TensorLean, from Columbia University and CertiK, formally verifies production GPU kernels at the PTX level using Lean. It checked 149 kernel instances, proved 83 correct, and found 38 bugs in FlashAttention, DeepGEMM an…
算力供给、能源与数据中心容量决定了 AI 运行的成本。基础设施的变化会在数周后体现在推理成本上。
摘要仅供参考,请点击来源链接查看全文。演示条目为示意。