely on real-time NPU inference for camera and perception workloads.<n>A single inference can induce an instantaneous voltage droop in the power-delivery network, causing the power management integrated circuit to invoke dynamic voltage and frequency scaling (DVFS) and increase latency.<n>We show that aggressive operator fusion in a mobile NPU compiler can create monolithic superlayers whose concentrated execution produces large peak-current bursts.
• Score: 49.34452620841728
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Mobile devices increasingly rely on real-time NPU inference for camera and perception workloads. Under low-voltage conditions, however, a single inference can induce an instantaneous voltage droop in the power-delivery network, causing the power management integrated circuit to invoke dynamic voltage and frequency scaling (DVFS) and increase latency. We present a measurement study of this effect on a commercial smartphone. We show that aggressive operator fusion in a mobile NPU compiler can create monolithic superlayers whose concentrated execution produces large peak-current bursts. These bursts shift the DVFS-onset voltage upward and reduce the low-voltage operating margin. We further evaluate a practical black-box mitigation: a measurement-guided, pre-compilation graph rewrite that inserts barriers at selected peak-to-average power ratio hot spots to prevent harmful superlayer merging in the vendor NPU compiler. On Snapdragon 8 Gen 3 with MobileNetV4 at 768 x 768 resolution on ImageNet-1k, this method reduces peak current from 3.12 A to 1.94 A with 3.76% latency overhead, preserves stable latency deeper into the low-voltage regime, and shifts the inferred DVFS margin by approximately 173 mV.

Related papers

• Operator Fusion for LLM Inference on the Tensix Architecture [13.215957873912181]
  This study addresses on-device inference of Transformer models on Tenstorrent's Tensix architecture.<n> RMSNorm is fused with matrix multiplication in self-attention and in the FFN, enabling back-to-back execution of memory-bound and compute-bound operators.
  arXiv  Detail & Related papers  (2026-06-03T02:48:34Z)
• SparseDVFS: Sparse-Aware DVFS for Energy-Efficient Edge Inference [24.06892697687466]
  This paper presents SparseDVFS, a fine-grained, sparse-aware DVFS framework for energy-efficient edge inference.<n>Our key insight is that operator sparsity is a primary metric for hardware frequency modulation.<n>SparseDVFS achieves an average 78.17% energy efficiency gain over state-of-the-art solutions.
  arXiv  Detail & Related papers  (2026-03-23T12:29:30Z)
• GAVINA: flexible aggressive undervolting for bit-serial mixed-precision DNN acceleration [1.168664284092282]
  Voltage overscaling, or undervolting, is an enticing approximate technique in the context of energy-efficient Deep Neural Network (DNN) acceleration.<n>We propose a new technique called Guarded Aggressive underVolting (GAV)<n>GAV supports arbitrary mixed precision and flexible undervolting, with an energy efficiency of up to 89 TOP/sW in its most aggressive configuration.
  arXiv  Detail & Related papers  (2025-11-28T14:06:08Z)
• Speeding Up MACE: Low-Precision Tricks for Equivarient Force Fields [51.95157731126864]
  Machine-learning force fields can deliver accurate molecular dynamics (MD) at high computational cost.<n>This thesis aims to make MACE cheaper and faster by identifying computational bottlenecks and evaluating low-precision execution policies.
  arXiv  Detail & Related papers  (2025-10-23T14:02:34Z)
• RAPID^3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformer [86.57077884971478]
  Diffusion Transformers (DiTs) excel at visual generation yet remain hampered by slow sampling.<n>We introduce RAPID3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformers.<n>It delivers image-wise acceleration with zero updates to the base generator.<n>It achieves nearly 3x faster sampling with competitive generation quality.
  arXiv  Detail & Related papers  (2025-09-26T13:20:52Z)
• Emulating Self-attention with Convolution for Efficient Image Super-Resolution [6.8410780175245165]
  We introduce a convolutionized self-attention module named Convolutional Attention(ConvAttn)<n>By utilizing the ConvAttn module, we significantly reduce the reliance on self-attention and its involved memory-bound operations.<n>We overcome the challenge of integrating flash attention into the lightweight SR regime, effectively mitigating self-attention's inherent memory bottleneck.
  arXiv  Detail & Related papers  (2025-03-09T15:45:53Z)
• Optimizing the Deployment of Tiny Transformers on Low-Power MCUs [12.905978154498499]
  This work aims to enable and optimize the flexible, multi-platform deployment of encoder Tiny Transformers on commercial MCUs. Our framework provides an optimized library of kernels to maximize data reuse and avoid data marshaling operations into the crucial attention block. We show that our MHSA depth-first tiling scheme reduces the memory peak by up to 6.19x, while the fused-weight attention can reduce the runtime by 1.53x, and number of parameters by 25%.
  arXiv  Detail & Related papers  (2024-04-03T14:14:08Z)
• Real-Time Image Demoireing on Mobile Devices [59.59997851375429]
  We propose a dynamic demoireing acceleration method (DDA) towards a real-time deployment on mobile devices. Our stimulus stems from a simple-yet-universal fact that moire patterns often unbalancedly distribute across an image. Our method can drastically reduce the inference time, leading to a real-time image demoireing on mobile devices.
  arXiv  Detail & Related papers  (2023-02-04T15:42:42Z)
• BottleFit: Learning Compressed Representations in Deep Neural Networks for Effective and Efficient Split Computing [48.11023234245863]
  We propose a new framework called BottleFit, which includes a novel training strategy to achieve high accuracy even with strong compression rates. BottleFit achieves 77.1% data compression with up to 0.6% accuracy loss on ImageNet dataset. We show that BottleFit decreases power consumption and latency respectively by up to 49% and 89% with respect to (w.r.t.) local computing and by 37% and 55% w.r.t. edge offloading.
  arXiv  Detail & Related papers  (2022-01-07T22:08:07Z)
• EdgeBERT: Sentence-Level Energy Optimizations for Latency-Aware Multi-Task NLP Inference [82.1584439276834]
  Transformer-based language models such as BERT provide significant accuracy improvement for a multitude of natural language processing (NLP) tasks. We present EdgeBERT, an in-depth algorithm- hardware co-design for latency-aware energy optimization for multi-task NLP.
  arXiv  Detail & Related papers  (2020-11-28T19:21:47Z)
• An Experimental Study of Reduced-Voltage Operation in Modern FPGAs for Neural Network Acceleration [9.06484009562659]
  Undervolting below a safe voltage level can lead to timing faults due to excessive circuit latency increase. We experimentally study the reduced-voltage operation of multiple components of real FPGAs. We propose techniques to minimize the drawbacks of reduced-voltage operation.
  arXiv  Detail & Related papers  (2020-05-04T22:59:07Z)

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.