New framework enhances fault tolerance in FPGA-based CNN accelerators

Researchers have developed ProWAFT, a novel fault-tolerance framework designed for CNN accelerators implemented on SRAM-based FPGAs. This system addresses the challenge of transient faults that can compromise reliability in edge computing environments. ProWAFT utilizes partial reconfiguration to dynamically apply Triple Modular Redundancy (TMR) across reconfigurable partitions, balancing workload criticality, fault propagation, and reconfiguration overhead to optimize latency, energy, and reliability. AI

IMPACT Enhances the reliability of AI inference hardware at the network edge, crucial for real-time applications.

RANK_REASON The item is an academic paper detailing a new technical framework for hardware accelerators. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• AMD Zynq UltraScale+
• CNN
• EfficientNet Lite
• field-programmable gate array
• MobileNetV2
• ProWAFT
• ResNet-18
• SRAM
• Xilinx
• ZCU104

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →