SemiAnalysis· @SemiAnalysis_ · X·· 3 天前AI 评分47
AI 导读
SemiAnalysis 称，AMD Anush Elangovan 团队通过调优 AITER 内核和 MonoKernel decode 等软件优化，在 Kimi K3 智能体推理上实现每美元性能最高达 B200 的 2 倍。MI355X 每服务器配备 2304GB HBM，可避免跨服务器的 8x400Gbit/s 慢速 scale-out 组网。
AI 生成摘要 · 以原文为准
正文
ALERT🚨🚨🚨: AMD's @AnushElangovan and his team have achieved up to 2x better performance per dollar than even B200 on Kimi K3 Agentic Inference! This is due to various software optimization efforts, including tuning AITER kernels and MonoKernel decode.
Along with MI355X having 2304GB of HBM per server, this allows it to avoid crossing slow 8x400Gbit/s scale-out networking between servers, which B200 needs to do due to having less than 1.5TB of HBM.
Great work to Anush, Peng Sun, Ramine, Emad, and the vLLM team for this amazing work!
来源：SemiAnalysis · x.com