SC26上AMD质疑Nvidia的FP64模拟,强调Instinct MI430X原生双精度价值
AMD attacks Nvidia's FP64 emulation at SC26, touts native double-precision on Instinct MI430X — says Ozaki schemes fit less than 10% of HPC workloads and triple memory use and emulated FP64 is a poor fit for many HPC workloads
AI 导读
在芝加哥开幕的SC26上,AMD宣称Nvidia依赖Ozaki方案模拟FP64并非原生双精度的合适替代,理由是其假设负载类似稠密GEMM,而按AMD的应用剖析,HPC工作负载中只有不到10%执行稠密GEMM操作,且模拟会将内存占用放大至最高三倍。