BitFair通过动态截断和自适应比特排序，实现超低功耗下毫秒级推理。
BitFair: A 12-nm Bit-Serial CNN Accelerator with Learnable Early Termination and Adaptive Bit Ordering for Ultra-Low-Power XR Vision
- 根据每层特征学习比特级提前终止阈值，减少无效计算
- 在IBM DVS128和N-MNIST上分别达96.5%和97.7%准确率
- 比已有方案能效提升4.0-22.1倍，适合XR设备实时视觉任务
扩展现实（XR）可穿戴设备需在几瓦功耗内实现持续感知，并满足低于20毫秒的端到端延迟，留给神经网络推理的时间不足几毫秒。位串行计算对这种高能效需求具有吸引力，但现有架构仍会处理所有比特，即使ReLU已确定输出为零。本文提出BitFair，一种软硬件协同设计的位串行卷积神经网络加速器，具备可学习的比特级提前终止与自适应比特排序能力，适用于超低功耗和严格延迟约束的XR场景。该设计通过学习每层阈值，在部分求和已能可靠预测最终ReLU输出为零时触发提前终止，利用动态比特级稀疏性。同时，搜索每层最优比特顺序以优先处理信息量高的比特，最大化提前终止率而不损失精度。基于格罗弗12纳米鳍式晶体管工艺实现，核心面积0.34 mm²，片上内存104 KB，工作电压0.55–0.70 V，达到亚毫秒延迟，最高能效达117.0 BTOPS/W，每操作功耗仅0.07 pJ/SOP。在IBM DVS128 Gesture和N-MNIST数据集上分别取得96.5%和97.7%准确率，相比先前已流片的XR视觉加速器，有效能效提升4.0–22.1倍，准确率最高提升9.2%。
原文摘要 · Abstract (English)
Extended Reality (XR) wearables require always-on perception within tight power envelopes of a few watts and motion-to-photon latency budgets below 20 ms, leaving only a few milliseconds for neural-network inference. Bit-serial computing is attractive for such energy-efficient neural network acceleration, but many existing architectures still process all bits even when ReLU sets the final output to zero. This paper presents BitFair, a software-hardware co-designed bit-serial CNN accelerator with learnable bit-level early termination and adaptive bit ordering, working under the ultra-low-power and strict latency requirements of XR applications. BitFair exploits dynamic bit-level sparsity by learning per-layer thresholds that trigger early termination when partial sums reliably predict that the final ReLU output will be zero. Furthermore, it searches for layer-wise bit orders that prioritize informative bits, maximizing early termination without sacrificing accuracy. A GlobalFoundries 12-nm FinFET implementation with a core area of 0.34 mm^2, 104 KB on-chip memory, and voltage scaling from 0.55 to 0.70 V achieves sub-millisecond latency, up to 117.0 BTOPS/W, and 0.07 pJ/SOP. On IBM DVS128 Gesture and N-MNIST, BitFair achieves 96.5% and 97.7% accuracy, respectively, while improving effective energy efficiency by 4.0-22.1x and accuracy by up to 9.2% over prior fabricated XR vision accelerators.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。