Silent data corruption (SDC) threatens the reliability of large-scale GPU clusters used for training large language models, yet its rarity and lack of explicit error signals make accurate high-level modeling challenging. To address this gap, we conducted a large-scale gate-level stuck-at fault injection on a production-class data-center GPU, consuming over three million simulator hours across 63 CUDA micro-benchmarks. We extracted GPU SDC characteristics in terms of corruption types, bit-flip behavior, and warp-aligned spatial correlation. Our results show that NaN/+INF/-INF account for only 1.01% of SDC outcomes, that single-bit flips constitute less than 40% of bit-flip events, and that corruption addresses exhibit periodicity. These statistics motivate distribution-aware high-level fault modeling and realistic software-based fault injection for resilience evaluation of production-class GPU architectures.

翻译：静默数据损坏（SDC）威胁着用于训练大语言模型的大规模GPU集群的可靠性，然而其罕见性及缺乏明确错误信号的特点使得精确的高层级建模面临挑战。为弥补这一空白，我们在生产级数据中心GPU上开展了大规模门级固定型故障注入实验，在63个CUDA微基准测试中累计消耗超过三百万模拟器小时。我们提取了GPU SDC在损坏类型、比特翻转行为及经线对齐空间相关性方面的特征。结果表明：NaN/+INF/-INF仅占SDC结果的1.01%，单比特翻转占比特翻转事件的不足40%，且损坏地址呈现周期性。这些统计数据为面向生产级GPU架构的分布感知型高层级故障建模及基于真实场景的软件故障注入实现弹性评估提供了动机。